[Frances Christie, J. R. Martin] Language, Knowled(Bookos.org)

Language, Knowledge and Pedagogy

This page intentionally left blank

Language, Knowledge and Pedagogy

Functional Linguistic and Sociological Perspectives

Edited by Frances Christie and J.R. Martin

ContinuumThe Tower Building 80 Maiden Lane 11 York Road Suite 704London SE1 7NX New York, NY 10038

© Frances Christie, J.R. Martin and contributors 2007

All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying,recording, or any information storage or retrieval system, without prior permission inwriting from the publishers.

British Library Cataloguing-in-Publication DataA catalogue record for this book is available from the British Library.

Published in paperback 2008

ISBN: 978-1-8470-6572-8

Library of Congress Cataloging-in-Publication DataTo

Typeset by Kenneth Burnley, Wirral, CheshirePrinted and bound in Great Britain by MPG Books, Bodmin, Cornwall

For Basil Bernstein, Michael Halliday and Ruqaiya Hasan,

in honour of the conversations they pursued.


Contents

List of tables viiiList of figures ixContributors xi

Section 1: Theoretical Foundations

1 Ongoing dialogue: functional linguistic and Bernsteinian sociological perspectives on education 3Frances Christie

2 A sociology for the transmission of knowledges 14Karl Maton and Johan Muller

3 Construing knowledge: a functional linguistic perspective 34J.R. Martin

4 On splitting hairs: hierarchy, knowledge and the school curriculum 65Johan Muller

5 Knowledge–knower structures in intellectual and educational fields 87Karl Maton

6 Hierarchical knowledge structures and the canon: a preference for judgements 109Rob Moore

Section 2: Fields of Discourse – Disciplines of Discourse

7 Language for learning in early childhood 131Clare Painter

8 Building verticality in subject English 156Frances Christie and Mary Macken-Horarik

9 Vertical and horizontal discourse and the social sciences 184Peter Wignell

10 Mathematical and scientific forms of knowledge: a systemic functional multimodal grammatical approach 205Kay O’Halloran

Section 3: Research Prospects – Exploring Uncommon Sense

11 Taking stock: future directions in research in knowledge structure 239Frances Christie, J.R. Martin, Karl Maton and Johan Muller

Index 259

List of tables

2.1 The arena of the pedagogic device 183.1 Factoring modernist history as genre 584.1 Maths content for grade 6 818.1 Models of school English in contemporary schools 1688.2 Semantic features with examples 178

List of figures

3.1 Strata and metafunctions in SFL 353.2 Social context as register and genre 353.3 Classification taxonomy (types of rock) 363.4 Compositional relations (volcanoes) 373.5 Common and uncommon sense fields 373.6 Genealogical organization of living things 393.7 Cochlear ear implant 403.8 Narrative image illustrating the cold front explanation 423.9 Stratal harmony – grammar matching semantics 523.10 Grammatical metaphor as stratal tension 533.11 Projection relations in pedagogic discourse 573.12 Typology of history genres 594.1 Kepler’s star tower 765.1 The two cultures as knowledge structures and knower structures 92 5.2 Legitimation codes of specialization for the two cultures 945.3 Legitimation codes of specialization 975.4 University students’ survey 1015.5 University students’ perceptions of basis of achievement 1027.1 Context-bound language for categorizing 1417.2 ‘Self-contextualised’ language for categorizing 1417.3 Grammatically congruent expression of meaning 1467.4 Defining clause with a metaphorical participant 1467.5 A definition involving grammatical metaphor 1479.1 Taxonomy of value (Smith and Ricardo) 1919.2 Taxonomy of value (Marx) 195

10.1a Bollmann map of New York 20810.1b Stable city landscape 20810.1c Horizontal and vertical lines 20910.2 Scientific experiment with frogs 21710.3 Newton: space, time, matter and prediction 21810.4 Newton: revisiting the moment mathematically 22010.5 Newton: suspending the instant 22210.6a Suspending the visual image: Newton 22310.6b Suspending the visual image: Prigent 22410.6c Suspending the visual image: Golovin 225


Contributors

Frances Christie is Emeritus Professor of Language and Literacy Education, theUniversity of Melbourne, and Honorary Professor of Education, University ofSydney. She has research interests in Language and Literacy Education gener-ally, including language curriculum theory and pedagogy, the history of Englishteaching, oral language and literacy development across the years of schooling,pedagogic grammar and classroom discourse analysis. Her publications haveincluded (edited with J. R. Martin), Genre and Institutions: Social Processes in theWorkplace and School, Cassell 1997; (edited with with Ray Misson), Literacy andSchooling, Routledge 1998; Pedagogy and the Shaping of Consciousness: Linguisticand Social Processes, Cassell 1999; Classroom Discourse Analysis. A Functional Perspec-tive, Continuum 2002; Language Education in the Primary Years, University of NSWPress 2005.

Mary Macken-Horarik is a Senior Lecturer in the School of Teacher Educationwithin the Division of Communication and Education at the University ofCanberra. She has worked for many years in the field of English and literacyeducation as a teacher, curriculum developer and researcher, with interests inteaching English as a mother tongue and as a second language. Mary has pub-lished widely in the field of systemic functional linguistics. Recent publicationsinclude ‘Negotiating Heteroglossia’ (a special issue of Text edited with J.R.Martin), Mouton de Gruyter 2003. Her other research interests include theanalysis of curriculum in school English, academic literacy education and mul-timodal discourse analysis. She is currently writing a book for English teachersabout systemic functional semiotics.

J.R. Martin is Professor of Linguistics (Personal Chair) at the University ofSydney. His research interests include systemic theory, functional grammar, dis-course semantics, register, genre, multimodality and critical discourse analysis,focusing on English and Tagalog – with special reference to the transdisciplin-ary fields of educational linguistics and social semiotics. Recent publicationsinclude Working with Discourse (with David Rose), Continuum 2003; Re/Readingthe Past (edited with Ruth Wodak), Benjamins 2003; ’Negotiating Heteroglossia’(a special issue of Text edited with Mary Macken-Horarik), Mouton de Gruyter2003; Language Typology: A Functional Perspective (edited with A. Caffarel andC. Matthiessen), Benjamins 2004; and ‘Interpreting Tragedy: the Language ofSeptember 11th, 2001’ (a special double issue of Discourse and Society edited withJohn Edwards), Sage 2004; The Language of Evaluation (with Peter White),Palgrave 2005. He has recently completed a book on genre (with David Rose, in

xii

press with Equinox). Professor Martin was elected a fellow the AustralianAcademy of the Humanities in 1998, and awarded a Centenary Medal for hisservices to Linguistics and Philology in 2003.

Karl Maton is Senior Research Fellow in Sociology and Education at the Uni-versity of Wollongong, Australia. He recently completed his doctoral thesis atthe University of Cambridge, ‘The field of higher education: A sociology ofreproduction, transformation, change and the conditions of emergence forcultural studies’, which develops Bernstein’s approach to create a sociology ofhigher education (http://www.KarlMaton.com). Karl has published in sociol-ogy, cultural studies, philosophy, linguistics and education, is co-editing avolume on cultural studies and global education, and is currently engaged incollaborative research into IT education, special educational needs, schoolmusic and the sociology of health.

Rob Moore is Senior Lecturer in Sociology of Education in the Faculty of Edu-cation of the University of Cambridge, UK. He is a fellow of Homerton Collegeand Director of Studies for Social and Political Sciences. His work includesresearch into youth and the labour market, educational decision-making, andissues of citizenship among school pupils and their parents and undergraduates(with John Beck and John Ahier). His current work is concerned with problemsin the sociology of knowledge and the structuring of intellectual fields. His mostrecent book is Education and Society: Issues and Explanations in the Sociology ofEducation (Polity 2004).

Johan Muller holds the chair of Curriculum at the University of Cape Town. Heis currently the Director of the Graduate School of Humanities and DeputyDean for Research and Postgraduate Affairs in the Faculty of Humanities. Hehas published in the areas of curriculum theory and policy, and in the sociologyof knowledge. His book Reclaiming Knowledge was published by RoutledgeFalmerin 2000.

Kay O’Halloran is Associate Professor in English linguistics and multimodalityat the National University of Singapore. Her teaching and research interestsinclude multimodality; systemic functional linguistics (SFL); mathematical andscientific discourse; multimodal approaches to hypertext, film and classroomdiscourse; digital media, visualization and technology. Her research interestsextend to the development of software for linguistic and multimodalanalysis. She has presented invited papers on many of these topics in interna-tional congresses in many parts of the world. She produced Systemics 1.0 softwarein 2002 with Kevin Judd, and recent publications include Multimodal DiscourseAnalysis (as editor), Continuum 2004, and Mathematical Discourse: Language,Symbolism and Visual Images, Continuum 2005.

Clare Painter is a Senior Lecturer in the School of English at UNSW, Australia.She has a longstanding research interest in systemic-functional (SF) linguisticsin relation to children’s language and literacy development and is the author of

CONTRIBUTORS

http://www.KarlMaton.com

a number of articles and books in the field, including Learning Through Languagein Early Childhood, Continuum 2000. Recent work in SF theory focuses on thesemiotics of children’s picture books and informs her forthcoming book,Reading Visual Narratives, to be published by Equinox.

Peter Wignell is Head of School of the Centre for Access and English as aSecond Language at Charles Darwin University in the Northern Territory ofAustralia. He has researched in the fields of linguistics, language and literacyeducation for the past 20 years and is currently working on a book on theevolution of the language of the social sciences.

xiiiCONTRIBUTORS


Section 1

Theoretical Foundations


Introduction

Over the last 45 to 50 years, systemic functional linguistic (SFL) theory andBernsteinian sociology have engaged in fruitful dialogue, exploring a numberof educational questions, each enhancing what the other discipline has to offer.It was perhaps an unlikely dialogue between two rather different disciplines,though those most instrumental in forging what became a very productivealliance had more in common than they might first have realized. The twofigures most fundamentally involved in the genesis of the dialogue thatemerged were the linguist M.A.K. Halliday and the sociologist Basil Bernstein,though they were soon joined by the linguist Ruqaiya Hasan, who went to workwith Bernstein in the 1960s. Halliday and Bernstein had both served in theforces in the Second World War, albeit in rather different capacities, and in theaftermath of the war both strove to make a difference in their respective areasof work by addressing issues of social justice, Bernstein initially as a schoolteacher, later a sociologist, Halliday as a teacher of Chinese, later of linguistics.

While in the 1950s Bernstein was troubled by issues of disadvantage whenobserving how working-class children performed badly in London schools,Halliday, teaching at Cambridge, was an active participant in the LinguisticsGroup of the British Communist Party, and like his colleagues he had a passionto use his linguistics to make a difference in exploring the nature of social expe-rience and in addressing questions of equity and social justice. In these waysHalliday and Bernstein thus had much in common, though there is at least oneother sense in which the two were to share a common fate. It was that each wasa remarkably original thinker and, as such, both were destined to be often mis-represented and misunderstood. Halliday (1986:1) once said that ‘one of theproblems with Bernstein was that he never fitted people’s stereotypes, theirready-made categories into which all thinkers are supposed to fit. So the leftbranded him as right-wing and the right-wing branded him as left-wing.’ WhileHalliday never made any secret of his left-wing sympathies, he was, like Bern-stein, often misrepresented and misunderstood, because his linguistics hasnever fitted what have been regarded as mainstream traditions of linguistics.Unlike linguists in more formalist traditions, he developed a linguistic paradigmsaid to be ‘extravagant’ principally because it engages with meaning andfunction as well as structure (thereby allowing a degree of messiness that formaltraditions don’t accept), and builds a strong sense of the social into accounts oflanguage.

1 Ongoing dialogue: functional linguistic and Bernsteinian sociological perspectives on education

Frances Christie

4

By the time of his death in 2000, Bernstein’s work had begun to receive amuch needed reassessment (e.g. Atkinson et al. 1995; Sadovnik 1995; Christie1999; Morais et al. 2001; Muller et al. 2004; Hasan 2005), and numbers ofscholars (some of whom contribute to this volume) continue to work with andpursue his theories and ideas. As for Halliday, he remains productive, whilerecent publications bear testimony to the major position he now occupies andthe major contribution he has made (e.g. Hasan et al. 2005; The Collected Worksof M.A.K. Halliday, edited Webster, from 2002). Thus, while Bernstein andHalliday both emerged as significant intellectuals in the often troubled years ofthe mid to later twentieth century, their scholarly contributions continue to berelevant and timely in the equally troubled years of the early twenty-first century.This book, then, seeks to maintain and extend the dialogue between two impor-tant traditions of scholarship by pursuing some significant educational issues.

Of course both Bernstein and Halliday pursued many questions other thaneducational ones. Bernstein regarded himself primarily as a sociologist, thoughhis was a Chair in the Sociology of Education, and Halliday was one of the twomost important linguists of the second half of the twentieth century. At the sametime, educational questions and issues have always been central to their respectiveconcerns, even where teaching and learning are not at the forefront of attention.

For Halliday, educational processes are part of the very warp and weft of life,and educational sites constitute major contexts for close analysis of language inuse. In addition – a lifelong preoccupation of his – language is an essentialresource in all teaching and learning. For Bernstein, pedagogic activity becameover time a major preoccupation, causing him to propose a social theory regard-ing cultural transmission and the manner in which different social groupsencode and communicate significant meanings and values. While both theoriststhus worked ‘beyond the classroom’, as it were, their concerns as linguist and associologist have remained bound up with questions of educational activity in thebroadest sense.

The concerns of this volume are with the nature of knowledge, a topic thatintrigued Bernstein in the last years of his life, though as Moore and Maton(2001: 154–5) have argued, this represented a necessary development upon theearlier work that Bernstein had undertaken over many years. For Bernstein (e.g.1990: 165–218), as he reflected on sociology of education dating from the1970s, what was missing in the various accounts available was any sense of thestructuring of knowledge. Sociology of education, he argued, had become inter-ested in matters to do with ‘relations (of educational knowledge) to’ otherthings, such as class, race or gender. It had lost sight of what he termed alter-natively ‘relations within’ knowledge, and thus had little to say about the natureof the knowledge transmitted in pedagogic enterprises. What was needed wasthe development of a theory that would deal with the structuring of knowledgeitself. The challenge he thus offered to sociologists was a bold one, and a verytimely one, for reasons to be developed in this volume.

To begin, however, it will be useful to sketch a little of the history of the emer-gence of the two traditions of scholarship involved here, and the sense in whichthey have joined together in common cause, while not overlooking the differ-ent methodologies each employs.

LANGUAGE, KNOWLEDGE AND PEDAGOGY

Developing a dialogue

Halliday and Bernstein did not meet until the early 1960s, when Halliday wasteaching linguistics at Edinburgh University and Bernstein was at the Universityof London. Halliday’s interest in linguistics developed from his work onChinese, which he taught at Cambridge, moving to linguistics when he went toEdinburgh. While he acknowledged a great debt to Firth in developing his firstaccounts of linguistic theory, he later observed that the work he and his col-leagues did with teachers while at Edinburgh was very significant (Halliday andHasan, in press) in shaping his early thinking both about language, and aboutlanguage education. In a subsequent move he made to the University ofLondon he was invited to develop an English curriculum proposal for theNuffield Foundation, and this led in time to the Nuffield/Schools Council Pro-gramme in Linguistics and English Teaching (1964–71). The latter programmeproduced extensive sets of curriculum materials for schools and resource booksfor teachers, all of which were influential in their day, and some of which remainin use. (See Hasan and Martin 1989 and Christie and Unsworth 2005.)

Early work on a theory of grammar (e.g. Halliday, 1976, 2002) led to a modelof grammar as systemic and functional. It is systemic, in that it offers sets of optionsor choices (not conscious) for making meaning; the choices provide a largenetwork of systems of choices, each with an entry condition. In practice, simul-taneous choices are said to apply when a language is used; in creating a clause,for example, one makes choices for transitivity, mood and theme. Language isfunctional, it was said, in that it serves basic human functions to represent expe-rience and knowledge, to construct relationship, and to create meaningfulmessages. The functional aspect of the theory evolved into a theory of metafunc-tions which operate across all natural languages: the ideational, interpersonal andtextual. The structures in which these metafunctions are realized are particular toany given language. Hence, in parsing and labelling of language, two kinds oflabels apply, functional and class. Class labels in English, for example, identifynouns, verbs, adverbs and so on, while functional labels identify, say, for tran-sitivity, participant, process and circumstance, and in each case, the type ofparticipant, process and circumstance is also specified.

And what of context? Working with notions of context of situation and later ofcontext of culture, taken from Malinowski (discussed by Halliday in Halliday andHasan 1985) Halliday and his colleagues (Halliday et al. 1964) pursued a theoryof register variation: for any context of situation, language choices are in part acondition of the social activity (field), the nature of the relationships (tenor) andthe role that language plays (mode). Text and context are mutually intelligible, forthe one constructs the other. One of the many remarkable contributions thatHalliday and his colleagues made in the developing theory of language wasto propose a ‘hook up’ between grammar and register. Some sets of choicesrelate primarily to activity, some primarily to the nature of the relationshipin construction, and others primarily to organization of the language asmessage.

A theme to which Halliday has returned often is his interest in language as asocial phenomenon. He once noted (Parret 1974) that when he studied with

5LINGUISTIC AND BERNSTEINIAN PERSPECTIVES • CHRISTIE

6

Firth, linguistics was ‘the study of language in society’, and he regretted thedevelopment of the term ‘sociolinguistics’, as he later regretted the emergenceof ‘pragmatics’ (Halliday 2002: 8). For Halliday, the study of language necessar-ily involved the study of language in use, as a part of the social system. Whatmattered about language was what people did with it. It was this concern forlanguage as social phenomenon, functioning to mediate social processes inprofound ways, which aroused the interest of Halliday – and that of his col-league Hasan – in the early work of Bernstein.

Bernstein, troubled, as we noted above, by the apparent educational failureof many working-class children, was seeking some explanation that would allowhim to make a difference. While he was not sure what kind of language theoryhe needed, he early realized he must account for the apparent differences inuses of language that he and his colleagues observed between working-class andmiddle-class children; he cast about among the available linguistic theories,including various early formulations of sociolinguistics (see Bernstein 2000:145–53). His early formulations (e.g. 1971) led him to propose the operation ofrestricted and elaborated language codes. Possession of the elaborated code enabled thespeaker to select from a relatively wide range of meanings, allowing expressionof personal attitude among other matters, as well as the capacity to expandupon and explain experience, and this code was rewarded for the purposes ofschool learning. The restricted code was associated with meanings that were morelocalized and immediate to situations than the elaborated code, and in conse-quence they were less amenable to expansion and explanation of experience ofa kind that schooling tended to reward and encourage. The codes wereacquired as part of socialization within one’s family, and while working-classchildren tended to use the restricted code, both codes were available to middle-class children. The learning of a code was profoundly important, for it involvedlearning the requirements of the social structure (ibid.: 124).

Code theory, intimately linked as it was to theories of social class, led in timeto a more extensive theory addressing cultural transmission. By 1981, Bernsteinwas to say that a code was ‘a regulative principle, tacitly acquired, which selectsand integrates relevant meanings, forms of realizations and evoking contexts’.Hasan (1973) explored the differences between code, register and socialdialect, making clear how code was to be distinguished from the other two.Dialects represented patterns of language spoken according to social and/orgeographical location, while registers were distinguished by their use in differ-ent contexts of situation; but codes must be recognized for their semantic prop-erties, for they had to do with orientations to meaning. Later on, in developingand enriching discussions of the coding orientations associated with differentsocial groups, Hasan (1989) proposed the term ‘semantic variation’ to refer to‘systematic variation in the meanings people select in similar contexts as afunction of their social positioning’ (Williams 2005: 457). With colleagueswho included Cloran (1994, 1999; Hasan 1986; Hasan and Cloran 1990)and Williams (1999, 2001) Hasan provided significant evidence to supportBernstein’s original claims about the manner in which meaning orientationsfunctioned as a condition of social – more specifically family – positioning.Hasan collected natural discourse between mothers and their pre-school


children drawn from ‘high autonomy professionals’ (HAPs) and ‘low autonomyprofessionals’ (LAPs); using delicately developed semantic networks based onSF linguistics, she was able to show systematic differences in the coding orient-ations of mothers and children from the two social groups. Such detailedanalysis using the functional grammar had not been possible at the time Bern-stein had advanced his proposals in the late 1960s and 1970s, for the theory wasinsufficiently developed at that time. Bernstein (2002: 128–30) acknowledged asignificant debt to Hasan and her colleagues.

Code theory was accompanied in the 1970s by other and related develop-ments in identifying models of school knowledge, such that distinctions weremade between ‘collection’ and ‘integrated’ codes in the school curriculum, andbetween ‘visible’ and ‘invisible’ pedagogies (Bernstein 1975). Both distinctionsserved to extend and enrich his discussion of the nature of cultural transmissionand the processes by which knowledge was shaped and pedagogic subject posi-tions were formed. Indeed, code theory and its corollaries were subsumed intoa more comprehensive theory of cultural transmission in which pedagogicactivity achieved significance in the 1980s, when Bernstein turned his attentionto the production of knowledge, and to the ‘symbolic control’ that was a featureof pedagogy.

For Bernstein (1990: 134–41), pedagogic relationships embraced more thanthose found in educational settings, while there were many agents of symboliccontrol other than teachers, such as architects, scientists and psychologists.Pedagogy, he argued, functioned through the operation of a pedagogic device.This was instrumental in creating pedagogic discourse, which was defined as ‘aprinciple for appropriating other discourses and bringing them into a specialrelation with each other for the purposes of their selective transmission andacquisition’ (Bernstein, ibid.: 183–4). Pedagogic discourse takes the discoursesof practice in identified sites (such as universities) and relocates or recontextu-alizes these for other purposes (such as teaching in schools). The pedagogicdevice works through distributive rules (which distribute forms of consciousnessthrough the distribution of forms of knowledge); recontextualizing rules (whichregulate the construction of specific instructional discourses for specific peda-gogic purposes) and evaluative rules (which provide the criteria used to legiti-mate the forms of knowledge communicated). Overall, Bernstein argued,pedagogic discourse serves to shape consciousness, differentially distributingknowledge and experience. (See Maton and Muller, this volume.)

Yet as Bernstein had also determined by the end of his life, the various dis-cussions of pedagogy offered by sociologists (including his own) had had littleto say of the nature of the knowledge communicated in pedagogic activities(Bernstein 2000: 155–74). He therefore began to develop his discussions ofknowledge, proposing a theory of horizontal and vertical discourses, out of which atheory of knowledge structure emerged. Horizontal discourse is segmentallyorganized, in that its various forms are localized and particular to specific activ-ities, creating what is often thought of as ‘common sense’ knowledge. Verticaldiscourse, by contrast, is not segmentally organized, for it ‘takes the form ofa coherent, explicit and systematically principled structure’, and it is ‘hier-archically organized’ as in the sciences, or consists of ‘a series of specialized


8

languages’, of the kind found in the social sciences or humanities’ (ibid.: 157).Vertical discourse could be understood in terms of the operation of two know-ledge structures: horizontal and hierarchical. Horizontal knowledge structures (e.g.English literature, philosophy) are apparent in the presence of a series of ‘lan-guages’, possessed of ‘grammars’, some stronger or weaker than others; suchstructures by their nature are not easily related and they are often incommen-surable. Hierarchical knowledge structures (e.g. physics) by contrast involve formsof knowledge which attempt to ‘create very general propositions and theories,which integrate knowledge at lower levels’ and which show ‘underlying unifor-mities across an expanding range of apparently different phenomena’. In thissense, they are produced using an integrating code (Bernstein, ibid.: 161).Overall, there is an internal logic which shapes the hierarchical knowledgestructure, giving it a unity in terms of which new knowledge claims may betested. (See Muller and Maton in this volume.)

These discussions of vertical discourse are both powerful and interesting, butBernstein was aware that the subject was by no means fully explored by the timehe died. A great deal more work remains to be done in refining what is meantby the notion of vertical discourse, and in particular in refining the nature ofhorizontal and hierarchical knowledge structures. This is the case for at leasttwo related reasons. Firstly, the twentieth century saw the appearance of manynew forms of knowledge, especially at university level, having consequences forthe ways knowledge in general was understood, though it also had conse-quences for the ways knowledge was later recontextualized in schools. Manyonce orthodox areas of knowledge were challenged, while the relationships ofthe various areas of inquiry available were often poorly defined. Not all newforms of knowledge were of equal merit, as Moore and Muller (1999) haveargued with respect to sociology of education, and Maton (2000) has arguedwith respect to cultural studies, as well as sociology of education.

The second and related reason why there is an urgent need to pursue furtherwork on the nature of knowledge structures concerns the claims of the schoolcurriculum. The steady advance of compulsory education led to a great deal oftheorizing both about how knowledge should be conceived for school purposes,and about the most appropriate pedagogies to support its transmission.Curriculum theorizing throughout the twentieth century was significantlyinfluenced by various progressivist and constructivist theories of knowledge andof the learner, whose effect was to diminish the status of knowledge structures,as well as the role of the teacher (e.g. Muller 2002; Christie 2002, 2004).Currently, in many parts of the English-speaking world there is considerableuncertainty, both about what forms of knowledge should be taught, and aboutthe most appropriate teaching practices. Bernsteinian theory and SFL theorycan together usefully address the problem.

Just as the sociology of Bernstein and his colleagues was extended andenriched over the last several decades, so too was SFL theory. Halliday’s versionof the functional grammar appeared in 1985, to be followed by a second editionin 1994, while the third edition (2004) was produced with his colleagueMatthiessen. Various discussions of the SF grammar were in time produced byothers, while several versions for teaching had appeared by the end of the


century (see Christie and Unsworth 2005, for some discussion). Alongside thesedevelopments Martin and his colleagues developed earlier work on cohesion(Halliday and Hasan 1976) as a stratum of semantics, mediating betweengrammar and context (Martin 1992; Martin and Rose 2003; Martin and White2005). This contributed to the development of significant work on registertheory, which had been outlined by Halliday, McIntosh and Strevens (ibid.) asreferred to above, though many others later contributed, offering more thanone framework (e.g. Benson and Greaves 1973; Hasan 1973; Halliday 1973;Gregory and Carroll 1978; Halliday and Hasan 1985; Hasan 1995; Matthiessen1993; Martin 1992). The rival claims of the various accounts of register will notbe addressed here. Suffice to note that probably the most influential of themodels of register, at least for educational theory, was that of Martin (1992). Hiswork, developed initially with Rothery (see Martin and Rose 2003, in press)offered an account of register and of genre. Text types or genres are said todiffer, depending on the immediate context of situation (field, tenor, mode),and the context of culture (genre systems). Work in the tradition of genre-basedresearch has led to many discussions of ‘subject specific literacies’ (e.g. Macken-Horarik 1996, 1998; Unsworth 2000; Schleppegrell 2004; Martin 1999, thisvolume). Such work, stimulated in addition by the research of Halliday into dif-ferences between speech and writing (Halliday 1985) and grammaticalmetaphor (e.g. Simon-Vandenbergen, Taverniers and Ravelli 2003), hasproduced extensive accounts of different types of genres, spoken and written.An accompanying theory of pedagogy has also emerged, some representativeaccounts being given by Feez and Joyce (1998), Martin (1999), Christie (2002,2005), and Martin and Rose (2005).

Paralleling these developments has been important associated work on alanguage-based theory of learning (Halliday 1993, 2003) to which colleaguessuch as Painter (1999, this volume) and Torr (1998) have contributed. A signif-icant additional development has been the entry into transdisciplinary discus-sion with scholars working on the human brain and on the emergence oflanguage in the human species (e.g. Williams and Lukin 2004; Hasan 2005:21–156). This promises to be a productive direction for future research.

About this book

Systemic functional theory has had a strong tradition of working in dialoguewith other areas of scholarship, and such transdisciplinary work has manymerits, not least that many of the most significant problems worthy of examin-ation require more than one scholarly tradition to do them justice. This volumeaims to continue the transdisciplinary work that has been established involvingboth SFL theory and Bernstein’s sociology. The volume is transdiciplinary, inthat each area of scholarship seeks to inform the other, in a common concernto pursue the nature of knowledge structures. Neither tradition tries to subsumethe other, nor do the two overlook their methodological differences, includingdifferences in uses of terms. A case in point is use of the term ‘grammar’, which,we have already noted, Bernstein used in referring to the ‘weak’ or ‘strong


10

grammars’ he associated with different horizontal knowledge structures. Froma linguist’s point of view he used the term metaphorically, to refer to the under-lying organizing principles that characterize knowledge structures. However,the issue points to a more general difference between the two traditions ofscholarship. For the linguist, the object of study is language, though at least inthe SF tradition language is always understood as part of social process. For thesociologist, the object of study is social structure, and though in the Bernsteiniantradition language is a very significant resource in the creation, reproductionand transformation of social structure, other resources, including educationand knowledge, are also significant objects of study.

The book is arranged in such a way that the transdisciplinary dialoguebetween the two traditions is early established and maintained. This chapter hassought to sketch in some major developments in the emergence of the two tra-ditions, and something of the history of their previous partnerships. TheoreticalFoundations are established on the one hand by the sociologists Maton andMuller, who examine Bernstein’s work on codes, the pedagogic device andknowledge structures, clarifying the connections between these areas of thetheory. Then, on the other hand, Martin offers an SFL account, demonstratinghow the theory is used to explore and identify specialist forms of knowledge.

In the subsequent section devoted to Knowledge Structure, Muller, Maton andMoore all explore aspects of horizontal and vertical discourse, examiningproblems for conceptualizing knowledge, given the dichotomous nature of thetheory, the notion of a ‘knower structure’ introduced to parallel that of theknowledge structure, and difficulties in dealing with hierarchical knowledgestructures.

The next section, devoted to Fields of Discourse – Disciplines of Discourse, com-mences with a chapter by Painter in which she examines the development ofhorizontal discourse in young children as an aspect of dealing with the‘common-sense’ knowledge of daily life, also suggesting its limitations as theyoung push towards the need to master vertical discourse. Christie andMacken-Horarik, Wignell and O’Halloran all go on to address, respectively,knowledge structures in subject English, the social sciences, and mathematicsand natural sciences. O’Halloran in particular looks at the impact of multi-modality in the evolving nature of the sciences. All these chapters considersome of the problems as well as the strengths of characterizing knowledgestructure in Bernstein’s terms.

The concluding section, devoted to Research Prospects, explores some of theissues we have identified and suggests possible directions for future research.

References

Atkinson, P., Davies, B. and Delamont, S. (eds) (1995), Discourse and Reproduction. Essaysin Honor of Basil Bernstein. Kresskill, NY: Hampton Press.

Benson, J.D. and Greaves, W.S. (1973), The Language People Really Use. Agincourt: TheBook Society of Canada.

Bernstein, B. (1971), Theoretical Studies towards a Sociology of Language. Class, Codes andControl, Vol. 1. London, Henley and Boston: Routledge & Kegan Paul.


Bernstein, B. (1975), Towards a Theory of Educational Transmissions. Class Codes and Control,Vol. 3. London and Boston: Routledge.

Bernstein, B. (1981), ‘Code modalities and the process of reproduction: a model’, inLanguage and Society 10, pp. 327–63.

Bernstein, B. (1990), The Structuring of Pedagogic Discourse. Class, Codes and Control, Vol. 4.London and NY: Routledge.

Bernstein, B. (2000), Pedagogy, Symbolic Control and Identity. Theory, Research, Critique (rev.edn). Lanham, Boulder, NY and Oxford: Rowman & Littlefield.

Christie, F. (ed.) (1999), Pedagogy and the Shaping of Consciousness. Linguistic and SocialProcesses. London and NY: Cassell.

Christie, F. (2002), Classroom Discourse Analysis. A Functional Perspective. London and NY:Continuum.

Christie, F. (2004), ‘Authority and its role in the pedagogic relationship of schooling’, inL. Young and C. Harrison (eds), Systemic Linguistics and Critical Discourse Analysis.Studies in Social Change. London and NY: Continuum, pp. 173–201.

Christie, F. (2005), Language Education in the Primary Years. Sydney: University of NSWPress.

Christie, F. and Martin, J.R. (eds) (1997), Genre and Institutions. Social Processes in the Work-place and School. London and NY: Continuum.

Christie, F. and Unsworth, L. (2005), ‘Developing dimensions of an educational linguis-tics’, in R. Hasan, C. Matthiessen and J. Webster (eds), Continuing Discourse onLanguage. A Functional Perspective, Vol. 1. London and Oakville: Equinox, pp. 217–50.

Cloran, C. (1994), ‘Rhetorical units and decontextualization: an inquiry into somerelations of context, meaning and grammar’, in Monographs in Systemic Linguistics 6.Nottingham: Department of English Studies, University of Nottingham.

Feez, S. with Joyce, H. (1998), Text-based Syllabus Design. Sydney: National Centre forEnglish Language Teaching and Research, Macquarie University.

Gregory, M. and Carroll, S. (1978), Language and Situation. Language Varieties and theirSocial Contexts (Language and Society Series). London: Routledge & Kegan Paul.

Halliday, M.A.K. (ed. G. Kress) (1976), Halliday: System and Function in Language. London:Oxford University Press.

Halliday, M.A.K. (1988), ‘Language and socialization: home and school’, in L. Gerot,J. Oldenburg and T. van Leeuwen (eds), Proceedings from the Working Conference onLanguage in Education, held at Macquarie University, 17th–21st November 1986, pp. 1–24.

Halliday, M.A.K. (1985), An Introduction to Functional Grammar. London: Edward Arnold.Halliday, M.A.K. (1993), ‘Some grammatical problems in scientific English’, in M.A.K.

Halliday and J.R. Martin (eds), Writing Science: Literacy and Discursive Power. London:Falmer, pp. 69–85.

Halliday, M.A.K. (1993), ‘The Analysis of scientific texts in English and Chinese’, inM.A.K. Halliday and J.R. Martin, Writing Science: Literacy and Discursive Power. London:Falmer, pp. 124–32.

Halliday, M.A.K. (1994), An Introduction to Functional Grammar. London: Edward Arnold.Halliday, M.A.K. (1973), Explorations in the Functions of Language (Explorations in Language

Study). London: Edward Arnold.Halliday, M.A.K. (1975), Learning How to Mean. Explorations in the Development of Language.

(Explorations in Language Study). London: Edward Arnold.Halliday, M.A.K (1985), Spoken and Written Language. Geelong, Victoria: Deakin University

Press.Halliday, M.A.K. (1993), ‘Towards a language-based theory of learning’, in Linguistics and

Education 5, 2, pp. 93–116.Halliday, M.A.K. (ed. J.J. Webster) (2002), On Grammar. Collected Works of M.A.K. Halliday,

Vol. 1. London and NY: Continuum.Halliday, M.A.K. (ed. J.J. Webster) (2004), The Language of Early Childhood: M.A.K.

Halliday, Vol. 4. London and NY: Continuum.


12

Halliday, M.A.K. and Hasan, R. (1976), Cohesion in English. London: Longman.Halliday, M.A.K. and Hasan, R. (1985), Language, Context and Text: Aspects of Language in

a Social Semiotic Perspective. Geelong, Victoria: Deakin University.Halliday, M.A.K and Hasan, R. (in press), ‘Retrospective on SFL and literacy’,in R. Whittaker, M. O’Donnell and A. McCabe (eds), Language and Literacy: Functional

Approaches. Continuum: London.Halliday, M.A.K., McIntosh, A. and Strevens, P. (1964), The Linguistic Sciences and Language

Teaching. London: Longman.Halliday, M.A.K. and Matthiessen, C. (3rd edn) (2004), An Introduction to Functional

Grammar. London and NY: Arnold.Hasan, R. (1986), ‘The ontogenesis of ideology: an interpretation of mother child talk’,

in T. Threadgold, E.A. Gross, G. Kress and M.A.K. Halliday (eds), Language, Ideology,Language. Sydney: Sydney Association for Studies in Society and Culture, pp. 125–46.

Hasan, R. (1989), ‘Semantic variation and sociolinguistics’, in Australian Journal of Lin-guistics 9, pp. 221–75.

Hasan, R. (1995), ‘The conception of context in text’, in P.H. Fries and M. Gregory (eds),Discourse in Society: Systemic Functional Perspectives. Norwood,NJ: Ablex, pp. 183–283.

Hasan, R. (ed. J. Webster) (2005), Language, Society and Consciousness. The Collected Worksof Ruqaiya Hasan, Vol. 1. London and Oakville: Equinox.

Hasan, R. and Cloran, C. (1987), ‘A sociolinguistic study of everyday talk betweenmothers and children’, in M.A.K. Halliday, J. Gibbons and H. Nicholas (eds),Learning, Keeping and Using Language, Selected Papers from the 8th World Congress ofApplied Linguistics, Sydney, 16–21st August, Vol. 1. Amsterdam and Philadelphia:Benjamins, pp. 67–100.

Hasan, R. and Martin, J.R. (eds) (1989), Language Development: Learning Language,Learning Culture. Meaning and Choice in Language. Studies for Michael Halliday.Norwood, NJ: Ablex.

Hasan, R., Matthiessen, C. and Webster, J. (eds) (2005), Continuing Discourse on Language.A Functional Perspective, Vol. 1. London and Oakville: Equinox.

Macken-Horarik, M. (1996), ‘Literacy and learning across the curriculum: towards amodel of register for secondary school teachers’, in R. Hasan and G. Williams (eds),Literacy in Society. Harlow: Addison Wesley Longman, pp. 232–78.

Macken-Horarik, M. (1998), ‘Exploring the requirements of critical literacy: a view fromtwo classrooms’, in F. Christie and R. Misson (eds), Literacy and Schooling, London:Routledge, pp. 74–103.

Martin, J.R. (1992), English Text. System and Structure. Philadelphia and Amsterdam:Benjamins.

Martin, J.R. (1999), ‘Mentoring semogenesis: “genre-based” literacy pedagogy’, inF. Christie, pp. 123–55.

Martin, J.R. and Rose, D. (2003), Working with Discourse. London and NY: Continuum.Martin, J.R. and Rose, D. (2005), ‘Designing literacy pedagogy: scaffolding democracy in

the classroom’, in R. Hasan et al., pp. 251–80.Martin, J.R. and Rose, D. (in press), Genre Relations: Mapping Culture. London: Equinox.Martin, J.R. and White, P.R.R. (2005), The Language of Evaluation: Appraisal in English. NY:

Palgrave Macmillan.Maton, K. (2000), ‘Recovering pedagogic discourse: a Bernsteinian approach to the soci-

ology of educational knowledge’, in Linguistics and Education 11, 1, pp. 79–98.Maton, K. (2004), ‘The wrong kind of knower: education, expansion and the epistemic

device’, in Muller, Davies (eds) (2004), pp. 218–31.Matthiessen, C. (1993), ‘Register in the round: diversity in a unified theory of register

analysis’, in M. Ghadessy (ed.), Register Analysis: Theory and Practice. London: Pinter,pp. 221–92.

Moore, R. and Muller, J. (1999), ‘The discourse of “voice” and the problem of knowledge


and identity in the sociology of education’, in British Journal of Sociology of Education20, 2, pp. 189–206.

Morais, A., Neves, I., Davies, B. and Daniels, H. (eds) (2001), Towards a Sociology ofPedagogy. The Contribution of Basil Bernstein to Research. NY: Peter Lang.

Muller, J. (2002), Reclaiming Knowledge: Social Theory, Curriculum and Education Policy(Knowledge, Identity and School Life series). London and NY: RoutledgeFalmer.

Muller, J., Davies, B.A. and Morais, A. (eds) (2004), Reading Bernstein, Researching Bernstein.London: RoutledgeFalmer.

Painter, C. (1999), Learning Language in Early Childhood. London and NY: Cassell. Parret, H. (1974), Discussing Language. The Hague: Mouton, pp. 81–120.Sadovnik, A.R. (1995), Knowledge and Pedagogy: the Sociology of Basil Bernstein. Norwood, NJ:

Ablex.Schleppegrell, M. (2004), The Language of Schooling: A Functional Linguistic Perspective.

Mawah, NJ and London: Erlbaum.Simon-Vandenbergen, A.M., Taverniers, M. and Ravelli, L. (eds) (2003), Grammatical

Metaphor. Views from Systemic Functional Linguistics (Amsterdam Studies in the Theoryand History of Linguistic Science). Amsterdam and Philadelphia: Benjamins.

Torr, J. (1998), From Child Tongue to Mother Tongue: Language Development in the First Two anda Half Years (Monographs in Systemic Linguistics). Nottingham: Nottingham Uni-versity.

Unsworth, L. (ed.) (2000), Researching Language in Schools and Communities. FunctionalLinguistic Perspectives. London and Washington: Cassell.

Williams, G. (1999), ‘The pedagogic device and the production of pedagogic discourse:a case example in early literacy education’, in F. Christie (ed.), pp. 88–122.

Williams, G. (2001), ‘Literacy pedagogy prior to schooling: relations between social posi-tioning and semantic variation’, in Morais et al., pp. 17–45.

Williams, G. (2005), ‘Semantic variation’, in Hasan, Matthiessen and Webster, pp. 457–80.Williams, G. and Lukin, A. (eds) (2004), The Development of Language. Functional Perspec-

tives on Species and Individuals. London and NY: Continuum.


Why knowledge?

In two of his final papers Basil Bernstein codified and extended a conceptual-ization of the different structures of knowledge associated with intellectualfields (1999a) and signalled a more general move in his theory ‘from peda-gogies to knowledges’ (2001). In so doing he returned to a longstandinginterest in discourse, a focus that has brought the approach associated withBernstein’s sociology into regular and fruitful relations with systemic functionallinguistics. Such cross-disciplinary dialogue has been ongoing since Bernstein inhis early work adapted the linguistic notion of ‘code’ to his own sociologicalpurposes, refining it over the years into a highly formal analytical concept:

Thus a code is a regulative principle, tacitly acquired, which selects and integratesmeanings, forms of realizations, and evoking contexts. (Bernstein 1990: 101;emphasis in original)

It may be useful to revisit those sociological purposes, addressing Bernstein’sparticular appropriation of ‘code’, his analysis of the ‘pedagogic device’ and hissubsequent move ‘from pedagogies to knowledges’. The intention is to makemore comprehensible the distinctive nature of Bernstein’s preoccupation withforms of discourse of which knowledge is one. To enquire into Bernstein’s soci-ological roots, therefore, is to broach the question: why knowledge?

In this chapter we briefly sketch a background to Bernstein’s theorization ofknowledge. Such an account can be valuable not simply to place this later focusin the context of the unfolding of his wider theory. Many understandings ofBernstein’s work are stuck in two principal time warps, focusing on either hisearly interest in sociolinguistics (in particular the notions of elaborated andrestricted codes) or his account of schooling in terms of pedagogic codes. Bothfreeze the theory to a time before knowledge itself became increasingly centralto his thinking. In recounting its subsequent development, however, ours is ofnecessity a partial account, for several reasons immanent to the theory itself.One aspect of Bernstein’s method is that he was always reworking and recastinghis ideas. Throughout his career theoretical developments made visible newobjects of study for empirical research, which in turn required development ofthe theory and which then, in turn, raised further issues for research. A secondaspect of the theory’s development is a form of excavation; Bernstein dugbeneath the empirical features of education to explore their underlying struc-

2 A sociology for the transmission of knowledges

Karl Maton and John Muller

turing principles (most famously in terms of codes) and then excavated furtherto analyse what generates these principles. Bernstein was, therefore, alwaysengaged in developing more general conceptual tools in the light of what wasbeing revealed by both empirical research and theoretical excavation. Lastly, theresulting theory is driven by an abiding interest in social order and the nature ofsymbolic control, one reaching from the macro-structure of society to the micro-level of individual consciousness. To recount fully the development of thisevolving and wide-ranging theory could lead to the temptation of drawing a mapas big as the country. Instead, we shall simply trace one path through his work toilluminate how Bernstein came to conceptualize knowledge, explore the mainaspects of these ideas, and thereby provide one context to their use in the studiescollected in this volume.1 Our guiding thread is the questions raised by the devel-oping theory as it unfolded over time.

The term ‘social base’ provides an initial key, both to this trajectory and torelations between Bernstein’s approach and systemic functional linguistics. Toask about the sociological significance of any symbolic or linguistic ensemble –such as a curriculum, pedagogy or discourse – is to enquire after its social base,its grounding in a material social form of life. As Bernstein put it when describ-ing the nature of his early interest in linguistics:

Language was the structuring interface by means of which a complex set ofordering and disordering processes were specialised by the social base of itsspeakers. What was paramount for me was the identification of origins of theseordering and disordering processes, their maintenance and change. (2001: 363)

As a Durkheimian, Bernstein considered the principal features of this ‘socialbase’ to derive from the forms taken by the economic division of labour insociety and the consequent forms of solidarity accompanying them.2 BothDurkheim and Bernstein described modern industrialized societies as havingdeveloped from a relatively unspecialized division of labour to a highly special-ized division of labour. As a direct consequence, they tend also to develop frommechanical modes of solidarity to organic or more specialized and interdepend-ent forms of solidarity. In tandem, societies develop specialized symbolic formsto give effect to the new specializations emerging from this growing division oflabour. In a class-based society, these symbolic resources are not only differen-tially valued and resourced but also differentially regulated and distributed. Thesociological task is to uncover how that differential valorization, regulation anddistribution occurs and thereby to explore the means and mechanisms by whichthe underlying structures of a complex, specialized society such as ours perpet-uates itself, develops and changes. As Bernstein stated:

I think like Durkheim one can identify and make explicit the social base ofthe pedagogic relation, its various contingent realisations, the agencies andagents of its enactments. One can begin to formulate a language for the descrip-tion of the production and reproduction of its discourses. At a more general levelsuch a study connects with the maintenance and change of the knowledge baseof society, and crucially with the maintenance and change of modalities of

15SOCIOLOGY FOR TRANSMISSION OF KNOWLEDGES • MATON & MULLER

16

symbolic control, especially those implicated in the process of cultural reproduc-tion. (2001: 364)

This was Bernstein’s quest, and it provides insight into the nature of his evolvingtheory and the questions it raised. The principal concept Bernstein placed atthe centre of his explanation of social and cultural reproduction, transform-ation and change was, at least initially, that of code.

Conceptualizing codes

The concept of code has undergone several transmutations in Bernstein’s workalthough the underlying conceptual continuum has remained stable, tracing arange from ‘less specialised’ to ‘more specialised’. In his early work leading upto the restricted-elaborated code couple, Bernstein (1971) identified code dif-ferences in terms of more or less complex lexical, semantic and grammaticalfeatures. Soon, however, he came to consider code less as a linguistic repertoireand more abstractly in terms of an orientation to meaning. Bernstein argued thatdifferent positions within the social base, understood in terms of their degree ofspecialization, create, as he later put it, ‘different modalities of communicationdifferentially valued by the school, and differentially effective in it, because ofthe school’s values, modes of practice and relations with its different communi-ties’ (1996: 91).

For Bernstein a restricted coding orientation, typical of someone in a rela-tively unspecialized context and with a relatively direct relation to the socialbase, predisposes that person to context-specific meanings; an elaborated codeorientation, typical of someone in a more specialized context and with a moremediated relation to the social base, predisposes that person to universalistic,non-local, context-independent meanings. The principal attribute of an elabor-ated coding orientation is that it is able to realize more combinatorial possibili-ties than a restricted coding orientation. Crucially, in societies with a specializeddivision of labour, such as is found in modern, industrialized countries, toprosper requires the possession of an elaborated orientation to meaning. Theprincipal sites for transmitting this privileged and privileging elaborated orien-tation are the home and the school, but not all homes and not all schools to thesame degree. In this way, code becomes a key point of cleavage in class society.Though the concepts of elaborated code and (especially) restricted code havebeen the subject of considerable misunderstanding and criticism, Bernstein’searly work thereby laid out a basis for understanding how modern societiessustain themselves.

In terms of the concept of code, much of this is common ground betweensociologists using Bernstein’s approach and systemic functional linguists, whoalone in the broader socio-linguistic community saw from the beginning that‘code’ was not dialect and that code theory was neither a bourgeois alibi formiddle-class speech nor a denigrating deficit account of working-class language,understandings Bernstein later somewhat mischievously described as ‘forpeople who not only won’t read but can’t read’ (2001: 371). Though quick to


state that he never saw himself as working in sociolinguistics (1996: 147–56),Bernstein often highlighted this shared understanding, paying tribute to the‘remarkable exception’ of a collection of papers edited by Frances Christie(1999) and to the ‘incalculable’ contribution of Michael Halliday and RuqaiyaHasan both to the development of this phase of the code theory and histhinking more generally (Bernstein 2000: 146).

As his work progressed, Bernstein continued to refine the theory of codetowards greater levels of delicacy and generality. His interest shifted increasinglyto exploring the elaborated code and the institutionalized sites of its dissemi-nation, principally the school. Bernstein’s first step was to see code as a princi-ple operating at a high level across a wide variety of contexts. The concept ofcode was thereby expanded beyond the possession of an individual to a moregeneral principle or set of rules for the regulation and distribution of meaning.His initial codification of this idea lay in the algebra of the language of classifi-cation and framing, where classification conceptualizes relations of power thatregulate relations between contexts or categories, and framing conceptualizesrelations of control within these contexts or categories (1975). By distinguish-ing between power and control in this way, Bernstein opened up the possibilityof exploring the different modalities an elaborated orientation to meaningmight take (such as ‘visible’ and ‘invisible’ forms of pedagogy) and addressingwhy a particular modality was institutionalized for particular groups of pupils,and with what consequences for their educational experiences and outcomes.

In the sociological work of his students and other researchers, the super-categories of classification and framing have been operationalized, into delicateobservational instruments that have proved remarkably fecund in empiricalstudies of classrooms in many different national contexts, including Australia,Chile, Finland, Portugal, South Africa, the UK and the USA.3 This work, togetherwith that conducted earlier at the Sociological Research Unit in London underBernstein’s direction, more than gives the lie to a common misconception thatthe theory lacks empirical application. From this ongoing body of work we nowhave a remarkably nuanced view of the operation and effects of various peda-gogic modalities with children from a variety of social backgrounds. As thesestudies show, Bernstein’s concepts of classification and framing enable not onlythe thick description prized in much educational research but also thick expla-nation. They offer a basis for researchers to address, for example, why particularsocial groups of pupils may do less well in particular classrooms or schools.Empirical research has been able to show that the group in question may notreadily be able to recognize and/or realize the code required for achievementwithin that specific educational context. This in turn has direct implications foreducation policy. Simply put, one is effectively presented with a choice: changethe underlying structuring principles of the school, curriculum or classroom tomatch the code already possessed by these pupils, or develop ways of providingthose pupils with the key to the code enabling success within those contexts.Unfortunately, the former, often advanced by well-intentioned but misguidededucationalists, would effectively relegate subordinate social groups to lower-status forms of educational knowledge and hence to the lower rungs of thedivision of labour; the latter is open to misreading as a deficit theory and requires


18

acknowledging the differential status of different forms of knowledge, an admis-sion considered beyond the pale in much contemporary social science.

This coding stage of Bernstein’s theory concentrated on the transmission andacquisition of elaborated orientations to meaning within schools and, to a lesserextent, within the home, and on relations between the two. However, they canin principle be applied to any ordered symbolic ensemble. Bernstein (1996:188–91) gave the example of using code theory to describe the form taken bythe layout and style of lavatories, their different modalities of usage, and theirsocial base. Moreover, he argued that the pedagogic nature of social relationsextended beyond the classroom to include, for example, doctor–patient, socialworker–client and lawyer–client relations (1999b). Nonetheless, though codetheory provided concepts that were highly applicable across a range of contextsfrom the macro to the micro, Bernstein’s principal ambitions also required thetheory to be able to account for what it is that schools actually do in the broaderrealm of the circulation and advancement of culture in society. The questionwas how schools act as relays of society’s distribution of power and principles ofcontrol. Put another way, given the now much expanded understanding of themicrophysics of school-based code transmission and acquisition, what is theDurkheimian big picture of society? Bernstein had, for example, argued thatstruggles between advocates of visible and invisible modalities of pedagogicpractice represented an ideological conflict on the wider social stage betweendifferent fractions of the middle class: how might these different levels ofanalysis be brought together to explore further how the differentiation and reg-ulation of symbolic forms shapes social structure? Bernstein’s answer to suchquestions was to attempt to relate the realms of knowledge production, peda-gogic recontextualization and meaning acquisition. Conceptualizing codesthereby raised new questions that would lead Bernstein closer to a focus onknowledge: how does a society circulate its various forms of knowledge, and howis consciousness specialized in society’s image?


Field of practice Production Recontextualization Reproduction

Form of regulation distributive rules recontextualizing evaluative rulesrules

Kinds of symbolic knowledge structure curriculum pedagogy andstructure evaluation

Principal types hierarchical and collection and visible andhorizontal integrated invisible pedagogiesknowledge curricular codesstructures

Typical sites research papers, curriculum policy, classrooms andconferences, textbooks, examinationslaboratories learning aids

Table 2.1 The arena of the pedagogic device (adapted)

The pedagogic device

Bernstein’s answer was by means of the pedagogic device (see Table 2.1). Named(potentially misleadingly) after the ‘language device’, the pedagogic deviceforms the basis of his account of: the ordered regulation and distribution of asociety’s worthwhile store of knowledge, ordered by a specifiable set of distribu-tive rules; the transformation of this store into a pedagogic discourse, a formamenable to pedagogic transmission, ordered by a specifiable set of recontextu-alizing rules; and the further transformation of this pedagogic discourse into aset of evaluative criteria to be attained, ordered by a specifiable set of evaluativerules.4 In Bernstein’s conceptualization each of these rules is associated with aspecific field of activity: a field of production where ‘new’ knowledge is con-structed and positioned; a field of recontextualization where discourses fromthe field of production are selected, appropriated and repositioned to become‘educational’ knowledge; and a field of reproduction where pedagogic trans-mission and acquisition takes place. The specific activities of each field are,Bernstein suggests, primarily, though not exclusively, associated with specificsites. As the above implies, the three rules and fields constituting the device aresaid to be hierarchically related: production precedes recontextualization,which precedes reproduction.

Together these three rules and their associated fields constitute an ‘arena’created by the pedagogic device. Taking each in turn, the distributive rules dis-tinguish between which knowledges are deemed more or less worthwhile andwhich of these forms of knowledge should be distributed to whom. FollowingDurkheim, Bernstein saw the fundamental division of labour – into mental andmanual forms of labour – as related to a corresponding symbolic cleavagebetween sacred and profane symbolic orders, where the higher status, morerewarded, sacred symbolic orders are differentially distributed, in the past byvarious kinds of priestly castes and the Church and in more recent times by theeducation system. Distributive rules regulate relations between these symbolicorders and how they are to be distributed – who enjoys access to what forms ofknowledge, and in particular, who enjoys access to the means of producingnew knowledge. The recontextualizing rules construct pedagogic discourse,the ‘what’ and ‘how’ of schoolwork; they comprise principles for ‘delocating adiscourse, for relocating it, for refocusing it’ (1996: 47), transforming know-ledge into pedagogic communication. Evaluative rules complete the circle byestablishing the evaluation nodal points that are to be acquired, stipulatingthe specialized consciousness that should result. These three sets of rulescould thus be said to integrate analyses of power (distributive rules), know-ledge (recontextualizing rules), and consciousness (evaluative rules).

As Sadovnik argues, what is critical about the notion of the pedagogic deviceis that ‘Bernstein is concerned with more than the description of the productionand transmission of knowledge; he is concerned with its consequences for dif-ferent groups’ (1995a: 10). Moreover, as ever, Bernstein’s sociological concernlay with how the differential regulation and distribution of knowledge is relatedto the evolving structure of society. Thus the theory aimed not only to bringtogether power/knowledge/consciousness, but to place this within an account of


20

cultural and social reproduction, transformation and change. Bernstein arguedthat the three fields of production, recontextualization and reproductiontogether represent an ‘arena of struggle’ (1990: 206) in which groups attempt toappropriate or control the pedagogic device. To control the device is to haveaccess to a ‘symbolic ruler of consciousness’, a ‘ruler’ in both senses of havingpower over consciousness and measuring the legitimacy of its realizations:

Groups attempt to appropriate the device to impose their rule by the constructionof particular code modalities. Thus the device or apparatus becomes the focus ofchallenge, resistance and conflict. (Bernstein 1996: 193; emphasis in original)

The code modality announces what should count as a marker of success orachievement and the pedagogic device is the means whereby this principle ofhierarchization is created, reproduced, transformed and changed. Those inpositions of power are, Bernstein suggests, able to metaphorically ‘set’ thedevice such that the dominant, higher-status code modality favours their own.Conversely, actors whose dispositions and practices are characterized by a dif-ferent code may experience difficulty in recognizing and/or realizing practicesthat are rewarded within the specific context. The question Bernstein posits ascrucial for research thereby became: ‘Whose ruler, what consciousness?’ (1996:193); that is, who controls the pedagogic device and what kind of principle ofhierarchization (code modality) are they attempting to impose as the measure?In Christie and Macken-Horarick (this volume) we can see a concrete effort toconstruct a counter-‘ruler for consciousness’, by means of constructing a lin-guistic metalanguage for the school English curriculum in a field currentlydominated by invisible pedagogy.

The ‘pedagogic device’ is an ambitious attempt to capture the role of educa-tion in the sociological big picture, reaching from social structure to individualconsciousness. It represents a synoptic perspective on the orders of symboliclife; if classification and framing began from the micro-physics of the classroom,the device comprises an attempt at beginning a grand unified theory. It alsoworks at a higher level of abstraction than codes. The rules regulated by the ped-agogic device are resources for the construction, reproduction, transformationand change of codes rather than the codes themselves – one sees the effects ofthe device and not the device itself. At the same time, this stage of Bernstein’swork subsumed earlier understandings of codes at a higher level of abstraction,which he now defined as:

where O refers to ‘orientation to meaning’ (elaborated or restricted) and theline refers to the embedding of this orientation in relative strengths (strongerand weaker) of classification and framing (1990: 43). Bernstein’s conceptual-ization thus operates at a high level of both generality and abstraction. It also,for that reason, lends itself less easily to empirical research. Indeed, Dowling(1999) argues that it cannot be operationalized, though Bernstein (2000:


O

± Ci/e ± Fi/e

116–20) cites studies by Cox Donoso (1986), Swope (1992) and Singh (1993)which have deployed it as an overarching framework for empirical studies.

The significance of the pedagogic device for this brief overview, however,lies more in the path it opens up to the issue of knowledge, a hitherto ratherneglected focus in Bernstein’s otherwise ambitious sociology of education. For,in setting forth these ideas and making a case for the necessity of a theory of thepedagogic device, Bernstein highlighted the absence of an analysis of pedagogicdiscourse itself and raised questions of the forms taken by knowledge, issues hecame to realize had not been answered by his account of the device.

Making knowledge visible

It is interesting to observe that though knowledge is ostensibly the basis of edu-cation, and we are said to now live in ‘knowledge societies’, it is often left unre-marked in studies of education. A potential barrier to seeing knowledge is aversion of what the sociologist Dennis Wrong (1994) once called the ‘over-socialised’ image of people, which in education becomes what may be called an‘over-ideologised’ image of knowledge and pedagogy. In its crudest forms thisleads to treating knowledge as if it exists only to reproduce various forms ofsocial inequality (which form depending on the perspective being advanced)or, in some more extreme positions, in claims that all knowledge is ideologizedand can be understood wholly in terms of either domination or subordination.Here knowledge is reduced to the knower perspective, the ruling ideas of anage are the ideas of the ruling class (or gender or ethnicity and so forth), andnothing but (see Maton 2000; Moore and Muller 1999). To all of this there issome truth, but it is not the whole truth. Approaches operating with an over-ideologized image of knowledge provide much of value; they offer insightfulaccounts of the ways in which social relations of power pervade the conditionsand contexts of the production, recontextualization and reproduction ofknowledge, and stand as a corrective to any temptation to decouple power andknowledge. However, the barrier that an unnuanced image of knowledge andpedagogy creates is simply that, from these perspectives, we do not have to takethe internal ordering of symbolic forms seriously. Education becomes a reflec-tion or epiphenomenon of social structure, one without any intrinsic powers,properties or tendencies of its own. All questions of ‘what knowledge is atstake?’ give way to the question: ‘whose knowledge?’.

This was the problem Bernstein felt was presented by most reproductiontheory approaches to education, and which he intended his account of thepedagogic device to help rectify. He argued that such approaches analyse only‘relations to’ education, such as relations of class, race or gender to pedagogicdiscourse:

It is as if pedagogic discourse is itself no more than a relay for power relationsexternal to itself; a relay whose form has no consequences for what is relayed.(1990: 166)


22

What was additionally required, he argued, was an analysis of ‘relations within’pedagogic discourse, that is, an analysis of ‘the intrinsic features constitutingand distinguishing the specialized form of communication realized by thepedagogic discourse of education’ (1990: 165).

It was just such an analysis that Bernstein’s theory aimed to offer. However, toachieve such an aim required bringing knowledge more firmly to the centre ofthe theory’s focus, for several reasons. First, Bernstein came to believe that theconcept of code ‘took for granted, and left unexamined, the form of the dis-course’ (1999b: 273) and the pedagogic device ‘didn’t actually show the natureof pedagogic discourse. It showed how it was put together but it didn’t show itsnature’ (2001: 373). In short, the pedagogic device was the condition for theconstruction of pedagogic discourse; what was still required was to address theforms this discourse might take, necessitating a focus more on the forms ofknowledge than solely on the forms of pedagogic communication. Second,having analysed the transmission and acquisition of educational knowledge (inthe field of reproduction) and its construction (in the field of recontextualiz-ation), the question remained as to the forms taken by the symbolic dimensionof the field of production, the knowledge from which pedagogic discourse iscreated (see Table 2.1). Lastly, though arguing that we are now entering anew ‘Totally Pedagogised Society’ based on lifelong learning, for Bernstein acontinuing focus on ‘pedagogy’ was paradoxically insufficient:

I think now, looking forward, that a sociology of pedagogy does not indicate orsuggest the conceptual development necessary to grasp the discursive culture forwhich we are being prepared. The term pedagogy has restrictive references,despite my attempt to expand its use. (2001: 367)

To understand the contemporary situation required a focus on what kinds ofknowledges are being distributed to which social groups and to shape whatforms of consciousness. With this, Bernstein signalled a move in his thinkingfrom pedagogies to knowledges: ‘I have lately been attempting what could becalled a sociology for the transmission of knowledges’ (2001: 368).

In coming to address the lack in his approach of an understanding of theforms taken by discourse, Bernstein was keenly aware of Durkheim’s character-ization of the internal properties of sacred symbolic ensembles, and how theydiffer from everyday modes of thought. How, asked Bernstein, do these differ-ently patterned symbolic ensembles relate to the social base, and in what way dothey differ in their specialization of consciousness? Bernstein had earlierasserted that elaborated codes differ from restricted codes in their ability torealize more combinatorial possibilities – could this be further stipulated?Moreover, his key notion of recontextualization highlighted how school subjectsare not simply a reflection of their associated fields of knowledge: what formsmight the latter take? In his last major contribution, Bernstein attempted toaddress these questions.


Discourses and knowledge structures

Bernstein’s conceptualization of the knowledges subject to pedagogic transfor-mation (1999a) begins from the wider perspective of distinguishing two formsof discourse: horizontal discourse and vertical discourse. Horizontal discourse refersto everyday or ‘common-sense’ knowledge and ‘entails a set of strategies whichare local, segmentally organized, context-specific and dependent’ (1999a: 159).The knowledges comprising this discourse ‘are related not by integration oftheir meanings by some co-ordinating principle, but through the functionalrelations of segments or contexts to the everyday life’ (1999a: 160). In contrast,vertical discourse ‘takes the form of a coherent, explicit, and systematically prin-cipled structure’ (1999a: 159). Where the knowledges of horizontal discourseare integrated at the level of relations between segments or contexts, the know-ledges of vertical discourse are integrated at the level of meanings which arerelated hierarchically: ‘context specificity through “segmentation” in horizontaldiscourse, but context specificity through recontextualisation in vertical dis-course’ (1999a: 161).

It is common to conflate this distinction. For example, some cognitive psy-chologists will refer to knowledge as ‘everything that goes on in the head’ or as‘everything that is a script for action’, and some linguists might consider know-ledge to be ‘everything that is marshalled in a person’s linguistic repertoire’;this tempting conflationary gesture inheres in common sense itself. However, inmaking the distinction Bernstein is not suggesting the discursive practices ofactors do not move between the two or that they represent an impassable faultline in the empirical world. Rather, the distinction is important for understand-ing social structure. If no such distinction is made, then relations betweenspecialized symbolic forms and the specialized division of labour cannot beexplored. The distinction is thus crucial for addressing the key sociologicalquestion: how are differently valorized and rewarded forms of knowledgedifferently distributed in society?

Bernstein starts his account by recognizing that distinctions between forms ofdiscourse had periodically been made in social science. However, he argued thata further attempt to account for knowledge is required because of a tendencywithin such discussions to offer a simple dichotomy – such as concrete/abstractthought, local/official knowledge or everyday/school knowledge – thatobscures differences within vertical discourse (2000: 156). Accordingly, Bern-stein then turns his attention to the different forms taken by vertical discourse,making a distinction between ‘hierarchical knowledge structures’ and ‘horizon-tal knowledge structures’.

Bernstein defines a hierarchical knowledge structure as ‘a coherent, explicit andsystematically principled structure, hierarchically organised’ which ‘attempts tocreate very general propositions and theories, which integrate knowledge atlower levels, and in this way shows underlying uniformities across an expandingrange of apparently different phenomena’ (1999a: 161, 162). This form, exem-plified by the natural sciences, Bernstein visually represents as a triangle ofknowledge, one motivated towards building an apex of greater integratingpropositions. In contrast, a horizontal knowledge structure is defined as ‘a series of


24

specialized languages with specialized modes of interrogation and criteria forthe construction and circulation of texts’ (1999a: 162), such as the disciplinesof the humanities and social sciences. For example, in sociology the languagesrefer to its wide array of competing theoretical approaches, such as functional-ism, structuralism, Marxism, post-modernism, and so forth. A key differencebetween the two knowledge structures lies in the form taken by their develop-ment. According to Bernstein, a hierarchical knowledge structure develops bymoves to widen the base and sharpen the tip of the triangle: theories are soughtthat embrace more empirical phenomena and comprise fewer axioms thanexisting theories. Intellectual progress is thus defined as the integration andsubsumption of existing ideas within more overarching and generalizing propo-sitions; for example, as the Nobel prize-winning physicist Richard Feymann putit: ‘Physics has a history of synthesising many phenomena into a few theories’(1990: 4). In contrast, a horizontal knowledge structure develops through theaddition of new languages. We thus have ‘integration of language in one caseand accumulation of languages in the other’ (1999a: 163).

These knowledge structures have ramifications for the intellectual shape offields of production. Bernstein suggests, for example, that the segments ofhorizontal knowledge structures tend to be characterized by short-term obso-lescence, only to reappear again some time in the future in a new guise. Theyare more vulnerable to the changing winds of intellectual fashion, but thoughnames and styles may change, a similar account is likely to recur within each newlanguage that emerges. Here, from the perspective of comparison with hier-archical knowledge structures, differences between these segments becomeakin to ‘fingerprints on identity cards which are otherwise exactly the same’(Adorno and Horkheimer 1947: 154). The capacity to create knowledge thatbuilds on and goes beyond existing knowledge is limited. These differentknowledge structures, as a consequence, specialize consciousness differently.They also shape social practices and forms of pedagogy, and differently special-ize consciousness within their intellectual fields. For example, Bernstein arguesthat in hierarchical knowledge structures acquirers do not have the problem ofknowing whether they are speaking or writing physics: ‘the passage from onetheory to another does not signal a break in the language; it is an extension ofits explanatory/descriptive powers’ (1999a: 164). In horizontal knowledgestructures acquirers are faced with an array of languages based on different,often opposed assumptions, making it less clear that one is indeed speaking orwriting sociology. Given all this, the question then becomes: who has access towhat form of knowledge?

In conceptualizing these knowledge structures Bernstein is concerned withexploring the properties of what Karl Popper (1972) terms ‘World 3’, theproducts of our human minds or ‘objective knowledge’, rather than ‘World 2’,our mental states or subjective knowledge. For, as Popper put it, ‘no theory ofsubjective knowledge will be able to account for objective knowledge’ (1994a:13). In contrast to the ‘over-ideologised’ image characterizing many approachesto education, Bernstein’s theory is thus an account of knowledge rather than ofknowers. This is not to say that Bernstein’s approach diminishes human agencyin favour of hypostasizing knowledge as an autonomous and freely floating


entity separate from the social practices of actors. Rather, his conceptualizationaims to make visible knowledge as an object, one with its own properties andpowers that are emergent from, but irreducible to, social practices and which,indeed, help shape those practices. To explore these properties and powers,Bernstein could be said to abstract knowledge from practices and socialcontexts, ‘rather like a figure relieved of its ground’ (1975: 2), as he describedhow his earlier work abstracted transmission and acquisition in the school fromits wider constraints and contexts in order to analyse those processes. In doingso, however, he brings to light the ways in which the structuring of knowledgeitself works to shape social practices, identity, relation and consciousness.

Exploring questions raised by ‘knowledge structures’

As we have outlined, each stage of the theory’s development and empiricalapplication raised questions which helped drive the theory forward. Bernstein’sconceptualization of knowledge structures is no exception; it raises questions,many addressed by the contributors to this volume. Here we shall explore twoissues as a context for these engagements: the nature of differences betweenknowledge structures; and relations between knowledge structures and the cur-riculum.

Differences between knowledge structuresThe first set of questions concern whether the distinctions in Bernstein’s con-ceptualization are too clear-cut: is there not a continuum between differentknowledge structures? Put another way, we might ask: do not intellectual disci-plines exhibit characteristics of both the forms Bernstein delineates? It can seemas if Bernstein is suggesting that the development of horizontal knowledgestructures is characterized by permanent cultural revolutions that leave no traceof the past. This would make discovering even the smallest degree of continuityof problems, themes or terms within a discipline appear to be a sign of a hier-archical knowledge structure. However, such a discovery is only to be expected:for an intellectual field to exist, it must have a degree of continuity across timeand space. Moreover, the addition of a new language within a horizontal know-ledge structure does not necessarily eliminate, and is likely to incorporate, atleast some of the terms of existing languages; when creating a new segmentaltheory authors are likely to start from established ideas. The issue for Bernsteinis not whether over time actors within an intellectual field gnaw away at similarproblems, in similar ways, using similar terms or referencing past authors.Rather, the question is whether new theories that emerge subsume and inte-grate past theories and aim for greater abstraction and generalizability, or areconsidered incommensurable with existing theories. It is a question not ofwhether authors use existing symbolic materials but of how they do so and withwhat epistemic outcomes. One would therefore expect to find the same themes,motifs, terms, styles and so forth recurring, being adapted, modified, recast andreworked over time within a horizontal knowledge structure, not only withina specific segment but also across segments. This in itself would not, for


26

Bernstein, signal a hierarchical knowledge structure. Though theories in a hor-izontal knowledge structure may overlap in their use of common terms, they‘make different and often opposing assumptions, with each language having itsown criteria for legitimate texts, what counts as evidence and what counts aslegitimate questions or a legitimate problematic’ (Bernstein 1999a: 163). So,though the segments of a horizontal knowledge structure may include the sameterms, such as ‘patriarchy’ or ‘social class’, their authors are not speaking thesame language – their assumptions and criteria for legitimate knowledge claimsare different. Similarly, one would expect to discover a degree of integration andsubsumption of past ideas within each language. However, the capacity for suchdevelopment across languages is limited. Even if one were to describe a theorythat develops in this way as a mini-triangle (see Wignell, this volume), in a hori-zontal knowledge structure each such mini-triangle does not subsume andintegrate its predecessors and compatriots within the knowledge structure toform part of a bigger triangle. The serial character of development of the knowl-edge structure as a whole thereby remains accumulation rather than integration.

If the two forms of knowledge structure are different, wherein then lie thebases of this difference? Put another way, is it possible for the social sciences toprogress in the same way as the natural sciences and, if so, what would makesuch development possible? Muller (this volume) highlights two attributes ofBernstein’s conceptualization of knowledge structures as key foci for addressingthese questions: ‘verticality’ and ‘grammaticality’. The first dimension refers tothe degree to which the development of a knowledge structure is characterizedby the integration and subsumption of knowledge into more overarching andgeneralizing propositions. What enables knowledge structures to develop in thismanner is a crucial area for research, one in which systemic functional linguis-tics is providing valuable insights; see, for example, Martin (this volume) on therole played by linguistic technicality, especially grammatical metaphor, inenabling hierarchical integration of knowledge, and O’Halloran (this volume)on how mathematics enables progress in the natural sciences to develop in thisway.

The second dimension highlights the role played by what Bernstein terms thestrength of ‘grammar’ or degree to which forms of knowledge exhibit ‘anexplicit conceptual syntax capable of “relatively” precise empirical descriptionsand/or of generating formal modelling of empirical relations’ (1999a: 164).Bernstein seems to suggest that, in any knowledge structure at any one time,more than one triangle or theory is likely to co-exist, suggesting a prima facie sim-ilarity to the segmented nature of horizontal knowledge structures. Both formsof knowledge structure are characterized by conflicts between advocates of alter-native theories. However, in hierarchical knowledge structures choices betweentheories are, Bernstein argues, at least possible on the basis of recourse toempirical research because these theories have stronger grammars. The emer-gence of a new theory within a hierarchical knowledge structure is thereby bothconflictual and integrative: to represent progress it must clash with its pre-decessor but also be able to explain that predecessor’s success. In other words,the new integrating theory includes but goes beyond its predecessor, or asPopper puts it:


In all those cases in which its predecessor was successful it must yield results atleast as good as those of its predecessor and, if possible, better results. Thus inthese cases the predecessor theory must appear as a good approximation to thenew theory, while there should be, preferably, other cases where the new theoryyields different and better results than the old theory. (1994b: 12)

In contrast, the relatively weaker grammars of horizontal knowledge structuresmean that relations between languages or segments cannot be settled byempirical research and are confined to critique. A key difference betweenknowledge structures is thus not one of stability–conflict, consensus–dissensusor orthodoxy–heterodoxy, but rather concerns the form taken by these con-flicts and their outcomes, in which the strength of grammar plays a role. Thisissue of grammaticality may prove enlightening for understanding how somesocial science disciplines are more capable of sustained intellectual progressthan others. Bernstein cites, for example, economics and linguistics asexamples of horizontal knowledge structures with relatively strongergrammars, and sociology and cultural studies as representing relatively weakergrammars. Within a discipline with a stronger grammar, where languagespurport to share the same empirical referents, there may be the possibility ofsomething akin to the relations between theories that characterize hierarchi-cal knowledge structures.

These two interrelated dimensions are key to the differences Bernstein high-lights between hierarchical and horizontal knowledge structures. Both play arole in determining the form taken by their development. In choosing betweenalternative theories, actors in hierarchical knowledge structures both criticallyexamine their consistency and compatibility with others theories (their internalstrengths of grammar) and their capacity to explain the results of empiricalresearch (their external strengths of grammar). Together these provide the pos-sibility of a rational basis for progress in hierarchical knowledge structures, tothe extent that it can be decided whether a new theory represents an advanceon existing theories by recourse to its integrative, subsumptive and explanatorypower and to its fit with the facts.5 In contrast, significant changes in horizontalknowledge structures are all too often ideological rather than rational revolu-tions. Here alternative theories are in a war of hearts and minds, and choicesbetween competing claims to insight are based more on a ‘knower code’, that isto say, on who is making knowledge claims rather than on what is being claimedand how.

Though Bernstein returned to the issue a number of times (there are threeversions of his paper in print [1996, 1999a, 2000] and at least one other longerversion that was never published), this question of the basis of differencesbetween knowledge structures could be clearer. There are two possible reasonsfor this. With Bernstein’s conceptualization we are, as Muller (this volume)highlights, ‘locked into an early (lexical) metaphorical stage of discussion,where the terms are more suggestive than they are explanatory’. Secondly,knowledge structures are not the only feature of intellectual fields of produc-tion; they represent the symbolic dimension of what are social fields of practice.To understand the development of any specific discipline one must also offer a


28

sociological account of ‘relations to’ knowledge, such as the roles played by thestate, economy, social structure, and struggles between actors within the field.6

A discipline is more than just its structuring of knowledge; the concepts of‘knowledge structures’, therefore, shed light on disciplinary development butare not the whole story.

Knowledge structures and curriculum structuresA second set of questions raised by Bernstein’s later ideas concerns relationsbetween knowledge structures and educational knowledge. With the notion of‘knowledge structures’ Bernstein is exploring the symbolic products of the fieldof production (see Table 2.1); the notion of recontextualization highlights thata knowledge structure is not necessarily a curriculum structure or pedagogicstructure, and his theorization is not simply a recasting of pedagogic codes.Thus, in terms of Bernstein’s concepts as they currently stand, one would notdescribe a school curriculum in terms of exhibiting a horizontal or hierarchicalknowledge structure.7 However, the caveat ‘as they currently stand’ highlights,of course, that Bernstein’s framework foresees its own reformulation and sowhether such concepts will be extended and developed in future is open. Specif-ically, this raises questions of how the issues highlighted by these concepts (suchas the degree to which knowledge structures develop through integrating andsubsuming past ideas) can be explored in the school context. For example,Christie and Macken-Horarik (this volume) highlight how pupils experience atrajectory of schooling in subject English in ways that often debilitate the inte-gration of already learned knowledge. Rather than explicitly building on pastlearning, the invisible pedagogy of the English curriculum renders the educa-tional knowledge structure less visible with the effect that pupils often experi-ence the equivalent of segmental rather than integrative acquisition ofeducational knowledge.

If knowledge structures are not curriculum structures, this also raises thequestion of the degree to which the latter reflect the former. What are therelations between the knowledge structures of physics, mathematics or Englishliterature and their respective curriculum structures? Bernstein argued thatwherever there is recontextualization there is a space for the play of ideology.Even so, are there limits to recontextualization and thus limits on the degree towhich ideology can shape the construction of pedagogic discourse? Bernstein,following Durkheim, saw that specialized divisions of labour demand specializedforms of consciousness. Since the specialized knowledges in the realm of pro-duction rest directly on the material base, there must surely be a limit to theamount of recontextualizing they can bear before defeating their purpose. Thisis made clear by the focus in Bernstein’s account of the pedagogic device on‘evaluative rules’; these may be pedagogized artefacts, but if the criteria theyconstruct bear no relation to their parent knowledges in the realm of produc-tion, then schooling will undermine its role as a relay of specialized knowledges.Relations between knowledge structures and their corresponding curriculumstructures is, in short, a key area for future exploration. It is to this elaborationthat existing work in systemic functional linguistics may have the most to con-


tribute (see, for example, Christie and Martin 1997, Halliday and Martin 1993and Unsworth 2000).

Conclusion

A key advantage of Bernstein’s approach, we have argued, lies in its capacity torender knowledge visible as an object of study. Many approaches to educationobscure the very thing that specializes education as a social field of practices. Interms of accounting for the forms taken by discourse and their relations tosocial structure, knowledge of knowledge has been in relatively short supply.With the concepts of codes, the pedagogic device and knowledge structures,Bernstein provides the basis for furthering that knowledge. In this chapter weoutlined one way of understanding how Bernstein’s theory arrived at his con-ceptualization of knowledge, highlighting the questions raised by the theory’sdevelopment. We began by outlining how the issue of how differently valorizedand rewarded forms of knowledge are differently distributed in society formeda key starting point of Bernstein’s problematic. We then traced how conceptu-alizing codes explored a key basis of social and cultural reproduction andchange, raising inter alia the question of what gives rise to these codes; and howthe pedagogic device modelled the construction of pedagogic discourse, raisingthe question of the nature of the knowledge subject to pedagogic transform-ation. Lastly, we outlined how Bernstein’s conceptualization of discourses andknowledge structures offers insights into the forms taken by knowledge, andsketched some of the questions these ideas have raised, issues opened upand addressed by many of the papers in this volume.

At the outset of this chapter we asked the question: ‘Why knowledge?’. Whydid Bernstein come to set out the basis of ‘a sociology for the transmission ofknowledges’? One answer lies in the broadening of Bernstein’s focus from howeducational knowledge is transmitted and acquired to how that educationalknowledge is constructed, and thence to the forms of knowledge from whicheducational knowledge is recontextualized. Bernstein was thereby tracingknowledge from the school upstream towards its epistemic sources, exploring inturn the fields of reproduction, recontextualization and production (Table 2.1).However, an issue we repeatedly highlighted is the sociological nature of thisquest. A second answer to the question ‘Why knowledge?’, therefore, lies inBernstein’s belief that the differential valorization, regulation and distributionof forms of knowledge in society is a crucial aspect of how societies maintainthemselves. Bernstein remained focused throughout his career on the socialreproduction function of cultural production, transformation and reproduc-tion. His unfolding account of knowledge sought to understand the role thatsymbolic forms play in the ordering of social life.

Finally, a third answer to ‘Why knowledge?’ may lie with the nature of thediscipline of sociology itself. Bernstein was repeatedly drawn to studying thesacred – whether understood as an elaborated orientation to meaning, verticaldiscourse or knowledge structures. Such forms of discourse have the singular


30

charm of being an accumulated sedimentation of symbolic extensions that hasbeen constructed by innumerable co-operating knowers, extended in time aswell as space, most of whom will remain unknown to one another. As Moore andMaton (2001: 172) put it when describing the community of mathematiciansinvolved over many centuries and across the globe in unravelling Fermat’s LastTheorem:

It represents an epistemic community with an extended existence in time andspace, a community where the past is present, one in which the living membersinteract with the dead to produce contributions which, when they die, will be inturn the living concern of future members.

Such an endeavour is a pure form of communism, as Robert Merton (1973)pointed out, but also ‘a very individualistic kind of communitarianism’ asArthur Stinchcombe (nd: 20) drily added. At its best, it comprises not only aknowledge structure where past insights are subsumed and integrated, standingon the shoulders of giants to see further (Muller 2006), but also a knowerstructure in which everyone in the scholarly community is potentially able tocontribute (Maton, this volume). It is likely that Bernstein yearned for that formof community, one which he rarely found in the lower reaches of his ownhorizontal knowledge structure, sociology, and that he has done more thananyone else in the sociology of education to change. As we have outlined, Bern-stein’s theory developed through the subsumption of past ideas within new,often more abstract, formulations as it sought to grasp an ever-widening rangeof phenomena. Bernstein’s own way of theorizing thereby attempted to helpanalyse and exemplify principles that could enable the building of knowledgereaching upwards, bringing together individuals within an epistemic commu-nity extended across time and space. The capacity of his ideas to do just that isillustrated by their continuing contribution to the ongoing, fruitful dialoguebetween sociologists and systemic functional linguists.

References

Adorno, T.W. and Horkheimer, M. (1947), Dialectic of Enlightenment. London: Verso. Atkinson, P. (1985), Language, Structure and Reproduction: An introduction to the sociology of

Basil Bernstein. London: Methuen.Atkinson, P., Delamont, S. and Davies, B. (eds) (1995), Discourse and Reproduction: Essays

in honor of Basil Bernstein. Cresskill, NJ: Hampton Press.Bernstein, B. (1971), Class, Codes and Control, Volume I: Theoretical studies towards a sociology

of language. London: Routledge & Kegan Paul.Bernstein, B. (1975), Class, Codes and Control, Volume III: Towards a theory of educational

transmissions. London: Routledge & Kegan Paul.Bernstein, B. (1990), Class, Codes and Control, Volume IV: The structuring of pedagogic dis-

course. London: Routledge.Bernstein, B. (1995), ‘A response’, in Savodnik, A. (ed.), Knowledge and Pedagogy: The

sociology of Basil Bernstein. Norwood, NJ: Ablex, pp. 385–424.


Bernstein, B. (1996), Pedagogy, Symbolic Control and Identity: Theory, research, critique.London: Taylor & Francis.

Bernstein, B. (1999a), ‘Vertical and horizontal discourse: An essay’, in British Journal ofSociology of Education 20(2): pp. 157–73.

Bernstein, B. (1999b), ‘Pedagogy, identity and the construction of a theory of symboliccontrol: Basil Bernstein questioned by Joseph Solomon’, in British Journal of Sociologyof Education, 20(2): pp. 265–79.

Bernstein, B. (2000), Class, Codes and Control, Volume V: Pedagogy, symbolic control andidentity. Oxford: Rowman & Littlefield.

Bernstein, B. (2001), ‘From pedagogies to knowledges’, in Morais, A. et al. (eds), Towardsa Sociology of Pedagogy: The contribution of Basil Bernstein to research. New York: PeterLang, pp. 363–84.

Christie, F. (ed.) (1999), Pedagogy and the Shaping of Consciousness: Linguistic and socialprocesses. London: Continuum.

Christie, F. and Martin, J.R. (eds) (1997), Genre and Institutions: Social processes in the work-place and school. London: Cassell.

Cox Donoso, C. (1986), ‘Continuity, conflict and change in state education in Chile:A study of the pedagogic projects of the Christian Democrat and Popular UnityGovernments’, in CORE 10(2).

Dowling, P. (1999), ‘Basil Bernstein in frame: “Oh dear, is this a structuralist analysis?”’,http://www.ioe.ac.uk/ccs/dowling/kings1999.

Feynmann, R. (1990), Q.E.D.: The Strange Theory of Light and Matter. London: Penguin.Halliday, M.A.K. and Martin, J.R. (1993), Writing Science: Literacy and discursive power.

London: Falmer.Maton, K. (2000), ‘Recovering pedagogic discourse: A Bernsteinian approach to the

sociology of educational knowledge’, in Linguistics and Education 11 (1), pp. 79–98.Maton, K. (2006), ‘On knowledge structures and knower structures’, in Moore, R., Arnot,

M., Beck, J. and Daniels, H. (eds), Knowledge, Power and Educational Reform: Applyingthe sociology of Basil Bernstein. London: RoutledgeFalmer.

Merton, R.K. (1973), The Sociology of Science: Theoretical and Empirical Investigations.Chicago: The University of Chicago Press.

Moore, R., Arnot, M., Beck, J. and Daniels, H. (eds) (2006), Knowledge, Power and Educa-tional Reform: Applying the sociology of Basil Bernstein. London: RoutledgeFalmer.

Moore, R. and Maton, K. (2001), ‘Founding the sociology of knowledge: Basil Bernstein,intellectual fields and the epistemic device’, in Morais, A., Neves, I., Davies, B. andDaniels, H. (eds) Towards a Sociology of Pedagogy: The contribution of Basil Bernstein toresearch. New York: Peter Lang.

Moore, R. and Muller, J. (1999), ‘The discourse of “voice” and the problem of knowledgeand identity in the sociology of education’, in British Journal of Sociology of Education20(2): pp. 189–206.

Morais, A., Neves, I., Davies, B. and Daniels, H. (eds)(2001), Towards a Sociology ofPedagogy: The contribution of Basil Bernstein to research. New York: Peter Lang.

Muller, J., Davies, B. and Morais, A. (eds), (2004), Reading Bernstein, Researching Bernstein.London: RoutledgeFalmer.

Muller, J. (2006), ‘On the shoulders of giants: verticality of knowledge and the school cur-riculum’, in Moore, R. et al. (eds) Knowledge, Power and Educational Reform: Applyingthe Sociology of Basil Bernstein. London: RoutledgeFalmer.

Popper, K. (1972), Objective Knowledge: An evolutionary approach. Oxford: Clarendon Press.Popper, K. (1994a), Knowledge and the Body-Mind Problem: In defence of interaction. London:

Routledge.Popper, K. (1994b), The Myth of the Framework. London: Routledge. Sadovnik, A. (1995a), ‘Basil Bernstein’s theory of pedagogic practice: A structuralist

approach’, in Sadovnik (ed.) Knowledge and Pedagogy. Norwood: Ablex.


http://www.ioe.ac.uk/ccs/dowling/kings1999

32

Sadovnik, A. (ed.) (1995b), Knowledge and Pedagogy: The sociology of Basil Bernstein.Norwood: Ablex.

Singh, P. (1993), ‘Institutional discourse: A case study of the social construction of tech-nical competence in the primary classroom’, in British Journal of Sociology of Education14(1): pp. 39–58.

Stinchcombe, A. (nd), ‘Universities and the communism of knowledge as the core of theworld information economy’, Northwestern University, mimeo.

Swope, J. (1992), ‘The production, recontextualising and popular transmission of reli-gious discourse in eight basic Christian communities in Santiago, Chile’, in CORE16(3).

Unsworth, L. (ed.) (2000), Researching Language in Schools and Communities: Functionallinguistic approaches. London: Cassell.

Wrong, D. (1994), The Problem of Order: What unites and divides society. London: The FreePress.

Notes

1 For informative introductions to Bernstein’s ideas, see Atkinson (1985), papers col-lected in Sadovnik (1995b) and in British Journal of Sociology of Education 23(4), 2002,and the website of the Estudos Sociológicos da Sala de Aula centre in Lisbon(http://essa.fc.ul.pt). Given the state of the intellectual field of sociology (whichBernstein would describe as a horizontal knowledge structure with a weak grammar),offering an introductory way into, rather than a critique or development of Bern-stein’s later ideas for a cross-disciplinary audience is not without its dangers: Bern-stein himself may have described the following as ‘schizzing’ or ‘creative replacing’(1990: 8–9); his critics, sympathetic or otherwise, may see only exegetic advocacy, thework of what Dowling (1999) describes as ‘disciples’.

2 In highlighting the undoubtedly Durkheimian nature of Bernstein’s sociologicalenterprise we are, of course, not intending to obscure his wide-ranging theoreticalinterests and integration of insights from, among others, neo-Marxist, Weberian andpost-structuralist approaches. Such intellectual ancestry is beyond the limited scopeof this chapter.

3 See, for example, studies discussed in Bernstein (2000) and those collected inAtkinson et al. (1995), Christie (1999), Moore et al. (2006), Morais et al. (2001),Muller et al. (2004) and Sadovnik (1995b).

4 Bernstein’s use of the term ‘rules’ has led some commentators to suggest his theoryargues that practices are deterministically rule-governed. However, for Bernstein,rules do not by themselves cause anything but rather direct our attention to thecontrols on the form taken by pedagogic discourse, i.e. to the principles which giverise to its structuring.

5 This is not to say that this possibility is always taken up in practice. The history ofscience is replete with examples of overlooked or dismissed theories that, in retro-spect, represented advances on existing theories. Whether this possibility is recog-nized and realized in any particular case is an empirical question in which the natureof an intellectual discipline as a social field of practice plays a key role.

6 Conversely, one must beware the temptation to sociological reductionism whereby, forexample, the astonishing expansion of science is explained solely in terms of theinterests it serves. Bernstein’s approach suggests that such social power is an insuffi-cient explanation; one must also take into account the epistemic power characteriz-ing forms of knowledge.

7 This point may appear nitpicking were it not for what could be termed pedagogicreductionism, the tendency in discussions of education for the school classroom to


http://essa.fc.ul.pt

embody the gravity well of a black hole into which all other foci are drawn and dis-tinctions crushed. Though we do not, of course, wish to obscure the significance ofthe chalkface, this reductionism can obscure the notion of recontextualization andissues specific to the production of knowledge. Similarly, analyses of textbooks orcurriculum guidelines are studies of recontextualized pedagogic discourse ratherthan of knowledge structures.


1 Knowledge in school

My interest in knowledge structure arose in the context of literacy development,around the evolving theory and practice of the ‘Sydney School’.1 As this workmoved from primary into secondary school, and reconsidered secondary schoolliteracy in relation to workplace and academic discourse (Christie and Martin1997), it became increasingly important to focus on the knowledge encoded invarious genres – from one discipline or profession to another. This was alsocritical by way of getting teachers focused on literacy across the curriculum,since by secondary school reading and writing is seen as a tool for learning.2 Ourinitial strategy was to focus on one key humanities discipline (history) and onekey science (physical geography), and push on from there (Eggins et al. 1993;Wignell et al. 1990; Martin 2002a, b; for related work see Coffin 2000; Halliday2004; Halliday and Martin 1993; Macken-Horarik 1998; Martin 1993b; Martinand Veel 1998; Martin and Wodak 2004; Schleppegrell 2004; Unsworth 1997a,b, 1998, 2000, 2001; van Leeuwen and Humphrey 1996; Veel 1992).

Our informing theory for this work was systemic functional linguistics (here-after SFL), drawing in particular on functional grammar (Halliday andMatthiessen, 2004), discourse semantics (Martin and Rose 2003; Martin andWhite 2005), register and genre theory (Ghadessy 1993, 1995; Martin and Rose2005) and multimodality (Kress and van Leeuwen 1996, 2001). Accordingly ourmodel of language mapped metafunctions (ideational, interpersonal andtextual) across strata (graphology, lexicogrammar and discourse semantics) asin Figure 3.1 below; following Martin 1992, we deployed a stratified model ofcontext, including register (field, tenor and mode) and genre, as in Figure 3.2below. In a model of this kind, the register variable field provides a socialsemiotic perspective on knowledge structure; and knowledge is by and largerealized through, construed by, and over time reconstrued through ideationalmeaning (via the modalities of language and image). So the rest of this papercan be read as an exploration of knowledge structure as field of discourse.

2 Field

Following Martin, 1992, we treated field as a set of activity sequences orientedto some global institutional purpose, including the taxonomies of participantsinvolved in these sequences (organized by both classification and composi-tion).3 Taking science for example (geology in particular), we can recognize the

3 Construing knowledge: a functional linguisticperspective

J.R. Martin

35CONSTRUING KNOWLEDGE • MARTIN

Figure 3.1 Strata and metafunctions in SFL

textual

ideational

interpersonal

discourse semantics

lexicogrammar

phono-/graphology

Figure 3.2 Social context as register (field, tenor, mode) and genre

textual

mode

genre

tenor

field

ideational

interpersonal

36

specialized activity sequence spelled out in the following text – explaining howvolcanoes are formed:

(1) In parts of the earth beneath the crust . . . heat accumulates to such an extentthat it does cause local melting of rocks to form a molten mass called magma. Thismolten material is under such enormous pressure that some of it is forced into anycracks and crevices that might form in the upper solid crust of the earth, and insurrounding solid rock. Some of this molten material can actually cool and solidifywithout reaching the earth’s surface; in other cases molten material is pushed rightthrough the earth’s surface and forms a volcano. When molten material is forcedout to the earth’s surface it is called lava. (Messel 1963: 12.1, 40.4–8)

Note that the activity sequence4 gives rise to technical terms naming participants(in bold face above – magma, volcano and lava). From the perspective oftaxonomy, such participants are additionally organized with respect to bothclassification (hyponymy) and composition (meronymy). The classificationtaxonomy complementing Text 1, which shows where igneous rocks come from,has to do with types of rock. An image for these relations is presented as Figure3.3 below, with general classes to the left and subclasses to the right – the direc-tion in which subclassification is portrayed in SFL (as opposed to biology, whichprefers general classes on top and subclasses underneath).

Turning to composition, the main relations relevant here have to do with thestructure of volcanoes and the earth’s crust, as outlined in Figure 3.4 (fromMessel 1963: 40.4). Whereas scientists (biologists excepted) seldom use imagesto model classification relations, schematic drawings with appropriate labellingare regularly used for compositional relation across sub-disciplines.

In Bernstein’s terms, we’ve illustrated field here using examples from a hier-archical knowledge structure (vertical discourse). But the notions of activitysequence, classification and composition can be used to map all discourses(whether horizontal or vertical); what varies of course is the nature of the par-ticipants and activities at risk, and the level of technicality, which we’ll explorein sections 3 and 4 below.


Figure 3.3 Classification taxonomy (types of rock)

rocks sedimentary

metamorphic

igneous

extrusions

intrusions

clastic

non-clastic


Figure 3.4 Compositional relations (volcanoes) (Messel 1963: 40.4)

Figure 3.5 Common and uncommon sense fields (Martin 1992: 544)

oral transmission(doing)

written transmission(studying)

domestic(guidance)

specialized(participation)

sport

Commonsense

Uncommonsense

hobby

recreational(‘coaching’)

trades(apprenticing)

humanities

social science

science

administration(co-operation)

exploration(instruction)

38

Martin 1992 (pp. 544–5), inspired by Bernstein’s earlier notions of commonand uncommon sense discourse, developed a crude mapping of fields accord-ing to the way he imagined they were learned5 and their degree of lexicalspecialization. His provisional outline of fields is reproduced as Figure 3.5below, by way of illustrating the affinity of Sydney School research with Bern-stein’s developing concern with knowledge structure. Martin’s common sense touncommon sense cline parallels Bernstein’s horizontal discourse to hierarchicalknowledge structure dimensionality (although the position of bureaucraticdiscourse in Bernstein’s model is unclear).

2 Science (hierarchical knowledge structure)

Since we’ve begun with science, let’s focus more closely on its discourse here.Compared with other fields, scientific taxonomies are relatively comprehensive,deep and precise. Taking biological classification as a case in point, the tax-onomies are comprehensive in that they set out to classify all of the phenomenain their gaze (all of life), deep in that they classify and subclassify at many levelsof generality, and precise in that they ultimately depend on categorical geneticcriteria. Classifying animals, for example, involves covering everything that ismobile, obtains food from other living things, has extensive movement of bodyparts, and so on; and it takes several steps to move from animals in general tohuman beings like us:

Kingdom (animalia – obtain food, mobile . . .) Phylum (chordata – with hollow dorsal nervous system)Sub-phylum (vertebrata – with a backbone)Class (mammalia – with hair and mammary glands)Order (primates – with grasping hands)Family (hominidae – man-like reasoning)Genus (homo – . . .)Species (sapiens – . . .)

At every step along the way, subclassification is categorical and exhaustive – forexample, there are three different types of mammals, no less and no more, andevery mammal belong to one group or another:

– the egg-laying mammals called monotremes– the pouched mammals called marsupials– the higher mammals called placentals

An overview is presented in Figure 3.6, based on genetic reasoning (as recon-textualized for secondary school science by Haire et al. 2005).

Scientific taxonomies contrast with everyday taxonomies which are relativelypiecemeal, shallow and fuzzy. For one purpose or another we divide livingthings into groups such as food, pets, pests, garden plants and predators, includ-ing a relatively open-ended list of plants and animals in each group, and some-


times wondering whether something properly belongs to one group or another,or more than one (Is a tomato a fruit or a vegetable? Is a peanut a nut? Is a horsea food?). The classification sets are segmentally rather than hierarchicallyorganized (i.e. lacking in higher level superordinates), and many living thingsbelong to more than one group:


Figure 3.6 Genealogical organization of living things (Haire et al. 2005: 202)

40

– food (meat, fish, fruit, vegetables . . .)– pets (cats, dogs, birds . . .)– gardens (trees, shrubs, flowers, grass, vegetables . . .)– pests (insects, vermin . . .)– predators (snake, spider, crocodile, shark, bear . . .). . .

Another important feature of these everyday categories is that they can belearned ostensively, by having things we can see, touch, taste, hear and/or smellpointed out to us as examples (Painter 1998). We don’t learn them like techni-cal terms, through definitions; we just learn them concretely, based on sensuousexperience.

The possibility of ostensive definition becomes an even more importantcriteria when we turn from classification to composition, and rely on uncommonsense images of various kinds to mimic ‘concrete’ learning. In Figure 3.7, forexample, the composition of the human ear and Australia’s famous cochlear earimplant technology is outlined, and specialized terminology is presented aslabels for the parts. But the image involved is not a photo of the human bodyand implant; rather it is a schematic idealization based on centuries of techno-logically assisted perception, by anatomists, using probing and recordinginstruments (including of course writing and images, to record results). Even ifwe had the stomach for it, most of us would be very hard pressed to recognizeparts involved if a human body were made available to us. One needs thescientific understandings as part of finding one’s way around.


Figure 3.7 Cochlear ear implant (Haire et al. 2005: 236)

At this point it may be useful to distinguish between technical and specializedlexis along the following lines. Technical lexis can be defined as lexis which hasto be learned by definition, through language (typically in institutionalizedlearning); specialized lexis on the other hand is lexis that can be learned byobservation, through gesture (it is ostensively ‘defined’). Specialized languageis a feature of technology (White, 1998), and learned as part of apprenticeshipin arts and crafts, trades, sports, hobbies and the like. As such it is not availableto the community as a whole, but to people operating hands-on in particularspheres of activity (see Rose, 1997, 1998 on technological discourse in science-based industry, for example). Like everyday lexis, specialized lexis is concrete;but like technical lexis it belongs to specific walks of life (and similarly enactsthe boundaries between common and less common sense discourses).

Turning to activity sequences, scientific sequences focus more on cause thantime, and lean towards a relatively deterministic view of causality. For this reasonscientific activity sequences are referred to as implication sequences (in Martin,1992, 1993c), and are designed to explain. In the following explanation of coldfronts (Messel 1963: 7.6–7.7), for example, a number of causal resources aredeployed, including connections between clauses (if, how), verbs (causes, result-ing), a noun (reasons) and preposition (as a result of).

(2) Cold fronts. A stream of comparatively cold, dense air tends to move alongclose to the ground as it flows towards regions in which warmer, less dense, air isrising. This rising air becomes cooler for the reasons mentioned earlier, and if it ishumid condensation of water vapour will take place. The resulting clouds areusually of the cumulus type. The front edge of the cold air mass is known as a coldfront. Much of the rain that falls in Australia occurs as a result of cold frontconditions.

Figure 3.8 shows how a cold front causes uplift and condensation in a warmer,humid, air mass.

The arrival of a cold front is marked by a sharp drop in temperature and asudden change of the wind direction.

Text 2 is accompanied by a ‘narrative’ image (Figure 3.8 below), includingtwo appropriately labelled vectors (cold air stream and boundary between air masses)which portray a cold air stream bumping into warm rising humid air and rainfalling from the cumulus clouds which form. Note that the image vectors afforda temporal ‘before and after’ reading of this process, which the verbal textfurther specifies as cause and effect. The application of Kress and vanLeeuwen’s 1996 work on images to vertical discourse of this kind is introducedin van Leeuwen and Humphrey 1996, and further developed in Unsworth 2001.

Text 2 also demonstrates the way in which implication sequences in scienceparticipate in the definition of technical participants, in this case cold fronts (asopposed to warm fronts), and cumulus clouds (as opposed to cirrus, altostratus,etc.). Here, the classification of the terms depends in part on their role in theimplication sequences in which they participate.


42

3 History (horizontal knowledge structure)

Unlike science, history is not a technical discipline. As far as participants areconcerned, uncommon sense classification and composition is simply borrowedfrom other fields. So for history dealing with the politics of migration in Aus-tralia we find borrowed classification for kinds of police force (South AustralianPolice, Australian Federal Police . . .) or types of asylum seekers (Indochinese,Vietnamese, Afghans, Iraqis, Iranians . . .), and borrowed composition forgovernment agencies:

DIMIA Dept of Immigration, Mutlicultural & Indigenous AffairsDFAT Dept of Foreign Affairs & TradeDEET Dept of Employment, Education and TrainingHREOC Human Rights & Equal Opportunity Commissionetc.

Note in passing the use of acronyms adopted from the relevant administrativediscourse here (something which is also a feature of the specialized lexis of tech-nological fields introduced above; White 1997).

Classification particular to history tends to be instantial, arising in the courseof the development of a particular discussion, but not transcending this textinto the field as a whole. In Text 3 below, for example, the historian reports onRuddock’s competing visions for Australia’s population; but this classificationdoes not become part of the technicality of the discipline, even for the specificdiscipline of Australian migration studies. The classification depends onreading this specific text, not on apprenticeship into the field.


Figure 3.8 Narrative image illustrating the cold front explanation (Messel1963: 7.6–7.7)

(3) The minister moves on to outline three competing visions for Australia’s pop-ulation in the century ahead. The first scenario is the high-immigration modelfavoured by some business groups, which call for Australia’s net migration intaketo be set at 1 per cent of existing population per year . . . The second scenario isnet zero migration, the model pushed by sections of the environmentalmovement and by groups such as One Nation, which say that Australia shouldtake just enough migrants to replace the number of people who permanentlydepart the country each year . . . The minister’s final forecast is reassuring –according to him, if we hold fast to the current government policy, Australia’spopulation will increase gradually for the next forty years before settling com-fortably at around 23 million . . . (Mares, P. 2001 Borderline: Australia’s treatmentof refugees and asylum seekers. Sydney: UNSW Press (Reportage Series) 141–2)

For technicality that is specific to history we need to look at activity sequenc-ing, and even here it takes historians a while to get going. Of course for histori-ans, chronology matters; but they construe different kinds of time. In Text 4,sequence in time is foregrounded – the relation of one event to another as hap-pening beforehand or afterwards (as indexed by the left and right pointingarrows in the text below: ‘→’ for ‘subsequently’, ‘←’ for ‘previously’).

(4) On 29 August the Tampa entered into Australian territorial waters approach-ing Christmas Island. → The prime minister told parliament that ← the captainhad decided on this course of action because ← a spokesman for the asylumseekers ‘had indicated that they would begin jumping overboard if medical assis-tance was not provided quickly’. → Captain Rinnan gave a different reason forthis decision: ‘We weren’t seaworthy to sail to Indonesia. There were lifejackets foronly 40 people. The sanitary conditions were terrible.’ → The SAS came aboardand took over Tampa. → An Australian Defence Force doctor was given 43minutes to make a medical assessment of the 433 asylum seekers. → He reported,‘Four persons required IV (2 urgent including 1 woman 8 months pregnant).’ →Captain Rinnan was surprised at the prompt medical assessment, because ← hiscrew had already identified ten people who were barely conscious lying in the sunon the deck of the ship. → The prime minister then made a finely timed ministe-rial statement to parliament insisting that ← ‘nobody – and I repeat nobody – haspresented as being in need of urgent medical assistance as would require theirremoval to the Australian mainland or to Christmas Island’. → One hundred andthirty-one fortunate asylum seekers were granted immediate asylum by the NewZealand government. → The rest, ← having been transported to Nauru, waitedprocessing under the evolving Pacific Solution. (Brennan 2003: 42–3)

In this case the precise timing of events and what was said is crucial, sinceBrennan is keen to show that the government was acting illegally out of self-interest for electoral purposes. In other cases, it is phases of time rather thanindividual events that matter, and chronology is organized through setting intime (realized through prepositional phrases of temporal location, typicallysentence initial) rather than sequence (realized through conjunctions and oneclause following another). In Text 5 below Brennan skips through time alongthese lines from 1978 to 1990.


44

(5) In 1978 the government set up a Determination of Refugee Status (DORS)Committee which would determine onshore refugee claims. A UNHCR represen-tative joined this committee of pubic servants from the key departments. Thecommittee made only recommendations to the minister. If it rejected an applic-ation, it could still recommend that the applicant be given temporary or perma-nent residence ‘on humanitarian or strong compassionate grounds’. In the early1980s the committee considered fewer than 200 applications a year, with less thana one third approval rating.

In 1982 the government decided that even offshore cases would be decided ona case by case basis. It would no longer accept the UNHCR’s blanket determina-tion that anyone from Indochina was a refugee. It was now seven years since theend of the Vietnam War and it was more likely that some of those departing Viet-namese were economic migrants unimpressed by their economic prospects undera communist regime rather than refugees who were fleeing in fear of persecution.At the same time the government set up a Special Humanitarian Program tocomplement the offshore refugee program. In the first year, there were 20,216offshore refugees and 1,701 applicants approved for migration to Australia underthe Special Humanitarian Program. Within eight years there were only 1,537under the offshore refugee category and 10,411 under the Special HumanitarianProgram . . . (Brennan 2003: 29–31)

Once time is phased in this way it can be nominalized, a process of ‘thingifi-cation’ whereby activity is reconstrued as abstract things. In Text 6 below, forexample, phases of Indochinese migration are enumerated as waves (first wave,second wave, third wave, etc.).

(6) The first wave of 2,077 Indochinese boat people came to Australia in 54 boatsbetween 1976 and 1981 . . . The first boatload of asylum seekers arrived in Darwinharbour on 28 July 1976 . . . At the end of that year another two boats arrivedcarrying 106 people . . . When the third Vietnamese boat of the first wave arrived,there was some media agitation . . . (p. 29)

The second wave of boat people commenced with the arrival of a Cambodianboat at Pender Bay near Broome on 25 November 1989 . . . (p. 32)

The third wave of boat people arrived between 1994 and 1998 . . . These Viet-namese and Chinese boat people were the last victims of the Comprehensive Planof Action . . . (p. 40)

The fourth and biggest wave of boat people in modern Australian history couldnot be so readily categorized as non-refugees or as refugees who had their claimsdetermined elsewhere. In late 1999 boat people started arriving fromAfghanistan, Iraq and Iran via Indonesia . . . (p. 40)

Once time is packaged as a thing it can be named; and where proper namesbecome established for phases of history, they transcend the text/s whichcreated them and enter into the field as technical terms. Examples include


Tampa, Mabo, The Sharpeville Massacre, The Long March, The Depression, The FirstGulf War. In short then, we can say that the technicality of history has to do withactivity not participants, but that this activity is reconstrued as a participant thatbecomes named and potentially technicalized. ‘Thingification’ of this order is arecurrent feature of technicality in any vertical discourse.

As well as chronicling, historians have to explain. More so than scientists, theyhave a strong preference for doing this within rather than between clauses. Thecontrast between explaining between clauses and inside them is well illustratedin Text 7 below. Manne begins explaining in the clause (unwillingness affectingresponse), and then unpacks this with three clarifications managing causebetween clauses (each beginning with a because clause preceding its effect).

(7) The Howard government’s unwillingness to apologise determined the natureof its response to other recommendations contained in Bringing them home. ↔Because it refused to consider the present generation of Australians legally ormorally responsible for the mistakes of the past, it refused altogether Bringing themhome’s recommendation for financial compensation for members of the stolengenerations. Because it thought the policies of child removal had been lawful andwell-intentioned, it treated almost with contempt the arguments in Bringing themhome . . . Because, nonetheless, it accepted . . . it was willing to allocate modestsums . . . (Manne 2001: 76)

More typical of historical explanation is a text like the following, in which themajority of cause and effect relations are construed inside a single clause.

(8) In another part of Australia, Aboriginal people were themselves acting toassert their rights. On 23 August 1966 Vincent Lingiari, a Gurindji elder, led hispeople off the cattle station operated by the giant Vesteys pastoral organisation inprotest against their wages and conditions. Their calls for Commonwealthinvolvement also strongly argued the case for land to establish their own cattlestation. They subsequently sent a petition to the Governor-General, with no imme-diate result. Their stand against injustice, however, attracted national publicityfor Aboriginal land rights grievances. The strike developed into a seven-yearcampaign by the Gurindji for the return of their traditional lands and became acause célèbre across Australia. The campaign was strongly supported by the tradeunion movement and sparked a campaign for human rights, including landrights, by many Aboriginal people. It was a cry for Commonwealth leadership thatwould not be acted upon until the election of the Whitlam government. (Tickner2001: 8)

In 8 there is only one explicitly temporal relation, realized through the con-junction subsequently. Otherwise Tickner draws on various clause-internalresources to explain what led to what. He uses causal circumstances (preposi-tional phrases) to explain why Lingiari led his people off (in protest against theirwages and conditions, for the return of their traditional lands), and to explain whatthe Gurindji were appealing for (for Commonwealth involvement . . . leadership).These nominalized appeals (for involvement, leadership) are verbally connected


46

(argued, acted) to further abstractions (the case, the election). And processes arealso used to causally connect the Gurindji walk-off to a burgeoning land rightsmovement (attracted national publicity, sparked a campaign).

Why this drift towards explanation inside rather than between clauses? Onething the ‘cause in the clause’ motif does is enable historians to fine tune causal-ity by deploying verbs which elsewhere literally construe material and verbalactivity (e.g. argue, act, attract, spark), but here enact finely differentiated types ofcause and effect relations. In horizontal discourse, processes tend to haveconcrete readings – with people acting and interacting, often involving otherpeople and things:

Frank argued with Mark.The pool attracted Mike.The lightning sparked a fire.The goal-keeper had to act.

So we find sensuous participants materially affecting sensuous participants inexamples such as these. Contrast the use of these processes in Tickner’s text,where abstractions affect abstractions – calls arguing the case, a stand attractingpublicity, one campaign sparking another, a cry not acted upon:

Their calls for Commonwealth involvement also strongly argued the case forland to establish their own cattle station.

Their stand against injustice, however, attracted national publicity for Abor-iginal land rights grievances.

The campaign . . . sparked a campaign for human rights, including landrights, by many Aboriginal people.

It was a cry for Commonwealth leadership that would not be acted upon untilthe election of the Whitlam government.

This indefinitely enhances historians’ resources for explaining how one eventaffects another. Between clauses English provides for a simple model of cause –one thing leading to another via conjunctions such as so, because and therefore.Inside the clause on the other hand the nominal (cause, reason, basis, source,motive, etc.), prepositional (due to, owing to, because of, thanks to, etc.) and mostimportantly the verbal resources of the language open up for us. This meansthat within the clause historians can fine tune their explanations about the waysin which events give rise to one another. A simple inter-clausal relation is trans-formed into a finely nuanced clause-internal repertoire for interpretation asthese resources are brought to bear. The possibilities are not endless, but indef-initely enlarged:


Their stand against injustice, however, attracted national publicity for Abor-iginal land rights grievances.

Their stand against injustice, however, initiated national publicity for Abor-iginal land rights grievances.

Their stand against injustice, however, kindled national publicity for Abor-iginal land rights grievances.

Their stand against injustice, however, accelerated national publicity for Abor-iginal land rights grievances.

‘Cause in the clause’ is also an important motif in science discourse, as illus-trated in Text 2 above. But unlike historians, scientists prefer a simple model ofcause and effect; they’re not interested in proliferating different kinds of cause(fuzzy cause would be a problem for scientists, not a solution). Even so, asHalliday has shown (Halliday, 1998, 2004), there is still an important pay-offinside the clause for scientists as far as managing argumentation and informa-tion flow is concerned. Note in (2) how the resulting clouds packages up the effectof the preceding sentence by way of participating in the definition of cumulusclouds; and in the next sentence as a result of cold front conditions parcels up thecause of Australian rainfall. The paragraph as a whole is then summed up inrelation to the accompanying figure, which shows how a cold front causes upliftand condensation. The causality may be simple but the explanation is sweet.

(2) . . . This rising air becomes cooler for the reasons mentioned earlier, and if itis humid condensation of water vapour will take place. The resulting clouds areusually of the cumulus type. Much of the rain that falls in Australia occurs as aresult of cold front conditions. Figure 7.7 shows how a cold front causes uplift andcondensation in a warmer, humid, air mass.

As Bernstein has pointed out, horizontal knowledge structures like historydevelop by engendering new languages of description, rather than by resolvingdifference by integrating new perspectives into a single over-arching theory. Sofar we’ve been looking at what we might refer to as modernist history – asegment of history discourse which works towards a grand narrative (Lyotard1984) involving men (sic) of wisdom and courage dicing with fate (lady luck) tosomehow get us where we are today. Thus Tickner immortalizes Lingiari’shistoric walk-off as abstractions affecting abstractions in Text 8 above. An alter-native Marxist stance is illustrated in Text 9 below, which focuses on the nation-alist movement in the Philippines beginning in the late nineteenth century.

(9) This local elite contributed to the growing intellectual ferment and for a timegave direction to the movement for nationhood. But because of their predisposi-tion to compromise and their capitulationist tendencies dictated by their materialaspirations, they ultimately became an impediment to the national struggle. ThePhilippine Revolution was the result of the conjuncture of the unarticulated


48

strivings of the people and the articulations of the ideologues of the emergingelite. The Revolution represented a temporary amalgam of the particular interestsof the elite and the general demands of the masses which eventually broke downinto its respective components during and after the attainment of a national stateand the subsequent incorporation of this new state into the American colonialempire. (Constantino and Constantino 1978: 1–2)

As highlighted in Text 9, Marxist history is more technical than modernisthistory because of the terms it borrows from Marxist economics and politicalthought. The modernist ‘cause in the clause’ syndrome is sustained, and sothese terms enter into explanations as agents of change. In Text 9, for example,we find a causative preposition (because of), a causative nominal group (result)and causative processes (e.g. contributed, dictated) all relating abstractions ofMarxist thought to one another. Where modernist history had abstractionsacting on abstractions, Marxist history technicalizes these abstractions (cf.Wignell, this volume) – with humanities leaning towards social science as onesegment of the field of history moves a little way down the cline of ‘verticality’(Muller, this volume) which distinguished modernist history from the sciencediscourse reviewed above.

prepositional phrase: because of their predisposition to compromise and their capitulationisttendencies. . .

nominal group: The Philippine Revolution was the result of the conjuncture of the unarticulatedstrivings of the people and the articulations of the ideologues of the emergingelite . . .

human agents affecting nominalized abstractions: This local elite contributed to the growing intellectual ferment . . .

abstractions affecting abstractions: . . . and their capitulationist tendencies dictated by their material aspirations . . .

Another important segment in contemporary history discourse is post-colonial readings of the past, illustrated in Text 10 below with a passage fromRafael’s reading of the colonization of large parts of the Philippines by SpanishCatholicism.

(10) The Spanish demand is that nothing be held back in confession. One is toexpend all that memory can hold in a discourse that will bring together both theself that recalls and that which is recalled. The present self that confronts thepriest in confession is thus expected to have managed to control his or her past –to reduce it, as it were, to discursive submission. Whereas the examination of con-science requires the division of the self into one that knows the Law and seeks outthe other self that deviates from it, a ‘good confession’ insists on the presentation


of a self in total control of its past. It is in this sense that confessional discourseimposes on the individual penitent what Roland Barthes called a ‘totalitarianeconomy’ involving the complete recuperation and submission of the past to thepresent, and by extension of the penitent to the priest. (Barthes 1976: 39–75;Rafael 1988: 101–3)

Once again the segment operates with borrowed technicality, this time fromcontemporary critical theory (in this case, the French masters Barthes, Lacanand Foucault). And once again the borrowed technicality is agentive, only thistime round what becomes agentive is not social categories but discursive ones:

a discoursethat will bring together both the self that recalls and that which is recalled

the examination of conscience requiresthe division of the self into one that knows the Law . . .

confessional discourseimposes on the individual penitent what Roland Barthes called a ‘totalitarian economy’.

Put simply, what holds these ‘languages’ together as discourses of history is theirfocus on explaining what happened over time, using cause in the clause to doso. What distinguishes them is what they use to explain. Modernist historynominalizes activity and gets the resulting abstractions acting on one another;Marxist history takes this step but also technicalizes abstractions, drawing onsocial science to do so; post-colonial history technicalizes abstractions as dis-course, drawing on an alternative ‘critical’ canon. The concern with agencyremains; what differs is what ‘acts’ – modernist abstractions, Marxist technical-ity or critical discourse.

4 Grammatical metaphor (the gatekeeper)

In one of his more scintillating appendices, Bernstein (1975: 153) writes:

(11) Imagine four lavatories. The first is stark, bare, pristine, the walls are painteda sharp white; the washbowl is like the apparatus, a gleaming white. A squareblock of soap sits cleanly in an indentation in the sink. A white towel (or perhapspink) is folded neatly on the chrome rail or hangs from a chrome ring. Thelavatory paper is hidden in a cover, and peeps through its slit.

This is pretty clear; even a dumb linguist can understand. Because in thispassage people and things are realized as nouns, descriptive qualities as adjec-tives and processes as verbs:


50

participants as nouns:lavatory, walls, washbowl, apparatus, soap, sink, towel, paper

qualities as adjectives:stark, bare, pristine, sharp, white, square

processes as verbs:imagine, are painted, sits, is folded, is hidden, peeps

Here Bernstein is writing as we speak; we can read the meaning directly off thegrammatical categories. There is no nominalization. The text could be under-stood by a pre-pubescent child. Bernstein moves on:

(11 continued) In the second lavatory there are books on a shelf, pictures on thewall, and some relaxing of the rigours of the first. In the third lavatory there arebooks on the shelf, pictures on the wall, and perhaps a scattering of tiny objects.In the fourth lavatory the rigour is totally relaxed. The walls are covered with amotley array of postcards, there is a wide assortment of reading matter and curios.The lavatory roll is likely to be uncovered and the holder may well fall apart inuse.

Meaning is less straightforward here. Processes are starting to be realized asnouns. We can’t trust the grammar any more. Not all nouns are realizingpeople, places and things; some are actually processes in disguise.

some relaxing of the rigours of the firsta scattering of tiny objectsin use

To get the meaning we have to unpack the nominalizations, translating what thegrammar actually says into what it means. So some relaxing of the rigours of the firstmeans that the second lavatory is not organized as strictly as the first; a scatteringof tiny objects means that tiny objects have been scattered about; and in use meanswhen used.

some relaxing of the rigours of the first(= ‘not organized as strictly’)

a scattering of tiny objects(= ‘tiny objects are scattered (about)’)

in use(= ‘when used’)

Now Bernstein is writing more like academic writing in general, with meaningof all kinds drifting towards the noun. And not surprisingly, as part of this driftto ‘verticality’, we find cause in the clause:


(11 continued) We can say that as we move from the first to the fourth lavatory weare moving from a strongly classified to a weakly classified space; from a spaceregulated by strong rules of exclusion to a space regulated by weak rules ofexclusion.

Bernstein’s sociological abstractions (strong and weak rules of exclusion) nowgovern space. And the text is getting much harder for outsiders to understand.

The process enabling the drift from spoken to written discourse we are mon-itoring here is referred to by Halliday as grammatical metaphor (Halliday 1998,2004; Halliday and Matthiessen 1999, 2004; Simon-Vandenbergen et al. 2003).Politically speaking, grammatical metaphor is Halliday’s most importantconcept, because it names the process which engenders vertical discourse. Whydoes he call it metaphor? What does this process do?

In discourse, grammatical metaphor complements lexical metaphor, so let’sbegin with the latter. Bernstein writes that The lavatory paper is hidden in a cover,and peeps through its slit. Here the lexical metaphor turns on the word peep,which is used in such a way that there are two meanings involved (‘peeps’ and‘just visible’), the meanings are layered with one readily available on the surfaceand another lurking deeper behind (‘peeps’ literal and ‘just visible’ trans-ferred), and one meaning implies the other (‘peep’, which we might define as‘look quickly and secretly’ symbolizing ‘just visible’).

Grammatical metaphor works in the same way, except that it involves struc-tures instead of words. For example, a structure such as a space regulated by strongrules of exclusion is used in such a way that there are two meanings involved, anabstract clause and a more concrete clause complex:

They are layered (literal/transferred; figure/ground; surface/deep).

Medium process Agent – literal &enhancing clause complex – transferred

One implies the other (symbolization)

‘written’ Medium process Agent implying‘spoken’ enhancing clause complex


Medium Process Agent

a space regulated by strong rules of exclusion

. the space was regulated

by using strong rules to exclude things

x β

α

52

In other words, what might well have been realized congruently in spokenlanguage as a sequence of clauses (the space was regulated by using strong rules toexclude things) comes out in the more written form as a nominal group with apost-modifier in which a nominalization (exclusion) is agentive.

What’s at issue here is whether grammar and semantics match. In a tri-stratalmodel of language such as that introduced in Figure 3.1, we’re asking how dis-course semantic categories such as processes, participants, qualities and logicalrelations are related to lexicogrammatical ones such as verb, noun, adjectiveand conjunction. Because of the way language has evolved and the way we alllearn it, there is a natural pairing off of processes with verbs, participants withnouns, qualities with adjectives and logical relation with conjunctions, asoutlined in Figure 3.9 below.

But as we have already seen, in abstract discourse we find processes, qualitiesand logical relations realized as nouns, and logical relations realized as verbs –the skewed realization outlined in Figure 3.10. From the perspective ofgrammar, the key derivational process here is nominalization. Looking acrossstrata, the result is stratal tension – a non-matching relation between grammarand semantics. The result is that meaning can’t be read directly off the wording.We have to read grammatical metaphors twice – once with respect to what the


Figure 3.9 Stratal harmony – grammar matching semantics

process

verb

participant

noun

adjective

conjunction

discoursesemantics

quality

logicalrelation

lexico-grammar

phonology/graphology

grammar is, and once again with respect to what lies behind. And the meaningof the metaphor is more than the sum of its parts (grammar ‘times’ semantics,if you will).

We have already touched on the pay-offs for the construction of vertical dis-course, using examples from science and history. Let’s consolidate them here,using examples from Bernstein’s sociology. First, participants. In vertical dis-course, uncommon sense classification and composition operates on entities. Sowhatever we choose to classify or decompose, we have to ‘thingify’ it first.Beyond this, in technical disciplines, terms have to be defined; and definitionsalso operate on entities. So technicality too depends on ‘thingification’. Here,for example, are Bernstein’s definitions of vertical and horizontal discourse, akey step in his classification of discourse and knowledge structures (2000: 157,his italics), annotated with Halliday’s labels for the technical term (Token) andits definition (Value) in identifying structures of this kind:

. . . a Vertical discourse (Token)

takes the form of (=)


Figure 3.10 Grammatical metaphor as stratal tension

process

verb

participant

noun

adjective

conjunction

discoursesemantics

quality

logicalrelation

lexico-grammar

phonology/graphology

54

a coherent, explicit and systematically principled structure, hierarchicallyorganized as in the sciences, or it takes the form of a series of specialized lan-guages with specialized modes of interrogation and specialized criteria forthe production and circulation of texts as in the social sciences and humani-ties. (Value)

A Horizontal discourse (Token)

entails (=)

a set of strategies which are local, segmentally organised, context specific anddependent, for maximising encounters with persons and habitats. (Value)

In short then, the uncommon sense organization of participants in verticaldiscourse depends on grammatical metaphor. From a functional linguisticperspective, if no grammatical metaphor, then no verticality.

Complementing this is the role of grammatical metaphor in explanation (i.e.vertical sequencing). In the following passage Bernstein focuses on strategiesfor the regulation of knowledge in a community (2000: 158):

(12) Consider a situation where a small holder meets another and complains thatwhat he/she had done every year with great success, this year failed completely.The other says that when this happened he/she finds that this ‘works’. He/shethen outlines the successful strategy. Now any restriction to circulation andexchange reduces effectiveness. Any restriction specialises, classifies and privatisesknowledge. Stratification procedures produce distributive rules which control theflow of procedures from reservoir to repertoire. Thus both Vertical and Horizon-tal discourses are likely to operate with distributive rules which set up positions ofdefence and challenge.

As with the history and science explanations reviewed above, we find abstrac-tions acting on abstracts, including technical abstractions indexing the field associal science. As with history, a wide repertoire of verbs is deployed, connectedcause to effect (with abstractions labelled as Agent and Medium, after Halliday):

Now any restriction to circulation and exchange (Agent)reduceseffectiveness (Medium)

Any restriction (Agent)specializes, classifies and privatizes knowledge (Medium)

Stratification procedures (Agent) producedistributive rules . . . (Medium)


distributive rules which (Agent)controlthe flow of procedures from reservoir to repertoire (Medium)

. . . distributive rules which (Agent)set up position of defence and challenge (Medium)

Once again then, this time as far as explaining is concerned, if no grammaticalmetaphor, then no verticality.

This means that from a functional linguistic perspective access to verticaldiscourse is bound up with control of grammatical metaphor, which in westernsocieties students are expected to master in secondary school. Failure to accessthis recourse entails exclusion from hierarchical and horizontal knowledgestructures. Here lies the social semiotic nub of institutionalized learning,educational failure and the distribution of knowledge in our expiring world(Martin 1993a, b, 1999, 2000; Muller 2000; Taylor et al. 2003).

5 Genre and field (and value)

As pointed out in section 1, my interest in field developed out of my interest ingenre, as part of designing literacy programmes for secondary school. And asoutlined in Figure 3.2, we used genre to model the interaction of the registervariables field, tenor and mode. In these terms genre theory is concerned withhow a culture maps ideational, interpersonal and textual meaning onto oneanother in phases as a text unfolds. Part of this interpersonal meaning is value,which SFL approaches from the perspective of appraisal theory (Macken-Horarik and Martin 2003; Martin and White 2005). So an important aspect oftext analysis has to be the infusion of field with value.

Returning to Text 3 above, Minister Ruddock doesn’t just classify Australia’sfuture into three scenarios; he evaluates each forecast as he goes. ‘Goldilocks’Ruddock tries the business position, which is too fast (trends alarmingly upswards,betrays a certain lack of realism); the extreme left green cum extreme right nation-alist position, which is too slow (population slump, sags in a depressing arc); andthen the government position, which is just right (reassuring, favourable outlook,rises gently and unthreateningly)!

(3) The minister moves on to outline three competing visions for Australia’s pop-ulation in the century ahead. The first scenario is the high-immigration modelfavoured by some business groups, which call for Australia’s net migration intaketo be set at 1 per cent of existing population per year. The red line on the graphtrends alarmingly upwards, predicting an ‘inevitably rising’ population, whichwould hit 65 million in the year 2007. According to the minister, the businessgroups’ goal of 1 per cent net migration betrays ‘a certain lack of realism’. Thesecond scenario is net zero migration, the model pushed by sections of the envi-ronmental movement and by groups such as One Nation, which say that Australia


56

should take just enough migrants to replace the number of people who perma-nently depart the country each year. With falling birthrates, this would see Aus-tralia’s population slump from 20 million to 14 million within the next century.The green line on the graph sags in a depressing downward arc, headinginevitably towards zero. The minister’s final forecast is reassuring – according tohim, if we hold fast to the current government policy, Australia’s population willincrease gradually for the next forty years before settling comfortably at around23 million. This favourable outlook is represented on the graph by a blue line,which rises gently and unthreateningly before charting a stable course between theextremes of green and red. (Mares 2001: 141–2)

This coupling of knowledge with value is an important dimension for any field.Because of the metafunctional organization of language and register in Figure3.2, the natural place to bring together ideational and interpersonal meaning isat the level of genre. We could argue for example that Ruddock’s classification(field) and ranking of scenarios (tenor) come together in an evaluativetaxonomizing report which positions readers to align themselves with a neo-conservative government migration and refugee agenda. Alternatively, it may bethat we should revise the model to allow for the coupling of content and valueswithin the register variable field, since apprenticeship to any field alwaysinvolves not just learning how to do things but learning what to feel about themas well.

With respect to Bernstein’s model, this makes me wonder whether valueswould be considered an essential dimension of vertical knowledge structures. Asfar as I can see, there is nothing in Bernstein’s programmatic discussion pre-cluding this, although from a linguistic perspective his analysis and exemplifi-cation seems to be purely ideational.

6 Linguistics as a knowledge structure

Bernstein (e.g. 1996: 46–50) defines pedagogic discourse as a rule embedding‘a discourse of skills and their relation to each other (instructional discourse)’in ‘a discourse of social order (regulative discourse)’, the purpose of the devicebeing to ‘produce a symbolic ruler for consciousness’. Following Christie 2002,I prefer the term projection to embedding, with the regulative discourse projectingthe instructional one. ‘Sydney School’ literacy programmes in fact involve adoubling of the instructional discourse, since disciplines are factored as systemsof genres (and thus of field, tenor, mode constellations and of their realizationin language and attendant non-verbal modalities of communication); so inclassroom practice a discipline such as history is projected by regulative dis-course via its linguistic deconstruction (actually its social semiotic deconstruc-tion since all relevant modalities of communication are considered) – inBernstein’s terms, history is ‘embedded’ in linguistics (so we have two instruc-tional discourses), which is in turn ‘embedded’ in regulative discourse.

As far as the design of our interventions is concerned, social semiotics is usedto deconstruct the pedagogic discourse (not just instructional discourse); so in


theory, history is projected by regulative discourse projected by social semiotic‘instructional’ discourse; some critical theorists have argued that pedagogy andcurriculum should be negotiated between teachers and students along theselines, by way of challenging the fundamental asymmetry of the teacher/learnerrelationship (a step we have not taken as far as I am aware). Bernstein’s con-ceptual notation for embedding in pedagogic discourse is adapted in Figure3.11 below.

Key: ID instructional discourse*SSID social semiotic instructional discourseRD regulative discourse

Factoring a discipline as a system of genres clearly recontextualizes that dis-cipline. In secondary school, content which a teacher conceives of in purelyideational terms is reworked with value and texture into the text types studentsneed to read and write. As part of this the genres need to be named, along withtheir stages; and then there is the question of how much register, discoursesemantics, lexicogrammar, graphology and image grammar is brought into play.What knowledge about language (KAL) do secondary history students need toknow to learn grammatical metaphor, for example, especially those studentswho can’t learn it by osmosis the way motivated middle class students apparentlydo? In our experience, outside of the typical English classroom where KAL isseemingly taboo, the names of key genres and their stages can be mobilized, butas a first step much more linguistic metalanguage than this is felt to be goingtoo far too fast.

An exemplary factoring of history as a set of genres is outlined in Table 3.1,which has been ordered as a learner pathway, beginning with genres similar tothose students will be familiar with in the oral culture outside of school, andmoving on through those particular to history, including various chroniclinggenres and also argumentation. The same genres are presented as a systemnetwork, SFL style, in Figure 3.12 – arranged vertically along a topologicalvector running from common through more and more uncommon sense. Forfurther discussion and illustraton see Coffin 1997, 2000; Martin 2002a; Martinand Wodak 2004; Martin and Rose 2005.


Figure 3.11 Projection relations in pedagogic discourse

ID

RDRD

ID

SSID

SSID

ID

RD

58

As far as vertical discourse is concerned, what kind of recontextualization isthis? One big transformation is the difference in the level of technicality. Lin-guistics is arguably the most technical discipline in the humanities, and possiblysocial sciences as well – the ‘science of language’ as it is often portrayed. Thereis also a fair degree of ‘verticality’ in its knowledge structure. As in science,networks for participants are deep, comprehensive and precise; and composi-tion is an exacting discipline, as anyone trained in text generation and parsingcan confirm. Explanation is perhaps less developed, since linguists are stillunsure what their data is (intuitions about grammaticality, selected texts orcorpora) and how to manage it (quantitatively, qualitatively); and all dimen-sions of language change are relatively poorly understood – logogenesis(unfolding text), ontogenesis (language development in the individual) andphylogenesis (language evolution and death). Teachers coming from horizon-tal discourses with relatively weak grammars (as Bernstein puts it) find this arrayof technicalization a shock; it’s hard to recognize one’s subject matter for thetrees as it were. So the argument that literacy and thus content learning in factdepends for many students on projecting subject matter through SFL is an easyone to lose. The new knowledge about language required costs time and there-fore money, teachers are busy, and stratified learning outcomes are blamed onthe ability level of individual students who don’t learn grammatical metaphor


Genre [staging] Informal description Hurdling

personal recount agnate to story genres;[OrientationˆRecord] what happened to me

autobiographical recount the story of my life[OrientationˆRecord] [oral history]

biographical recount the story of someone[OrientationˆRecord] else’s life

historical recount; establishing the time line of[BackgroundˆRecord] ‘grand narrative’

historical account; naturalizing linearization of[BackgroundˆAccount] ‘grand narrative’ as causal

factorial explanation complexifying notion of[OutcomeˆFactors] what leads on to what

consequential explanation complexifying notion of[InputˆConsequences] what leads on from what

exposition – one sides; promote problematic interpretation[ThesisˆArguments] that needs justifying

challenge – one sided; rebut problematic interpretation[PositionˆRebuttal] that needs demolishing

discussion – multi-sided; adjudicate more than one[IssueˆSidesˆResolution] interpretation considered

Table 3.1 Factoring modernist history as genre (learner pathway)

because they can’t (or perhaps won’t: Rose 2004). But democracy aside, arethese stratified outcomes something we can still afford?

In Bernstein’s terms, for all its scientific aspirationalism, linguistics is a hori-zontal knowledge structure, made up of competing languages of description –formal and functional schools for example, and many sub-varieties within,between and around blunt categorizations such as this. As Muller (this volume)comments, specific segments of a horizontal knowledge structure with stronggrammars may exhibit a degree of verticality that approximates that idealized byphilosophers of science for science proper (especially physics). And SFL is cer-tainly a segment of this kind – not just in its technicality, but in its evolutionarypath of development. The theory is robust enough to fit new perspectives in (forexample, genre as a deeper stratum, appraisal in interpersonal discourse seman-tics, multimodality as mode, corpus counting in relation to probabilitisticsystems, and so on).

Bernstein (2000: 161) uses the image of a triangle as a model for hierarchicalknowledge structures, to capture the sense in which they strive for general andabstract propositions integrating knowledge at lower levels. This metaphor istaken up in Wignell’s (this volume) characterization of social sciences aswarring triangles. And this seems an apt characterization of SFL and its rivaltheories of language, each developing what they see as the kind of verticalitythey need to wipe out the others and become THE science6 of language linguistsare bred to need. Lack of progress in this war confirms linguistics as preciselythe kind of horizontal knowledge structure Bernstein recognizes it to be.


Figure 3.12 Typology of history genres

fieldtime

text time

explain

argue

Commonsense

Uncommonsense

temporal

biographical

first person autobio-graphicalrecount

third person biographicalrecount

historical historicalrecount

one-sided

multi-sided discussion

promote exposition

rebut challenge

inputs factorialexplanation

outcomes consequentialexplanation

causal historicalaccount

60

One comment I would make here, however, is that Bernstein’s typology ofknowledge structures, with hard science at one end and soft humanities at theother, is a relatively modernist conceit. It may well be that a social semiotictheory is a different kind of knowledge structure from those dealing with thephysical, biological and social world. In this regard, the notion of an apex inBernstein’s triangle metaphor is a potentially misleading one, at least as far asSFL is concerned. The theory is more of a network of parameters than apyramid with a broad bottom and a narrow tip – including several scales (e.g.delicacy, rank, realization, instantiation and individuation), and complemen-tarities (e.g. metafunction, agnation, modality, axis, perspective, and of coursethe scales just reviewed). A rough characterization of the bits and pieces of thisevolving theoretical web is offered below (including in parentheses somecomparable conceptualizations from Bernstein’s work).

SFL scales (magnification):

rank clause–group/phrase–word–morphemedelicacy primary–secondary–tertiary . . . (forms of discourse: knowledge)realization phonology–lexicogrammar–discourse semantics–

register–genreinstantiation system–register–text type–text–reading (real/empirical)individuation community–coding orientation–identity (reservoir/repertoire)

SFL complementarities (multinocular vision):

metafunction ideational/interpersonal/textual (classification/framing)axis system/structureagnation typology/topology (knowledge map/pedagogy matrix)perspective synoptic/dynamicmodality (MDA) language/image/music/architecture . . . (discourse/lavatory)(and scales above)

For an overview of these parameters see Halliday and Matthissen 1999, whichtaken alongside Halliday 2003 suggests that knowledge structures dealing withsocial semiosis are much more sophisticated than anything developed byscience because of the complexity of the phenomena in their gaze. Thus oneday, paraphrasing Whorf, science (and I would extend this to include all verticaldiscourse) will have to look to linguistics to find out what it means.

7 Envoi

In this chapter I have attempted to sketch out various parameters involved in asocial semiotic interpretation of discourses and knowledge structures as field –using science and history as exemplars. Central to this overview was the intro-duction of grammatical metaphor as the critical linguistic resource which hasevolved in written cultures for the construction of vertical discourse. Because of


its ability to reconstrue any meaning as a thing, and thereby manage causalityinside the clause, it is grammatical metaphor that enables uncommon senseclassification, composition and explanation right across humanities, socialscience and science. The power of these discourses to consume our ecosystem(via technology) and manage populations (via bureaucracy) engenders boththeir prestige and the alienation their privileging verticality entails.

Thanks to Bernstein and Halliday our legacy comprises a sociology with a placefor language and a linguistics with a commitment to the social. These will neverbe one theory, one knowledge structure; nor would this loss of incommensura-bility be in any way desirable. But shared political commitment, especially with afocus on education, has maintained dialogue through to a second generation ofresearch. I sense that if we can sustain this mutual interrogation, our under-standings can grow; and more importantly, our interventions can prosper.

References

Bernstein, B. (1975), Towards a Theory of Educational Transmissions. Class, Codes and ControlVol. 3. London: Routledge & Kegan Paul.

Bernstein, B. (1990), The Structuring of Pedagogic Discourse. London: Routledge.Bernstein, B. (1996), Pedagogy, Symbolic Control and Identity: theory, research, critique (Revised

edition, 2000). London: Taylor and Francis. Christie, F. (1992), ‘Literacy in Australia’, in ARAL 12, pp. 142–55.Christie, F. (ed.) (1999), Pedagogy and the Shaping of Consciousness: linguistic and social

processes. London: Cassell (Open Linguistics Series). Christie, F. (2002), Classroom Discourse Analysis. London: Continuum.Christie, F. and Martin, J.R. (eds) (1997), Genre and Institutions: social processes in the work-

place and school. London: Cassell.Coffin, C. (1997), ‘Constructing and giving value to the past: an investigation into

secondary school history’, in Christie and Martin (eds), pp. 196–230.Coffin, C. (2000), ‘Defending and challenging interpretations of the past’, in Revista

Canaria de Estudios Ingleses 40, pp. 135–54.Cope, W. and Kalantzis, M. (eds) (1993), The Powers of Literacy: a genre approach to teaching

literacy. London: Falmer and Pittsburgh: University of Pittsburgh Press. Eggins, S., Martin, J.R. and Wignell, P. (1993), ‘The discourse of history: distancing the

recoverable past’, in Ghadessy, M. (ed.), Register Analysis: theory and practice. London:Pinter (Open Linguistics Series), pp. 75–109.

Feez, S. (1998), Text-based Syllabus Design. Sydney: National Centre for English LanguageTeaching and Research (NELTR), Macquarie University.

Feez, S. (2002), ‘Heritage and innovation in second language education’, in Johns, A.M.(ed.), pp. 43–69.

Ghadessy, M. (ed.) (1993), Register Analysis: theory and practice. London: Pinter.Ghadessy, M. (ed.) (1999), Text and Context in Functional Linguistics (CILT Series IV).

Amsterdam: Benjamins.Haire, M., Kennedy, E., Lofts, G. and Evergreen, M. (2005), Core Science: Stage 5 Essential

Content (2nd edition). Milton, Qld: John Wiley & Sons.Halliday, M.A.K. (1998), ‘Things and relations: regrammaticising experience as technical

knowledge’, in Martin and Veel (eds), pp. 185–235.Halliday, M.A.K. (2003), On Language and Linguistics (Vol. 3 in The Collected Works of

M.A.K. Halliday, ed. J. Webster). London: Continuum.Halliday, M.A.K. (2004), The Language of Science. (Vol. 5 in The Collected Works of M.A.K.

Halliday, ed. J. Webster). London: Continuum.


62

Halliday, M.A.K. and Martin, J.R. (1993), Writing Science: literacy and discursive power(Critical Perspectives on Literacy and Education). London: Falmer.

Halliday, M.A.K. and Matthiessen, C. (1999), Construing Experience through Language: alanguage-based approach to cognition. London: Cassell.

Halliday, M.A.K. and Matthiessen, C. (2004), An Introduction to Functional Grammar (3rdedition). London: Arnold.

Hasan, R. and Williams, G. (eds) (1996), Literacy in Society. London: Longman.Johns, A.M. (ed.) (2002), Genres in the Classroom: applying theory and research to practice.

Mahwah, NJ: Lawrence Erlbaum.Kress, G. and van Leeuwen, T. (1996), Reading Images: the grammar of visual design. London:

Routledge.Kress, G. and van Leeuwen, T. (2001), Multimodal Discourse – The Modes and Media of

Contemporary Communication. London: Arnold.Kress, G., Jewitt, C., Ogborn, J. and Tsatsarelis, C. (2001), Multimodal Teaching and

Learning: the rhetorics of the classroom. London: Continuum.Lemke, J.L. (1990), Talking Science: language, learning and values (Language and Educa-

tional Processes). Norwood, NJ: Ablex.Lyotard, J. (1984), The Postmodern Condition. Minneapolis: University of Minnesota Press.Macken-Horarik, M. (2002), ‘Something to shoot for: a systemic functional approach to

teaching genre in secondary school science’, in Johns (ed.), pp. 17–42.Macken-Horarik, M. (1998), ‘Exploring the requirements of critical school literacy: a view

from two classrooms’, in F. Christie and R. Misson (eds), Literacy and Schooling.London: Routledge, pp. 74–103.

Macken-Horarik, M. and Martin, J.R. (eds) (2003), ‘Negotiating heteroglossia: socialperspectives on evaluation’, (Special Issue of Text 23.2).

Martin, J.R. (1992), English Text: system & structure. Amsterdam: Benjamins.Martin, J.R. (1993a), ‘Genre and literacy – modelling context in educational linguistics’,

in Annual Review of Applied Linguistics 13, pp. 141–72.Martin, J.R. (1993b), ‘Technology, bureaucracy and schooling: discursive resources and

control’, in Cultural Dynamics 6.1, pp. 84–130.Martin, J.R. (1993c), ‘Life as a noun’, in Halliday and Martin, pp. 221–67.Martin, J.R. (1995), ‘Text and clause: fractal resonance’, in Text 15.1, pp. 5–42.Martin, J.R. (1998), ‘Practice into theory: catalysing change’, in S. Hunston (ed.),

Language at work. Clevedon: Multilingual Matters (British Studies in Applied Lin-guistics 13 [BAAL 13]), pp. 151–67.

Martin, J.R. (1999), ‘Linguistics and the consumer: theory in practice’, in Linguistics andEducation 9.3, pp. 409–46.

Martin, J.R. (2000), ‘Design and practice: enacting functional linguistics in Australia’, inAnnual Review of Applied Linguistics 20 (20th Anniversary Volume: ‘Applied Linguis-tics as an Emerging Discipline’), pp. 116–26.

Martin, J.R. (2001), ‘Giving the game away: explicitness, diversity and genre-based literacyin Australia’, in R. de Cilla, H. Krumm and R. Wodak et al. (eds), Functional Il/literacy.Vienna: Verlag der Osterreichischen Akadamie der Wissenschaften, pp. 155–74.

Martin, J.R. (2002a), ‘Writing history: construing time and value in discourses of thepast’, in C. Colombi and M. Schleppegrell (eds), Developing Advanced Literacy in Firstand Second Languages. Mahwah, NJ: Erlbaum, pp. 87–118.

Martin, J.R. (2002b), ‘From little things big things grow: ecogenesis in school geography’,in R. Coe et al. (eds), The Rhetoric and Ideology of Genre: strategies for stability and change.Cresskill, NJ: Hampton Press, pp. 243–71.

Martin, J.R. and Rose, D. (2003), Working with Discourse: meaning beyond the clause. London:Continuum.

Martin, J.R. and Rose, D. (2006), Genre Relations: mapping culture. London: Equinox.Martin, J.R. and Veel, R. (eds) (1998), Reading Science: critical and functional perspectives on

discourses of science. London: Routledge.


Martin, J.R. and White, P.R. (2005), The Language of Evaluation: appraisal in English.London: Palgrave.

Martin, J.R. and Wodak, R. (eds) (2004), Re/reading the past: critical and functional perspec-tives on discourses of history. Amsterdam: Benjamins.

Muller, J. (2000), Reclaiming Knowledge: social theory, curriculum and education policy (Knowl-edge, Identity and School Life Series 8). London: Routledge.

Muller, J., Davies, B. and Morais, A. (eds) (2004), Reading Bernstein, Researching Bernstein.London: RoutledgeFalmer.

Ogburn, J., Kress, G., Martins, I. and McGillicuddy, K. (1996), Explaining Science in theClassroom. Buckingham: OUP.

Painter, C. (1998), Learning through Language in Early Childhood. London: Cassell.Rose, D. (1997), ‘Science, technology and technical literacies’, in Christie and Martin,

pp. 40–72.Rose, D. (1998), ‘Science discourse and industrial hierarchy’, in Martin and Veel, pp.

236–65.Rose, D. (2004), ‘Sequencing and Pacing of the Hidden Curriculum: how indigenous

children are left out of the chain’, in Muller et al., pp. 91–107.Schleppegrell, M.J. (2004), The Language of Schooling: a functional linguistics perspective.

Mahwah, NJ: Erlbaum.Simon-Vandenbergen, A.-M., Taverniers, M. and Ravelli, L.J. (eds) (2003), Metaphor:

systemic and functional perspectives. Amsterdam: Benjamins.Taylor, N., Muller, J. and Vinjevold, P. (2003), Getting School Working: research and systemic

school reform in South Africa. Cape Town: Peason Education.Unsworth, L. (1997a), ‘Scaffolding reading of science explanations: accessing the gram-

matical and visual forms of specialised knowledge’, in Reading 31.3, pp. 30–42.Unsworth, L. (1997b), ‘Explaining explanations: enhancing scientific learning and

literacy development’, in Australian Science Teachers Journal 43.1, pp. 34–49.Unsworth, L. (1998), ‘“Sound” explanations in school science’, in Linguistics and Educa-

tion 9.2, pp. 199–226.Unsworth, L. (ed.) (2000), Researching Language in Schools and Communities: functional

linguistics approaches. London: Cassell. Unsworth, L. (2001), Teaching Multiliteracies across the Curriculum: changing contexts of text

and image in classroom practice. Buckingham: Open University Press.White, P. (1998), ‘Extended reality, proto-nouns and the vernacular: distinguishing the

technological from the scientific’, in Martin and Veel, pp. 266–96.van Leeuwen, T. and Humphrey, S. (1996), ‘On learning to look through a geographer’s

eyes’, in Hasan and Williams, pp. 29–49.Veel, R. (1992), ‘Engaging with scientific language: A functional approach to the

language of school science’, in Australian Science Teachers’ Journal 38.4, pp. 31–5.Wignell, P., Martin, J.R. and Eggins, S. (1990), ‘The discourse of geography: ordering and

explaining the experiential world’, in Linguistics and Education 1.4, pp. 359–92.(Republished in Halliday and Martin 1993, pp. 136–65.)

WORKPLACE LITERACY REPORTS:1 Rose, D., McInnes, D. and Korner, H. (1992), Scientific Literacy (Write it Right Literacy

in Industry Research Project – Stage 1). Sydney: Metropolitan East DisadvantagedSchools Program.

2 Iedema, R., Feez, S. and White, P. (1994), Media Literacy (Write it Right Literacy inIndustry Research Project – Stage 2). Sydney: Metropolitan East Disadvantaged SchoolsProgram.

3 Iedema, R. (1995), Literacy of Administration (Write it Right Literacy in Industry ResearchProject – Stage 3). Sydney: Metropolitan East Disadvantaged Schools Program.


64

SECONDARY SCHOOL LITERACY PACKAGES:

EnglishRothery, J. (1994), Exploring Literacy in School English (Write it Right Resources for Literacy and

Learning). Sydney: Metropolitan East Disadvantaged Schools Program. Rothery, J. and Stenglin, M. (1994), Spine-Chilling Stories: a unit of work for Junior Secondary

English (Write it Right Resources for Literacy and Learning). Sydney: Metropolitan EastDisadvantaged Schools Program.

Rothery, J. and Stenglin, M. (1994), Exploring Narrative in Video: a unit of work for Junior Sec-ondary English (Write it Right Resources for Literacy and Learning). Sydney: MetropolitanEast Disadvantaged Schools Program.

Rothery, J. and Stenglin, M. (1994), Writing a Book Review a unit of work for Junior SecondaryEnglish (Write it Right Resources for Literacy and Learning). Sydney: Metropolitan EastDisadvantaged Schools Program.

Geography1 Humphrey, S. (1996), Exploring Literacy in School Geography. Sydney: Metropolitan East

Disadvantaged Schools Program.2 Humphrey, S. and Takans, P. (1996), Explaining the Weather: a unit of work for Junior

Secondary Geography. Sydney: Metropolitan East Disadvantaged Schools Program.3 Sikes, D. and Humphrey, S. (1996), Australia – Place and Space: a unit of work for Junior

Secondary Geography. Sydney: Metropolitan East Disadvantaged Schools Program.

History1 Coffin, C. (1996), Exploring Literacy in School History. Sydney: Metropolitan East Dis-

advantaged Schools Program.2 Brook, R., Coffin, C. and Humphrey, S. (1996), Australian Identity: a unit of work for

Junior Secondary History. Sydney: Metropolitan East Disadvantaged Schools Program.

Notes

1 For overviews of this genre-based literacy pedagogy see Christie 1992; Cope andKalantzis 1993; Feez 1998, 2002; Macken-Horarik 2002; Martin 1993a, 1998, 1999,2000, 2001.

2. The Write it Right project workplace and secondary school literacy reports docu-menting this research are listed at the end of the references.

3 The ways in which participants configure with participants and circumstances viaprocesses as part of activity sequences is of course part of this picture, but we will passover this dimension of analysis here; see Martin 1992 for discussion.

4 Generally referred to as implication sequences in our work on science because oftheir ‘if a happens, then b ensues’ structure; see Halliday and Martin 1993 for dis-cussion.

5 For discussion of pre-school learning in the home, see Painter 1998.6 Somewhat ironically, it’s formal linguists, who try to present themselves as more sci-

entific than others, who are most responsible for proliferating languages of descrip-tion in linguistics, rolling their theories over in a post-Fordist carnival to ensure jobsfor each new wave of graduate students; since these theories aren’t actually aboutvery much of language, and not really used for anything outside the discipline, theyare easy to change – and the power of their self-righteous exclusionary authority ismaintained.


To me, wherever there is pedagogy there is hierarchy. What is interesting, it’s thelanguage of description that we use, because the language of description maskshierarchy, whereas the language of description should attempt to sharpen its pos-sibility of appearance.

(Bernstein, in Morais et al. (eds) (2001), p. 375)

In every hair there are an infinite number of lions, and in addition all the singlehairs, together with their infinite number of lions, in turn enter into a single hair.In this way the progression is infinite, like the jewels in Celestial Lord Indra’s net.

(Fa-Tsang, T’ang Dynasty, quoted in Collins 1998, p. v)

The trouble with hierarchy

The set of papers in this volume can loosely be said to be about hierarchy indiscourse – in language, knowledge, and pedagogy. The two disciplinary com-munities represented in this volume do not deal with hierarchy in the same way,but they are at least endeavouring to find common ground for representing itin the same discursive space. This is a large step forward. In fact, hierarchy is aword we generally shun in our lexicons, infused as they are with one or otherkind of egalitarianism, which is generally speaking a good thing. Deeplyembedded in our egalitarian zeitgeist is the unassailable assumption that hier-archy in discourse and hierarchy in society are connected, together with a corol-lary assumption that one inoculates oneself from the latter by avoiding talkabout the former. Alas, sympathetic magic of this sort works about as well insocial science as it does in folk medicine.

Our two communities are on the face of it engaged in nicely complementaryprojects. The linguists are engaged in establishing what the building blocks ofhierarchy are, while the sociologists of education are engaged in establishinghow hierarchy is distributed. Both communities have made some progress, butneither community has settled the issue of what exactly discursive hierarchyderives from, or what knowledge hierarchy is. There are at least three obstaclesthat I can see. The first is that we are locked into an early (lexical) metaphori-cal stage of discussion, where the terms are more suggestive than they areexplanatory, and where use of the same term does not guarantee equivalentmeaning. Secondly, the egalitarian ethos seems to drive us to use terms thatobdurately suggest variety without hierarchy – discovery, variation, and the like.Thirdly, a currently influential trend in social science that I call the ‘New

4 On splitting hairs: hierarchy, knowledge and theschool curriculum

Johan Muller

66

Cynicism’ below, denies the very existence of discursive hierarchy, making alltalk about it sacrilegious. Fortunately, the two communities gathered togetherin this volume seem united in the view that the devil you know is better than theone you don’t, and that understanding hierarchy in discourse is a necessaryprelude to combating it in society. That is promising indeed.

Though we routinely avoid addressing it directly, we willy-nilly imply hierarchy:to take an example almost at random, Hasan (2001) refers to ‘higher mentalfunctions’, after Vygotsky. One infers that ‘higher’ functions enable greaterabstraction, hence greater combinatorial power. But what about the levels in theparent discourse within which this higher abstraction is embedded? From a con-structivist ‘knower’ perspective, it is almost as if each neophyte knower inventsthe levels ab initio. From a ‘knowledge’ perspective, where it is the distributivepotential of different kinds of knowledge that is at stake, it is precisely knowledgehierarchy that conditions distributive potential. So there is no avoiding it.

Martin (this volume) tackles it head on. Starting with Bernstein’s commonsense-uncommon sense cline, Martin produces a branching diagram that,despite proceeding in the opposite direction to Bernstein’s (2000: 168) ownbranching diagram, anticipates its principal features in a quite remarkable way.1

The principal similarity is that both Martin and Bernstein use the branchingtree device to complicate and fill in the continuum between the two ends of thespectrum. With this, the explanatory implication of the tree shifts fromdichotomy to continuity. The question then becomes: how far do we want topush the implication? Do we want to imply that there is a complete continuum,or are there categorical differences between some of the symbolic ensemblesunbranched by the tree?

The answer depends upon whether one is primarily interested in knowledgeas meaning or knowledge as distributed social goods. If one is interested in know-ledge as meaning, then one is primarily interested in describing the universalsemantic building blocks that enable transition from one form to the other.Martin goes on exemplarily to show how grammatical metaphor ‘engenders’the ‘drift’ from spoken to written discourse, from horizontal to vertical dis-course (Martin, this volume). If one is primarily interested in knowledge as dis-tributed social goods, then one is interested in describing the way both formshave distributive rules which are in turn conditioned by discontinuities insemiotic structure that mirror, sustain and reproduce inequalities in society. Forthe first, the explanatory task is to uncover the most basic universal processesand hence to reveal the ideal underlying unity of semiosis; for the secondproject, the task is to delineate the social limits to distributive equality. The firstis, if you like, a classical, ultimately optimistic enterprise; the second is a funda-mentally tragic enterprise, no matter how optimistically driven.

My intention is not to separate the enterprises. Quite the opposite. I hope tosharpen the difference in the projects and starting points the better to makevisible where the projects at present overlap and where bridge-building shouldconcentrate. Both Bernstein and Martin distinguish between discourse formstied to empirical particulars (to ‘context’), and those which do not dependupon the world in order to make sense, which in fact float free from it. For Bern-stein, a fundamental distinction between what he called horizontal and vertical


discourse is that the latter is integrated at the level of meanings allowing de-contextualization, the former at the level of (culturally specialized) segmentsbinding the language to context. In his contrast between ‘everyday’ and ‘scien-tific’ taxonomies, Martin makes the same point. A key characteristic of everydaycategories is that they are largely learnt ostensively or by modelling, that is, bymodes of discursive action that require the coincidence of time and space, forBernstein placing a constraint on their distribution. In middle class homes, asPainter (1999) shows, the transition from ostensive definition to decontextual-ized definition (to technical lexis), ‘allowing the speaker to attend to themeaning rather than the referent’ (Painter 1999: 82), is pretty well seamless, atleast for young Stephen busy acquiring his semantic style (see also Painter, thisvolume). Bernstein would be in full agreement, but would want to make some-thing more of the social gulf between the two forms of discourse. To see this, itis useful to take a small detour via Durkheim.

In The Elementary Forms of the Religious Life Durkheim (1976) famously draws adistinction between two orders of existence which relate thought and practicein two fundamentally different ways. The first order is the profane world, theeveryday world of ‘sensual representations’, the world of matter and sense,where meaning arises directly out of bodily encounters with the world, withother people, with reality. It is a world of flux and of particulars, and it is drivenby the most practical and direct wisdom: proverbs, prudence, street lore, on-the-job knowledge, the rhythmic language and wisdom of the domestic community.

The second order is the sacred (originally religious) world, one of prescrip-tions and interdicts that are not pragmatically modifiable but are ‘fixed andcrystallised’ (Durkheim 1915: 433). This sacred world is an order of verities notoriginating in bodily hexis, and therefore arbitrary, in Pierce’s sense of unmoti-vated: taboos, for example, can be attached to any object. The religious world isthus a world of arbitrary conceptual relations, a symbolic order constructed byan accretion of ‘collective representations’ (op. cit.: 434) that are a collectiveaccomplishment, the ‘work of the community’, in contrast to the ‘sensual rep-resentations’ of the everyday world that are the work of continually changingexperiential particulars shared traditionally in face-to-face encounters (seeMuller 2000).

Religion is for Durkheim the primary cognitive classificatory scheme of thesacred, the primary form of ordering social representations in non-empirical,formal ways. The force of the ordering comes from ‘outside of the object inwhich it resides’ (quoted in Thompson 1994: 125), not from the object itself. Itis the result of a process of ‘examination and elaboration’ (op. cit.: 126): it is theresult of a cognitive process of idealization, or ‘schematic idealization’ inMartin’s terms (p. 8).

Durkheim means at least two things with this faculty of idealization. The firstis the purely cognitive or speculative sense of being able ‘to connect things witheach other, to establish internal relations between them, to classify them andto systematize them’ (op. cit.: 133). The second is that of forward projectiontowards an order and a world more desirable, more felicitous, more powerful –in a word, better – than the one we have in hand at any specific point in history.

Durkheim plays upon the double sense of ideal: ideal first as the facility to

67HIERARCHY, KNOWLEDGE & THE SCHOOL CURRICULUM • MULLER

68

manipulate objects and relations in non-empirical virtual space; ideal secondlyas the projection into and towards that which is more desirable. Both togetherallow us to break with empirical facticity and to imagine an ordering of objectsthat is ‘logical’ and ‘hierarchical’ (op. cit.: 137). This is a key feature of virtualconnections that allows, as Foucault (1981: 59) says in a related idiom when dis-cussing disciplinarity, ‘the possibility of formulating new propositions, ad infini-tum’, or as Hacking (1985: 156) says when discussing styles of reasoning, to‘generate new classes of possibilities’. No surprise then that the exemplary formthat the sacred takes in societies with a complex division of labour is science.

The sacred and the profane are thus two fundamental orders present in allforms of society. The principal purpose of the strong classification between thetwo is to sustain the fundamental social cleavage between mental and manuallabour, and to reproduce it. As easy as it is for Stephen with the help ofStephen’s mother to traverse the semiotic gap (Painter 1999), as difficult is it toclose the social gap. For Bernstein, this is because of the bias in the distributiverules that regulate access to powerful discourses, to highly specialized forms ofconsciousness. The more differentiated the division of labour, the more differ-entiated will be the distribution of these sacred goods.

This may seem to damn both Bernstein and Durkheim as incorrigible pes-simists, yet, though they were both frequently derided as functionalists, theywere both anything but pessimists. To say that power arises from the social baseis not to say that power is monolithically and automatically reproduced. It issimply to say that symbolic configurations and their distributive potential areallied to socio-economic conditions. To understand how the symbolic configu-rations are reproduced is to enquire into the way these symbolic ensembles arerecontextualized, and how the recontextualized discourses are acquired. This isto enquire into the workings of symbolic control and its agencies, which is to say,the education system, which is, as we used to say, a site of struggle within whichmuch can be done. David Rose has shown this at the level of the classroom(Rose 2004), and Cuba has shown it at the level of the state, outperforming itsfellow Latin neighbours for reasons that Durkheim and Bernstein (and Rose)would have applauded, namely, because the state has realized that reducingsocial inequality across the society was an important corollary to a large-scaleequalizing of the distribution of educational competencies (Carnoy, Gove andMarshall 2004).

Returning briefly to Martin’s cline, it is worth observing that the initial cat-egories of his cline are oral transmission and written transmission. Bernsteinalmost seems to shy away from this contentious area,2 but it is noteworthy thathis exemplification of the way everyday wisdom circulates has everything to dowith its face-to-face (oral) nature (Bernstein 2000: 158). In both Martin andBernstein thus we find echoes of the dual advantages of writing, both cognitiveand social. On the one hand, as Collins (1998: 27) reminds us after Goody, Ongand Havelock, sustained writing is a ‘gateway to abstraction and generality’; onthe other, writing breaks the strictures of time/space coincidence that mark oralcultures: ‘What is needed is a social arrangement for writing texts of somelength and distributing them to readers at a distance . . .’ (op. cit.). I have beenmaking the claim in this introductory section that while both sociologists and


linguists share this dual interest, linguists concentrate on the former – how textsare specialized – while sociologists concentrate on the latter – how texts aredistributed.

Not all sociologists concentrate on distribution from the field of productionto the fields of reproduction and acquisition, as Bernstein does. It may be inter-esting to examine briefly how an exemplary sociologist of science, RobertMerton, describes the distributional consequences of discursive structure forthe discourse producers (rather than acquirers) – how disciplinary differencesaffect their practising scholars. Merton first establishes in a terse formulation hisversion of (a part of) the cline, the terms in which specialized disciplines differ:‘Codification refers to the consolidation of empirical knowledge into succinctand interdependent theoretical formulations. The various sciences and special-ities within them differ in the extent to which they are codified’ (Merton 1973:507; see also Foray and Hargreaves 2003). Degree of codification has a series ofconsequences, three of which are the following:3

• There is a higher rate of obsolescence in C+ than in C– disciplines, becausethey display a greater tendency to subsume past work. One consequence isthat there is a greater percentage of references to recent rather than olderwork in C+ than is the case in C– disciplines.

• In C+ disciplines innovative work by young scholars is more easily recognizedthan in C– disciplines, where it is easier to be overlooked, leading to whatMerton calls the Matthew Effect (Merton, op. cit.: 516), from St Matthew: tohim who hath shall be given, etc. Young scholars find it difficult to break intoC– disciplines. One consequence is that there are age differences in discov-ery patterns, summed up famously by Caius Asinius Pollio in Robert Graves’I Claudius: ‘Science is a young man’s game’ while ‘history is an old man’sgame’ (op. cit., fn. 39: 513), or Merton, with the irony for which he wasfamous: ‘This sort of thing can thus foster the illusion that good mathemati-cians die young, but that, say, good sociologists linger on forever’ (op. cit.).

• Induction into C+ entails grasping high-level propositions; into C–, intolearning masses of particulars. Induction opportunity costs consequentlydiffer.

Merton’s discussion of degrees of codification rings many bells in popular aswell as esoteric literature: recall Bertrand Russell’s distinction between ‘hard’and ‘soft’ disciplines, a distinction formalized by Tony Becher (Becher 1989; seealso Becher and Parry 2005) using a version of Bernstein to aid him. It ringsbells in the work of self-styled Bernsteinian heretic Paul Dowling (1999), whohas coined the concept of discursive saturation to replace Bernstein’s code anddiscourse cline. Dowling’s objection to code/discourse involves an objection tothe separation of the dimension of classification from that of framing. ForDowling, classification, the carrier of power and the distributive rules, is not tobe separated from its semiotic enactments (framing), thus there is only highdiscursive saturation (DS+) and low discursive saturation (DS–); the former isspecialized by generalizing strategies, the latter by localizing strategies. This is astunningly original elaboration of the theory in its pedagogic dimension, that is,


70

in establishing which pedagogic strategies belong with which kinds of DS mode,although the core distinctions are arguably latent in Bernstein’s tacit andexplicit transmission distinction. Yet in its predominant focus on texts andpedagogy, it takes us away, once again, from the field of production, from know-ledge and how it grows, from hierarchy.4

The codification cline rings bells for Bernstein and Martin too, but I believethe affinity is greater with Martin than with Bernstein, the sociological congru-ence around distribution notwithstanding. This is because both Merton andMartin are pursuing a notion of textual specialization within knowledge struc-ture across a single graded continuum, while Bernstein, as I hope to makeclearer below, seeks to find the heart of the discontinuity between the way thetwo ends of the codified spectrum grow, and progress. The rest of the chapteris an inquiry into hierarchy from the point of view of the question of knowledgeprogress.

Knowledge and the dilemma of progress5

The reluctance to speak directly about hierarchy and its cognate, progress,referred to above, is an old one, and below I will trace its roots to the terms ofa debate in the eighteenth century at the advent of the age of science, continu-ing to the present. This debate is about the idea of progress in general, and theidea of progress in knowledge in particular. We are, it would seem, exceedinglyreluctant to speak about the social dimensions of knowledge hierarchy, not onlyin terms of relations between different knowledge forms, but particularly in termsof relations within knowledge forms. The idea of hierarchy haunts us, nowheremore so than in regard to the question of knowledge progression and growth.

Bernstein has intervened decisively in the discussion about the forms ofsymbolic systems, setting out to delineate the ‘internal principles of their con-struction and their social base’ (Bernstein 2000: 155). As is by now well known,he distinguishes between two forms of discourse, horizontal and vertical, as thediscussion above made clear. From here on, this chapter will not discuss thequestion of discourses further and will concentrate on the question of variationbetween knowledge structures within vertical discourse. Here Bernstein distin-guishes between two kinds of knowledge structure, hierarchical and horizontal.

For Bernstein, knowledge structures differ in two ways. The first way is interms of what I call verticality. Verticality has to do with how theory develops. Inhierarchical knowledge structures, it develops through integration, towards evermore integrative or general propositions, the trajectory of development ofwhich lends hierarchical knowledge structures a unitary convergent shape. Hor-izontal knowledge structures, on the other hand, are not unitary but plural, con-sisting of a series of parallel incommensurable languages. Progress in horizontalknowledge structures occurs not (or at least not primarily) through theory inte-gration but rather through the introduction of a new language which constructsa ‘fresh perspective, a new set of questions, a new set of connections, and anapparently new problematic, and most importantly a new set of speakers’ (ibid.:162). Because these languages are incommensurable, they defy incorporation.


The level of integration, and the possibility for knowledge progress in the senseof greater generality and hence explanatory reach, is thus strictly pegged.

Before I proceed to discuss the second form of knowledge form variation,grammaticality, it is worth making a few observations on verticality. The firstobservation is that it artfully incorporates and recapitulates the fierce dispute inthe philosophy and sociology of science between the logical positivists and thenon-realists, a dispute I selectively revisit below. Bernstein is implicitly assertingthat the logical positivists (or realists) were right, but only in respect of hierar-chical knowledge structures; the non-realists (Kuhn and after) were likewiseright, but only in respect of horizontal knowledge structures. In other words,encoded into Bernstein’s principle of verticality are the terms of debate in thephilosophy of science since the romantic revolt of the eighteenth century.Secondly though, we should note that the category of horizontal knowledgestructures span a very broad range, from mathematics to sociology and thehumanities. Although there is more than one mathematical language, andmathematics is in this sense a ‘horizontal’ knowledge structure, this examplemakes clear that verticality is certainly possible within the discrete languagesconstituting horizontal knowledge structures, verticality of a kind approachingthe triangular form obtained in hierarchical knowledge structures, as Wignell(this volume) argues. The germane question then becomes, not so much whatconstrains progression in horizontal knowledge structures, but rather, whatinternal characteristics fail to constrain it in those that proliferate languagescompared to those where language proliferation is constrained. For Bernstein,this is the difference between ‘strong’ and ‘weak grammar’ horizontal know-ledge structures. In this usage of the term ‘grammar’, Bernstein is referring tointernal properties of the knowledge structure. This should not be confusedwith the external sense of grammaticality discussed below.

If verticality has to do with how theory develops internally, with what Bern-stein later called the internal language of description, grammaticality (in theexternal sense) has to do with how theory deals with the world, or how theoret-ical statements deal with their empirical predicates, the external language ofdescription (Bernstein 2000). The stronger the (external) grammaticality of alanguage, the more stably it is able to generate empirical correlates and themore unambiguous because more restricted the field of referents; the weaker itis, the weaker is its capacity to stably identify empirical correlates and the moreambiguous because much broader is the field of referents, thus depriving suchweak grammar knowledge structures of a principal means of generatingprogress, namely empirical disconfirmation: ‘Weak powers of empirical descrip-tions removes a crucial resource for either development or rejection of a par-ticular language and so contribute to its stability as a frozen form’ (Bernstein2000: 167–8). In other words, grammaticality determines the capacity of atheory or a language to progress through worldly corroboration; verticalitydetermines the capacity of a theory or language to progress integrativelythrough explanatory sophistication. Together, we may say that these two criteriadetermine the capacity of a particular knowledge structure to progress.

The precise nature of the relation between verticality and grammaticalityis unclear. A plausible surmise could be the following: that verticality is a


72

categorical principal, consigning knowledge structures to either a theory-integrating or a theory-proliferating category. The latter can in turn be brokendown into a constrained proliferation or an unconstrained proliferationcategory. Grammaticality on the other hand is an ordinal principal, construct-ing a continuum of grammaticality within each category of knowledge structure,or perhaps across the entire spectrum.6 In what follows, I will concentratemainly though not exclusively on verticality, on the internal characteristics ofthe internal language of description.

Why would one want to elaborate a theory of knowledge forms? After all, weseem to have got along reasonably well without one for a long time. Bernsteinonly turned to the issue towards the end of his work. The contention here is thatthis lacuna in the study of knowledge and education was not accidental. Rather,I suggest, it was produced by the failure of social thought to deal with thedilemma of progress and the distributive strictures of hierarchy. The failure toreckon with the material structural differences in knowledge forms has becomesomething of an obstacle in educational thinking. This can briefly be illustratedin two domains of education practice: namely, curriculum planning andresearch administration.

Curriculum planning has been thrust into the limelight by internationallearner performance comparisons, most vividly displayed by the Third Interna-tional Mathemetics and Science Study (TIMSS).7 A central tenet of assessmentis that the instrument measuring performance is valid to the degree that itassesses what has been made available for acquisition. TIMSS has made visiblethe fact that not all children of the same age cohort across the globe learn thesame things in the same order at the same level of cognitive demand. This hasput a spotlight on the stipulation, sequence and progression requirements ofcurricula, and has begun to suggest that not all subjects in the curriculum havethe same requirements. Could this be because their parent knowledge forms aredifferent and take different distributional forms which in turn lead to differentrecontextualizing requirements?

As for the question of research administration, research assessments ofindividuals and bodies of work have made possible comparisons betweenindividuals, faculties, universities and countries. As more and more comes todepend on assessments of innovation and novelty (‘Is this paper really a contri-bution to new knowledge, or a re-hash of the known?’), the question arises as towhat exactly constitutes innovation in different areas of research endeavour,and whether they are at all comparable. This is only the tip of the iceberg: itsoon becomes clear that there are different epistemic cultures, different kindsof collaboration, different publishing traditions, and so on. In short, theglobally emergent audit culture compels us to reflect on our knowledgepractices, at the centre of which sits the question of their likeness, their com-parability, and their compatibility. Once again, we realize how little we reallyknow about how they may be alike or different, and what difference this mightmake. At the centre of this conundrum lies the question of knowledge hierar-chy and progression.


Progress: the very idea, and its sceptics

The foundation of the Cartesian revolution in the seventeenth century was anaxiom that appeared to be radically new, namely, that ‘true’ knowledge wascharacterized by knowledge progression. As Berlin put it, that ‘which had oncebeen established did not need to be proved again, that is to say, in which scien-tific progress, universally recognised as such by rational thinkers, was possible’(Berlin 2000a: 28). The conventional account depicts this as the decisivemoment in the emergence from the closed tautological world of antiquity, andthe birth of the modern (Shapin 1996).

There are a number of entailments to this view. First, Descartes believed thatonly in a bona fide branch of knowledge can we find ‘clear and distinct ideas’(Berlin 2000a: op. cit.).

The paradigm of true knowledge, according to the Cartesian school, consisted inbeginning from truths so clear and distinct that they could be contradicted onlyon pain of falling into absurdities; and in proceeding thence, by strict deductiverules, to conclusions whose truth was guaranteed by the unbreakable rules ofdeduction . . . (ibid.)

This was indeed a lofty aim for knowledge, and it meant that Descartes viewedthe knowledge array then available in a particular way. For example, the humansciences might generate edification and improvement, but were otherwise oflittle enduring social value because they could not produce ‘strict deductiverules’. Here lies the foundation of the distinction between science and all othersymbolic ensembles, and it rests on the notion of what may be called strongprogression, that is, the step-by-step accretion of certainties.

No one today is a thorough-going Cartesian; no one today believes in strongprogression. Challenges to Cartesian rationalism have come from both withinand outside of science. One challenge to this idea of strong progression fromwithin science has culminated in the generally accepted position in sciencetoday of what may be called weak progression, or what Haack (2003) would callprogress ‘within reason’, which I will return to below. This is a revision whichaccepts the postulate of progression (and hence of the division of the field ofrepresentations into ‘true’ or progressive knowledge, and belief or mere narra-tive), but recognizing at the same time that the ‘true’ in true knowledge did notequal absolute knowledge, and that progress in knowledge, if based on the bestcertainty to hand at that time, could always and in principle be revised – hence,weak progression.

The dominant challenge to strong progression from outside science hassought to overturn the distinction between knowledge that progresses(‘science’) and knowledge that does not. The first brilliantly original formula-tion can be traced back to Giambattista Vico who, with his 7th Inaugural lecturein 1708, and later with the publication of the first edition of Scienza nuova/NewScience in 1725, rejected the fundamental premise of Cartesian rationalism, thedistinction between the true (verum) and the artificial (factum). Vico begins byarguing their essential unity: ‘We demonstrate geometry because we make it’


74

(Berlin 2000a: 31). What he meant by this was that we can be said to fully knowsomething, not only because we know what it is (i.e. through rational recon-struction) but because we know how it came to be (i.e. through historical orgenetic reconstruction), which he called per caussas. By this logic, we only knowwhat we create. If we did not create it, we cannot know it, because it then has nohuman history. ‘The true (verum) and the made (factum) are convertible’ (ibid.:35), or, ‘The criterion of truth is to have made it’ (ibid.: 36). In other words,with this argument, truth becomes a human artefact, and Vico becomes the firstconstructivist.

Thus, whereas Descartes with his criterion of ‘clear and distinct ideas’ funda-mentally sundered verum from factum, Vico fundamentally subverts it by re-uniting them. The ‘demarcation debate’ in the philosophy of science, as towhether there is or is not a significant distinction between ‘science’ and otherknowledge forms, begins here.

Vico’s revolt has come to be a mere dress rehearsal for the more thorough-going romantic revolt of the nineteenth century, and the anti-realist one of ourown time. The European romantics took up Vico’s anti-demarcatory premise ofthe make-ability of truth and of the world: . . . ‘the common assumption of theromantics that runs counter to the philosophia perennis is that the answers to thegreat questions are not to be discovered so much as to be invented. They are notsomething found; they are something literally made’ (Berlin 2000b: 202, 203).Amongst the romantics and their contemporary successors there are strong andweak traditions of make-ability. Common to all, however, is the following:

Hence that new emphasis on the subjective [the maker] and ideal rather than theobjective and the real, on the process of creation rather than its effects, onmotives rather than consequences; and, as a necessary corollary of this, on thequality of the vision, the state of mind or soul of the acting agent – purity of heart,innocence of intention, sincerity of purpose rather than getting the answer right,that is, accurate correspondence to the ‘given’. (ibid.: 203)

It is this tradition of ‘make-ability’ that I will trace below, first as it snakes its waythrough the debates on knowledge and then, in a homologous way, throughthe debates on curriculum. As we will see below, the tradition of ‘make-ability’translates into a ‘knower’ as opposed to a ‘knowledge’ perspective (Moore andMaton 2001).

On the shoulders of giants

What the scientific revolution of the seventeenth century accomplished was adecisive challenge by the self-styled Moderns to the ‘human-centred universe’(Shapin 1996: 20) of the Ancients, as consecrated by Aristotelianism, which hadbecome a hermetic dogma of a priori truth. The cultural shock should not beunderestimated. By 1611, the clerical poet John Donne could write: ‘And NewPhilosophy calls all in doubt’, ending with:


’Tis all in pieces, all coherence gone;All just supply, and all Relation.

(ibid.: 28)

A century before Vico, Donne was registering a commonly felt shock at thecultural displacement of a deeply cherished worldview. What was gone was themeasure of man, man as the measure of nature, and with it, the dominance ofthe humanistic Trivium over the scientific Quadrivium was decisively broken(Durkheim 1977; Bernstein 2000). At the heart of it was the entirely novelnotion of progression towards a not-yet-attained truth that was not determinedby man, but could be discovered by him through rational methods and intel-lectual daring. With the future and man’s fate loosed from the comfortingembrace of classical-Christian teleology, the classicists, Christians and humanistsof every stripe discerned a cultural abyss. The threat was felt to every formof social authority that depended on that worldview. The humanist counter-revolution would be only a matter of time.

In the meantime, the intoxication of the expanded temporal horizon that theidea of progress suddenly constituted can be graphically seen in drawings andsayings of the time. In the frontispiece to his 1620 Instauration magna/The GreatInstauration, an already provocative title, Francis Bacon depicted a ship boldlysailing out beyond the Pillars of Hercules, symbolizing the traditional limits ofknowledge, below which was written the biblical text from Daniel: ‘Many shallpass to and fro, and science shall be increased’ (cited by Shapin 1996: 21). Butif the forward horizon was extended, so too was the horizon backwards into thepast in like manner re-constituted.

This expanded retrospect is given iconographical expression in JohannesKepler’s tower (Figure 4.1). At the front of the tower are two modern columnsnamed for Copernicus and Tycho Brahe. Further back are more rough-hewncolumns representing the earlier knowledge of Ptolemy and Agrippa; at theback are crude columns representing ancient astronomy. Right at the front,seated, is Kepler himself: the more recent, hence, the more sophisticated. TheModerns are separated from the Ancients by two dimensions, thus. The first istime. But in order to express progress across time, a second dimension is crucialto the first, namely, greater differentiation. In order to express this, then, thefurther towers are less differentiated, the nearer ones more so. What evolves, orprogresses, is differentiation. The condition of that progression is a progressivecapitulation and building on previous knowledge, a greater differentiation ofknowledge. This quintessentially ‘modern’ idea, progressive differentiation anddual temporality, is embodied in Newton’s famous aphorism, ‘if I have seenfarther, it is by standing on the shoulders of giants’ (Merton 1993: 1).

In one of the richest ironies in the history of science, it turns out that theaphorism, everywhere attributed to Newton and hence taken as emblematic ofthe modern view of progress, does not originate with him at all. Indeed, itappears to have been common currency in knowledge circles since at leastBernard of Chartres in 1126, who probably got it from his Priscian predecessors,and was used in various forms, at regular intervals. Merton (1993: 268, 269)records at least 27 usages in print before Newton. The aphorism continues to be


76

used today in a wide variety of more or less appropriate contexts.8 The originalpoint of the aphorism was to highlight inter-generational co-operation as theground for scientific progress, usually expressed in humble or mock-humbleterms: ‘A dwarf standing on the shoulders of a giant may see farther than a gianthimself’ (ibid.: 4).

As Merton himself puts it:

When you come right down to it, the essential point is that the dwarfs-on-the-shoulders-of-giants Aphorism is a rough equivalent to the twentieth-century soci-ological conception that scientific discoveries emerge from the existing socialbase and consequently become, under conditions that can be reasonably welldefined, practically inevitable. (ibid.: 267)

What we learn from this pithy story is that scientists probably had a pragmaticview of themselves and the process of science from relatively early on – they cer-tainly didn’t have to wait for social studies of science to discover messiness and


Figure 4.1 The Astro-poecilo-pyrgium (the variegated star tower) from the titlepage of Johannes Kepler’s Tabulae Rudolphinae, 1627. (Shapin 1996: 77)

serendipity in the twentieth century – but it was only in the seventeenth centurythat the threat to social order, up until then clearly contained by the hegemonicview, provoked the first in a series of backlashes that have continued withgreater and greater sophistication to this day. The debate is variously figured,but it is probably fair to say that it has been between various kinds of realism andirrealism, those for a more or less strong view of progress in science (called byHaack 2003, the ‘Old Deferentialists’), and those against (the ‘New Cynics’).Before being sucked into the complexities of contemporary philosophy ofscience, I should immediately declare that my aim in what follows is solely to geta grip on the way that knowledge systems are internally differentiated, or aresaid to progress, and to answer the questions ‘what is it that progresses whenthey do progress? And is it only science that progresses? Has only science gothierarchy?’ The philosophy of science debate detracts from asking the questionabout the non-science disciplines. Yet the point must hold for the non-sciencedisciplines too, at least in one form or another, if these disciplines are not tosurrender all claims to progress and relevance in an information or knowledgeage (Foray and Hargreaves 2003).

Kinds of verticality

In answer to the question, ‘what is it that accumulates when knowledge pro-gresses?’ the logical positivists gave an unequivocal answer: it was the piled-upstructure of laws related to one another by strict definition, in strict order ofexplanatory integratedness, that is, in strict order of their approximation to thetruth. Here is to be found the definition of Basil Bernstein’s verticality, namely,the degree of integratedness and ‘subsume-ability’ of theory.

The logical positivist idea of progression has come under universal attack.The debate has been intricate and technical, a far cry from the ‘village ortabloid’ scapegoat of positivism (Matthews, 2004: 2) that has taken centre stagein the social sciences. The principal objection has been to the founding idea ofprogression, to the idea of a single convergent system of knowledge. There wasno single progression path, went one criticism, not least because the explan-atory reach and range of application of most covering laws was much moremodest, leading to a notion of a cluster of ‘languages’ rather than a single con-verged-upon structure, an idea conceded already by leading logical positivistOtto Neurath, he of the modest title of ‘Director of the Agency for Full SocialPlanning’ in ante bellum socialist Vienna (Cartwright 2001). Nor is it a mattersimply of ‘inductive scepticism’, of asserting the messiness of discovery againstthe seeming imperialism of the subsumptive structure: after all, the ‘Old Defer-entialists’, following Reichenbach in the early twentieth century, maintained adistinction between the ‘context of discovery’ as distinct from the ‘context ofjustification’ to deal with just this objection (Phillips 2004), though the distinc-tion proved difficult to sustain. The contemporary disavowal of progress by the‘New Cynics’ cuts far deeper, against the heart of the impersonality of this visionof progress (Popper’s ‘epistemology without a knowing subject’), against thenotion of a knowledge about the world that exists without man at its centre, in


78

its most extreme reduction, against the idea that there is a real independentworld to be known at all, a disavowal that eliminates progress by disavowing theworld and the possibility of real knowledge about it (Moore and Muller 1999).

Just as all forms of realism have built into them some or other form ofprogress, thus, all forms of the ‘New Cynicism’ have built into them the idea thatknowledge progress is incoherent. While there are many alternative accounts(meaning holism and constructive empiricism to name but two), the landmarkaccount belongs to Thomas Kuhn and his account of innovation by ‘revolution’or paradigm change (Kuhn 1962). This can be grasped by seeing that Kuhnturns the tables on the imperialism of the ‘Old Deferentialists’ who depicted allknowledge in the image of science, by depicting all knowledge, scienceincluded, as behaving like Basil Bernstein’s horizontal knowledge structures,advancing up to a point, only to break off into an alternative theoreticallanguage or paradigm. For Kuhn, as for Bernstein, the crucial point was that thelanguages were incommensurable.

One of Bernstein’s great contributions, as I suggested above, is to have re-capitulated realism’s loss of innocence, as staged in the literature of the phil-osophy of science, and recast the terms of debate into a taxonomy of knowledgeforms, with verticality, or ‘subsume-ability’, as one principal criterion. Earlier Isuggested that verticality was a categorical principal. In the discussion above itis apparent that though this criterion has been considerably weakened, thoughit is now conceded that knowledge grows by virtue of different kinds and pathsof conceptual change, these are still categorically distinguished as to whetherthey are commensurable or not. Or are there degrees of commensurability?Does incommensurability of a more absolute sort characterize what Bernsteincalled languages with weak internal grammars (and I called with ‘unconstrainedproliferation’)? Is incommensurability less absolute among ‘constrained prolif-eration’ languages? We could put this to the following test: can SFL – a languageby all accounts (including those of Bernstein, Dowling and Martin) that has astrong internal grammar, is theory-integrating, and is discursively saturated –commune with, say, other forms of sociolinguistics? Or is that conversation ofthe deaf as absolute as anything in the social sciences? I suspect that, in com-parison with the hapless sociology, there is greater seepage between languagesin linguistics than in sociology.

The discussion may be summed up as follows. As far as the internal structureof internal languages of description is concerned, that is, as regards their verti-cality, languages vary as to their:

• capacity to subsume statements into logical types (syntactic/semantic axis);• therefore their relative expressibility in terms of general and particular

statements (general/particular axis);• and therefore their relative expressibility in terms of propositional content

and stylistic content (content/form axis).

One could go on. The question now for this paper is: how are different powersof subsumption, of verticality, expressed by pedagogy? Does hierarchy specifypedagogy, and if so how?


The knowledge structure/pedagogic structure link

Does knowledge structure constrain pedagogic structure, does it place any onuson the way that the ‘what is to be learnt’ is recontextualized? Do these internalcharacteristics of knowledge structures place limits on the form their curricularoffspring optimally could and should take? We know that pedagogic structurehas distributive potential, but does knowledge structure come with an alreadyencoded distributive potential, placing structural limits on pedagogic form? Weknow that it takes a specialized language to specialize consciousness: but can wedetermine what kind of specialization a pedagogy must encode to effectivelyrealize a specialized consciousness in a specialized language? All these differentways of posing the question presuppose answering the question one way oranother.

One affirmative response to these questions is given in the Review Commit-tee’s (2000) report on a review of the South African grade 1 to 9 national cur-riculum, known as Curriculum 2005 (C2005). The Review found that thecurricular form of C2005 was under-stipulated, under-sequenced and its pacingrequirements under-signalled. It was a form of invisible or competencepedagogy (Bernstein, 2000) which provided minimal markers as to what shouldbe learnt or evaluated at what level. Unsurprisingly, the Review found thatpoorly trained teachers, teaching largely disadvantaged children, fared worstwith this pedagogy: they had not covered what the curriculum expected them tohave covered by the end of each grade. Consequently children entered the nextgrade with knowledge gaps, elements of knowledge presupposed by the cur-riculum of the next grade. These knowledge gaps had more serious conse-quences in what the Review called ‘content-rich’ subjects (maths, science andlanguage) than in more skills-based subjects like life-skills, because the formerhad content, sequence and progression requirements deriving from theirparent disciplines.

How does Bernstein answer the question? Equivocally, it turns out. In thepaper on the pedagogic device, the answer is negative:

As physics is appropriated by the recontextualising agents, the results cannotformally be derived from the logic of that discourse. Irrespective of the intrinsiclogic which constitutes the specialised discourse and activities called physics, therecontextualising agents will select from the totality of practices which is calledphysics . . . But these selections cannot be derived from the logic of the discourse of physics. . . (Bernstein 2000: 34, emphasis added)

Here Bernstein appears to contradict the Review, but we should note thecontext, where Bernstein is asserting the theoretical priority of the regulativeover the instructional, meaning that the internal order of school physics iswholly derived from normative social order. In one sense this is undoubtedlycorrect. Any state can, on the basis of its ideology, decide what pedagogicmodality to impose. It can even appear to choose a de-specializing pedagogy indefiance or denial of the requirements of its specialized division of labour, atleast for a while. This is indeed what happened in South Africa with C2005 in


80

the 1990s, as it had in the UK with the Plowden Report in the late 1960s (Fitz,Davies and Evans 2006). But in another sense, this is quite misleading. Indeed,the argument can be made that Bernstein came to the view, late in his career,that the instructional domain – or knowledge – has an internal determinativelogic of its own, one which cannot simply be reduced to subordination to theregulative order. Although the sea change was presaged in the ‘pedagogicdevice’ paper, the strongest evidence for this view can be found in Bernstein’s‘vertical and horizontal discourse’ paper (Bernstein 2000).

The principal reason Bernstein opposed the idea of a link in his early work isbecause he maintained that the recontextualized discourse (e.g. school physics)was a wholly separate discourse from that of research-based physics. It was nec-essary to maintain this position in order to stay true to the postulate that allsymbolic formations were specific to a context with its specializing practices.That context is conditioned by a society’s regulative or moral order. But if recon-textualization totally severs any relation, then how are specialized knowledgesever reproduced? After all, school maths performance predicts (imperfectly tobe sure) university maths performance; and that predicts in turn proper mathsadeptness. The only way this can be intelligible is by conceiving that schoolmaths competence ‘precurses’ (Gee 2001) university maths competence, which‘precurses’ real maths adeptness. There has to be some form of specialization ofconsciousness continuum in play; this could be called a founding assumption ofmodern education, and it is strongly suggested by Bernstein’s view that thesymbolic device is ‘condensed’ in the evaluative rules which binds the logic ofthe parent discipline to the attainment requirements for acquirers (Bernstein2000). After all, this idea of the interpenetration of symbolic competence isbuilt into Bernstein’s explanation of how the middle-class home code precursesits young into the school code better than does the working-class home code.So, a relation there must be. One might pursue the exact nature of the relation.A preliminary question for this chapter is: what effects the relation? Whatactivates it?

There are two typical answers: let us call them a knowledge answer, and a knoweranswer. A knowledge approach is epitomized by TIMSS and their principal intel-lectual construct, ‘opportunity to learn’ (OTL). OTL in its simplest form isdefined as coverage of the curriculum, and the original TIMSS project definedcoverage in terms of a serial list of topics only (Porter 2002). But is that suffi-cient to map coverage in a vertical curriculum? It certainly wasn’t adequate inTIMSS. A syntactical view will say: list the principal propositional steps in theknowledge hierarchy. Each propositional step will function like a rule with rulesof combination, each cluster of which can generate an indefinite number ofpossibilities; for example, ‘odd numbers’ in arithmetic (see Pinker 1999: 318).A complete list will describe the internal grammar of the internal language ofthe subject. Table 4.1 shows one way of representing this, drawn from Reevesand Muller (2005).

This table displays an extract from a finite list of maths content presumed tobe learnt in grade 6, listed in order of cognitive complexity. The information inbrackets (g4, g5, g6) indicates that these knowledge elements are consideredessential at grade 4, 5 or 6 levels – in other words, they reflect work that learners


are, at a minimum, expected to cover at this level. However, although certainelements of topics or subtopics are considered essential for a particular gradelevel, there are other elements of topics or subtopics that are considered essen-tial at all or more than one grade level.

But here a pedagogical question arises. Does a list of topics in order of com-plexity, in other words in order of disciplinary progression, constitute anoptimal learning path? The reconstituted logic of a discipline and the optimalpedagogical learning sequence might overlap only by default. The reason forthis lies with the way subsumption works: the same semantic topics (the sameparticulars) play different roles in different generals. The upshot is that partic-ular topics, even for the most hierarchical of subjects, are repeated acrosslearning levels, but differently. In short, imperfect subsumption has so far stymiedthe linear representation of content in a curriculum, and the relation of curric-ular structure to disciplinary structure has remained an open, more usually an


Representing and comparing whole numbers including zero and fractions including:

Whole numbers to

• 4-digit numbers (g4)



Odd and even numbers to 1,000 (g4)

Common fractions in diagrammatic form (g4)

Common fractions with different denominators including

• halves (g4)

• thirds (g4)

• quarters (g4)

• fifths (g4)

• sixths (g4)

• sevenths (g4)

• eighths (g4)

• tenths (g6)

• twelfths (g5, 6)

• hundreds (g6)

Table 4.1 Extract of hierarchy of maths content for grade 6

82

avoided, question. The same question may be posed to the linguists: does a cur-riculum following a genre-sequential logic constitute an optimal learning path?How are these different cognitive logics to be braided into the artifice calledcurriculum and pedagogy? This is the nub of pedagogy.

All in all, it is not too surprising that by far the most common way of repre-senting verticality in the literature has been distilled from what teachers do orexpect, in terms of an index of ‘cognitive demand’, usually depicted as a scalefrom ‘memorization’, through ‘routine procedures’, ‘communication of under-standing’, ‘problem solving’, to ‘conjecture/generalise/prove’ (Porter 2002: 4).Each listed topic is given a ‘cognitive demand’ rating. Proportion of coverage bydegree of demand yields a proportion of instructional time spent. Thistemporal proportion becomes a proxy for ‘opportunity to learn verticality’.There are clearly other ways of compiling a demand index (see Morais et al.2004, for a good example). All of them, as far as I can see, shift the focus fromthe knowledge itself to the knower-actors, from a knowledge approach to aknower approach. They shift the focus from what knowledge is made available, towhat levels of complexity teachers make available. Hierarchy moves back into theshadows.

In the wake of the demise of logical positivism, and the discrediting of the dis-tinction between the contexts of discovery and justification, a new orthodoxyunder the aegis of the ‘New Cynics’ and their pedagogical confrères the educa-tional progressives, took hold, to the effect that sequence and progression inpedagogy simply didn’t matter, that there was no one royal road to learning, andthat only the most minimal stipulations were necessary to signal the learningend points as measured by common assessment instruments (Muller 2002;Labaree 2004). This turned out to be wrong, in South Africa and elsewhere.Such radically invisible pedagogies can work, but only for middle-class pupilsand usually only in the lower grades, and exactly how they do that is still beingunearthed. For the majority of poorer children, the evidence increasinglysuggests, clear and explicit articulation of evaluation criteria is sine qua non (seethe various papers in Muller et al. (eds) 2004).

What are the evaluation criteria evaluations of? Of the knowledge steps to betraversed; it is hard to avoid this conclusion. To be sure, there is not only oneset of steps per discipline, nor need we assume that these steps are always to betraversed in exactly the same order: in practical situations they simply won’t be.Nevertheless, insofar as the idea of theory integration means anything at all, itdoes, qua hierarchy, specify the formal, minimal steps to be acquired in orderfor sense to be made at all. So, making concessions to messiness and agreeingthat we cannot stipulate a once-and-for-all-path, we would still have to concede,retrospectively considered, that there are a specifiable necessary minimum setof steps that must be pedagogically traversed. In a world of disadvantage, to losesight of this is to lose sight of the target for the social equality of outcomes.

Does this argument do away with the necessity of the teacher? Not at all. Whatit does is to emphasize the knowledge dimension of what makes a teacher ateacher. The condition for a teacher being an authoritative pedagogical agentis, at the minimum, an internalized map of the conceptual structure of thesubject, acquired through disciplinary training (this is perhaps why Morais et al.


2004, dub the index of the what ‘teacher competence’ in their study). In otherwords, the condition for teachers to be able to induct pupils into strong internalgrammar subjects is that they themselves already stand on the shoulders ofgiants, that they can speak with the disciplinary grammar. But if they can’t? I aminclined to say: let us then train them so that they can. The difficulty is, it is oftennot clear what they don’t know. Two things stand out in the global literature oneffective learning. The first is that teacher competence is by far the most impor-tant factor in learner attainment; the second is that in-service teacher traininghas had almost no effect. I fear this will continue unless we pay as much atten-tion in future to knowledge as we have up until now to knowers.

Conclusion

I have proceeded in this chapter on the hunch that the contemporary avoid-ance of knowledge structure, in this chapter principally of the question of hier-archy and progression, lies at the heart of many current pedagogical dilemmas,in particular those having to do with providing to poor children access to thetools of powerful knowledge. That this can be done has been incontrovertiblyshown by Rose (2004) and Carnoy et al. (2004). I have not kept that in the fore-ground of the discussion, but it nevertheless remains a principal motive force.It remains plausible, perhaps even likely, that knowledge structure has distribu-tional implications, which interact with distributional alignments of pedagogy.It has proved easier discussing verticality in hierarchical disciplines than inthose with horizontal knowledge structures. I have suggested, nudging Bern-stein and Martin, that not all horizontal structures have languages that areequally incommensurable. I have suggested that some, like linguistics perhaps,have languages whose boundaries are more permeable than those, like sociol-ogy perhaps, that are inclined to proliferate languages without constraint. Con-straint must at least in part be promoted by hierarchy. This remains a task to becontinued, not least because it seems likely to me that the exceptionalism thatthe social sciences and humanities have claimed for so long will prove increas-ingly implausible in the global networks of the knowledge society. When thathappens, I fear that the ‘New Cynicism’ and its pedagogical analogues will beabout as effective a cloak against the cold winds of global comparability as fash-ionable decadence was for the artistic fringe in ante bellum Berlin.

References

Becher, T. (1989), Academic Tribes and Territories: Intellectual Enquiry and the Cultures ofDisciplines. Buckinghamshire: SRHE and Open University Press.

Becher, T. and Parry, S. (2005), ‘The endurance of the disciplines’, in I. Bleiklie andM. Henkel (eds), Governing Knowledge. Dordrecht: Springer.

Berlin, I. (2000a), Three Critics of the Enlightenment: Vico, Hammann, Herder. Princeton:Princeton University Press.

Berlin, I. (2000b), ‘The essence of European romanticism’, in I. Berlin, The Power of Ideas.London: Pimlico.


84

Bernstein, B. (1994), ‘Discourse, knowledge structure and fields: some arbitrary consid-erations’. Mimeo.

Bernstein, B. (1996), Pedagogy, Symbolic Control and Identity: Theory, Research, Critique.London: Taylor & Francis.

Bernstein, B. (2000), Pedagogy, Symbolic Control and Identity: Theory, Research, Critique.Oxford: Rowman & Litttlefield.

Carnoy, M., Gove, A. and Marshall, T. (2004), Why do some Students Achieve More in someCountries than in Others? A Comparative Study of Brazil, Chile and Cuba, in press.

Cartwright, N. (2001), The Dappled World: A Study of the Boundaries of Science. Cambridge:Cambridge University Press.

Collins, R. (1998), The Sociology of Philosophies: A Global Theory of Intellectual Change.Cambridge, MA: The Bellknap Press of Harvard University Press.

Dowling, P. (1999), ‘Basil Bernstein in frame: “Oh dear, is this a structuralist analysis?”’http://www.ioe.ac.uk/ccs/dowling/kings1999/index.html

Dowling, P. (2005), ‘A timely utterance’. http://homepage.mac.com/paulcdowling/ioe/publications/dowling2005/timely_utterance/index.htm

Durkheim, E. (1976: first pub. 1915), The Elementary Forms of Religious Life: A Study in Reli-gious Sociology. Trans. J. W. Swain. New York: The Free Press.

Durkheim, E. (1977: first pub. 1938), The Evolution of Educational Thought: Lectures on theFormation and Development of Secondary Education in France. Trans. P. Collins. London:Routledge & Kegan Paul.

Fitz, J., Davies, B. and Evans, J. (2006), Educational Policy and Social Reproduction: ClassInscription and Symbolic Control. London: Routledge.

Foray, D. and Hargreaves, D. (2003), ‘The production of knowledge in different sectors’,in London Review of Education 1 (1), pp. 7–19.

Foucault, M. (1981), ‘The order of discourse’, in R. Young (ed.), Untying the Text. London:Routledge & Kegan Paul.

Gee, J. (2001), ‘Learning in semiotic domains: a social and situated account’, paper pre-sented to the International Literacy Conference, Cape Town, 13–17 November, 2001.

Goldman, A. (1999), Knowledge in a Social World. Oxford: Oxford University Press.Haack, S. (2003), Defending Science – Within Reason. New York: Prometheus Books.Hacking, I. (1985), ‘Styles of scientific reasoning’, in J. Rajchman and C. West (eds), Post-

Analytic Philosophy. New York: Columbia University Press.Hasan, R. (2001), ‘The ontogenesis of decontextualised language: some achievements of

classification and framing’, in A. Morais, I. Neves, B. Davies and H. Daniels (eds),Towards a Sociology of Pedagogy. New York: Peter Lang.

Kuhn, T. (1962), The Structure of Scientific Revolutions. Chicago: Chicago University Press.Labaree, D. (2004), The Trouble with Ed Schools. New Haven: Yale University Press.Matthews, M.R. (2004), ‘Editorial’, Science and Education, 13, pp. 1–4.Merton, R.K. (1973), The Sociology of Science: Theoretical and Empirical Investigations.

Chicago: The University of Chicago Press.Merton, R.K. (1993: first pub. 1965), On the Shoulders of Giants. Chicago: The University of

Chicago Press.Moore, R. and Maton, K. (2001), ‘Founding the sociology of knowledge: Basil Bernstein,

epistemic fields and the epistemic device’, in A. Morais, I. Neves, B. Davies andH. Daniels (eds), Towards a Sociology of Pedagogy: The Contribution of Basil Bernstein toResearch. New York: Peter Lang.

Moore, R. and Muller, J. (1999), ‘The discourse of “voice” and the problem of knowledgeand identity in the sociology of education’, in British Journal of Sociology of Education20 (2), pp. 189–206.

Moore, R. and Muller, J. (2002), ‘The growth of knowledge and the discursive gap’, inBritish Journal of Sociology of Education 23 (4), pp. 627–37.

Morais, A., Neves, I. and Pires, D. (2004), ‘The what and the how of teaching and


http://www.ioe.ac.uk/ccs/dowling/kings1999/index.html

http://homepage.mac.com/paulcdowling/ioe/publications/dowling2005/timely_utterance/index.htm

http://homepage.mac.com/paulcdowling/ioe/publications/dowling2005/timely_utterance/index.htm

learning’, in J. Muller, B. Davies and A. Morais (eds), Reading Bernstein, ResearchingBernstein. London: RoutledgeFalmer.

Muller, J. (2000), Reclaiming Knowledge: Social Theory, Curriculum and Education Policy.London: RoutledgeFalmer.

Muller, J. (2002), ‘Progressivism redux: ethos, policy, pathos’, in A. Kraak and M. Young(eds), Education in Retrospect: Policy and Implementation Since 1990. Pretoria andLondon: HSRC Publishers and Institute of Education, University of London.

Muller, J., Davies, B. and Morais, A. (eds) (2004), Reading Bernstein, Researching Bernstein.London: RoutledgeFalmer.

Mullis, I., Martin, M., Gonzales, E., Gregory, K., Garden, R., O’Connor, K., Chrostowski,S. and Smith, T. (2000), TIMSS 1999 International Mathematics Report: Findings fromIEA’s Repeat of the Third International Mathematics and Science Study at the Eighth Grade.Chestnut Hill: Boston College.

Nagel, E. (1961), The Structure of Science: Problems in the Logic of Scientific Explanation.London: Routledge & Kegan Paul.

Painter, C. (1999), ‘Preparing for school: developing a semantic style for educationalknowledge’, in F. Christie (ed.), Pedagogy and the Shaping of Consciousness. London:Continuum.

Phillips, D.C. (2004), ‘Two decades after “After the wake: postpositivistic educationalthought”’, Science and Education 13, pp. 67–84.

Popper, K. (2002: first pub. 1963), Conjectures and Refutations: The Growth of Scientific Knowl-edge. London: Routledge Classics.

Porter, A. (2002), ‘Measuring the content of instruction: uses in research and practice’,in Educational Researcher 31 (7), pp. 3–14.

Pinker, S. (1999), Words and Rules: The Ingredients of Language. Cambridge: CambridgeUniversity Press.

Reeves, C. and Muller, J. (2005), ‘Picking up the pace: variation in the structure andorganisation of learning school mathematics’, in Journal of Education 37, pp. 103–130.

Rose, D. (2004), ‘Sequencing and pacing of the hidden curriculum: how Indigenouslearners are left out of the chain’, in J. Muller, B. Davies and A. Morais (eds), ReadingBernstein, Researching Bernstein. London: RoutledgeFalmer.

Shapin, S. (1996), The Scientific Revolution. Chicago: The University of Chicago Press.Thompson, K. (ed.) (1994), Readings from Emile Durkheim. London: Routledge & Kegan

Paul.Weinberg, S. (1993: first pub. 1992), Dreams of a Final Theory. New York: Vintage.

Notes

1 The first record in print of Bernstein’s tree is in a mimeo from March 1994. If Martinpublished his in 1992, as the text referred to above implies, then he is indeed respon-sible for a critical breakthrough.

2 Wariness of the written-oral cline is explicit in the early forms of the ‘discourses’paper (see Bernstein 1994), implicit in the later ones (see the version in Bernstein2000).

3 Highly codified disciplines are labelled C+; weakly codified disciplines C–.4 This does not stop him from commenting on disciplines, and although he professes

relative ignorance, Dowling (2005: 6) is unusually generous about SFL: ‘SystemicFunctional Linguistics, by contrast (again with the hapless sociology), is now wellestablished, which is to say, strongly institutionalized, internationally and is clearlyDS+ as is readily apparent from even an amateur’s perusal of Halliday and Math-iessen (2004). We may legitimately refer to SFL as a discourse in its own right . . .’


86

Oddly enough he seems unaware of Kay O’Halloran’s work (see this volume), anSF linguist not only working in his own area, maths education, but who uses a term,symbolic saturation, close in spirit to his own.

5 Sections of what follows will be published as ‘On the shoulders of giants: verticalityof knowledge and the school curriculum’, in R. Moore, M. Arnot, J. Beck andH. Daniels (eds), Knowledge, Power and Educational Reform: Applying the Sociology of BasilBernstein. London: RoutledgeFalmer 2006.

6 Though Bernstein depicts grammaticality as a feature only of horizontal knowledgestructures in the diagram on p. 168, he refers to physics as having a ‘strong grammar’on p. 163. He is probably using it here in its internal sense of ‘explicit conceptualsyntax’.

7 TIMSS and the TIMSS Repeat studies are synonymous with international learnerattainment comparisons, having generated comparable data from 38 countries forMaths and Science, and leading to the first international league tables (see Mulliset al. 2000).

8 The cheekiest is its use as the heading on the opening page of the Google Scholarsearch engine.


For every knowledge structure there is also a knower structure.

Introduction

Why bother reading this chapter? On what grounds am I claiming insight orunderstanding? How can I claim to be a ‘sociologist’? These are the kinds ofquestions I shall focus on: the basis of achievement, status and membership;i.e. the issue of how knowledge and knowers are specialized. The work of BasilBernstein is particularly valuable for addressing such questions. Where mostapproaches in the sociology of knowledge and education focus on relations toknowledge (of class, race, gender and so forth), Bernstein’s approach paysattention to relations within knowledge. Instead of simply showing how identityshapes knowledge, this approach also reveals how knowledge itself specializesidentity, consciousness and relations. In this chapter I explore how Bernstein’sconceptual framework sheds light on these issues and how his ideas can bedeveloped to create further insights.1

The reasons for developing the approach further are immanent in the formof its development. One trajectory that can be traced through Basil Bernstein’ssociology is from the analysis of the pedagogic practices of educational fields ofreproduction (1977), through an account of the construction of educationalknowledge (1990), to the study of the intellectual fields of production fromwhich this knowledge is selected and recontextualized (1999). With theconcepts of educational knowledge codes, the pedagogic device and knowledgestructures, respectively, Bernstein showed how structurings of intellectual andeducational knowledge specialize actors and discourses in ways that shape socialrelations, institutional organization, disciplinary and curricular change,identity, consciousness and habitus (Singh 2002; Moore 2004). This trajectoryhas been characterized by an unusually intimate dialectic between theory andresearch (Bernstein 1996). Thanks to the form taken by the theory (Moore andMuller 2002), in each case its development (in the light of previous research)created new ways of seeing existing objects of study and highlighted new objectsfor empirical research to explore. In turn, subsequent research spoke back tothe theory, raising new questions and necessitating further conceptual develop-ment. Bernstein was fond of saying one should aim for ‘productive imperfec-tion’ and stated that his concepts represented a provisional mapping ofintellectual fields. One can thereby rest assured that the notion of ‘knowledgestructures’ is not the end of the matter. In this chapter I elaborate on a meansof continuing this neverending story.

5 Knowledge–knower structures in intellectual andeducational fields

Karl Maton

88

Specifically, I address two questions raised by the concepts of knowledgestructures (see Maton and Muller, this volume). Bernstein’s language of descrip-tion provides a means of systematically describing differences between the dis-cursive practices of intellectual fields, raising the question of conceptualizingthe generative principles underlying these fields of production. This formed astarting point for the development of legitimation codes (Maton 2000a, 2000b)and the epistemic device (Moore and Maton 2001). Legitimation codes provide ameans of conceptualizing the structuring principles underlying intellectualfields; the epistemic device is the means whereby these codes (and so the formtaken by intellectual fields) are created, maintained, reproduced, transformedand changed. The epistemic device was intended to complement rather thandisplace the pedagogic device; rather than being specific to intellectual andeducational fields respectively, both devices form the basis for production, recon-textualization and reproduction of knowledge. Though developed throughstudies of knowledge production, the concepts were thereby intended to illu-minate educational knowledge and practice more generally, and in a number ofstudies are being used to analyse educational fields.2 However, in terms of Bern-stein’s theory, a second question remains of relations between the concepts ofknowledge structures and educational knowledge codes or, put another way,how intellectual fields of production and educational fields of reproduction canbe analysed within the same conceptual framework.3

I shall argue that the concepts of legitimation codes and the epistemic deviceprovide a means of addressing these questions, and that their answers are to befound by thinking in terms of Knowledge–knower structures. My basic argument issummarized in the opening motif: for every knowledge structure there is also aknower structure. I elaborate the implications of this claim in two stages. First, Iintroduce the concepts of knowledge structures and knower structures andshow how they can be brought together and their underlying structuring prin-ciples analysed. I do so through considering fields of knowledge production,focusing on the example of the famous ‘two cultures’ debate on relationsbetween science and the humanities. Secondly, I explore how these conceptsrelate to educational knowledge codes, elaborate on the forms taken byknowledge–knower structures, and show how the concepts can be applied tofields of reproduction, focusing on the example of studies addressing themarginalized position of Music in the English school curriculum.

Knowledge–knower structures in fields of production

A useful way of introducing the notion of knowledge–knower structures is toconsider the well-known ‘two cultures’ debate. This was sparked by C.P. Snow’sfamous 1959 lecture in which he claimed the intellectual life of western societywas being split into ‘two polar groups’ that ‘had almost ceased to communicateat all’ with ‘between the two a gulf of mutual incomprehension – sometimes. . . hostility and dislike, but most of all lack of understanding’. (Snow 1959: 3,2, 4). Snow’s focus lay beyond the academy, but the ‘two cultures’ became asso-ciated with the humanities and science and the ensuing debate was constructed


as a struggle over status within higher education. Though this picture of twocultures was already well established, the debate was ferocious, bitter, spreadwidely, and remains a source of contention. Why Snow’s lecture sparked suchdepth of feeling is clear from the ways in which participants portrayed scienceand the humanities as enjoying contrasting fortunes. On the one hand, whatSnow termed ‘scientific culture’ was portrayed as enjoying a meteoric rise instature; as one commentator tartly expressed:

You cannot open a newspaper, let alone the ‘quality’ journals, without the impor-tance of science and technology being trumpeted at you from the headlines.(Morris 1959: 374)

Feted by and enjoying massive funding from industry and the state, reveredby the media and worshipped by the public, by the late 1950s scientists were saidto be enjoying unprecedented prestige. In contrast, the humanities wereportrayed as embattled, in decline and insecure. For example, an influentialcollection of essays entitled Crisis in the Humanities (Plumb 1964a) includedaccounts of proclaimed crises within Classics, history, philosophy, Divinity,literary education, sociology, the fine arts, and economics, as well as the human-ities in schools. They were said to be unwanted by better quality students, con-sidered irrelevant to a modern economy by industrialists, increasingly excludedfrom the corridors of power by politicians, no longer considered the repositoryof culture, and publicly ridiculed as offering little genuine knowledge. One his-torian, for example, claimed that ninety per cent of his colleagues believed theirsubject to be ‘meaningless in any ultimate sense’ (Plumb 1964b: 25).

According to participants, the disciplinary map was undergoing a fundamen-tal shift of power between humanist and scientific cultures in their long-acknowledged struggle for status and resources. Two questions this raises are:what was the basis of their differences, and why was this shift of power occur-ring? A common contemporary explanation of their differences held thatscientists and humanist intellectuals ‘speak different languages’ (Editorial, TheListener, 3 Sept. 1959: p. 344). Using Bernstein’s approach would suggestfocusing instead on the underlying structuring principles of their languages. I shallexplore these principles in terms of knowledge structures and then knowerstructures, before bringing them together to show how an analysis ofknowledge–knower structures can shed light on the bases of intellectual fields.

Knowledge structuresAnalysing the form taken by knowledge in intellectual fields, Bernstein (1996,1999) distinguishes first between horizontal discourse (everyday or ‘common-sense’ knowledge) and vertical discourse (scholarly or professional knowledge),and secondly within vertical discourse between horizontal and hierarchicalknowledge structures. The latter exemplify well one aspect of the way the twocultures were portrayed in the debate. Humanist culture was described byproponents as riven by competing claims for status between strongly boundeddisciplines. Commentators argued that Classics had served as the basis of a‘common culture’ or ‘unifying force’ (Lee 1955) and their decline had

89KNOWLEDGE–KNOWER STRUCTURES • MATON

90

fragmented a single, organic culture into a series of rival subcultures, with littledialogue between disciplines and no means of adjudicating between competingclaims to be a new unifying centre. Humanist culture thereby resembled whatBernstein defines as a horizontal knowledge structure:

a series of specialised languages, each with its own specialised modes of interro-gation and specialised criteria . . . with non-comparable principles of descriptionbased on different, often opposed, assumptions. (Bernstein 1996: 172–3)

This comprises a series of segmented, strongly bounded languages which, devel-oping Bernstein (1999: 162), can be visually represented as:

Bernstein further distinguished between horizontal knowledge structureswith stronger grammars, ‘whose languages have an explicit conceptual syntaxcapable of “relatively” precise empirical descriptions and/or of generatingformal modelling of empirical relations’ (1999: 164), such as mathematics, lin-guistics and economics, and those where these powers are weaker, such asanthropology, cultural studies and sociology. Humanist culture, as characterizedby proponents, possessed a weaker grammar – its objects of and procedures forstudy were defined in ethereal, nebulous, even mystical terms, most famouslyand widely expressed as immersion in ‘the best that has been known andthought in the world’.

In contrast to the segmentation of humanist culture, proponents of scientificculture claimed that scientists comprised an organic community; as Snow put it,they shared ‘common attitudes, common standards and patterns of behaviour,common approaches and assumptions’ (1959: 9). Unlike the pluralized human-ities, science was often referred to in the singular and portrayed as integratedand whole, and though scientists were proliferating new knowledge andcreating sub-disciplinary specialisms at a prolific rate, they were said to knowhow to integrate this knowledge. Scientific culture thereby resembled whatBernstein describes as a hierarchical knowledge structure: ‘an explicit, coherent,systematically principled and hierarchical organisation of knowledge’ whichdevelops through the integration of knowledge at lower levels and across anexpanding range of phenomena (1996: 172–3). This Bernstein represents as:

where the point of the pyramid represents the smallest number of axioms ortheories and the base represents the maximal number of empirical phenomenaexplainable by these propositions.


L1 L2 L3 L4 L5

Knower structuresBernstein’s concepts enable the form taken by the knowledge structurescharacterizing the two cultures to be described. To reach an understanding oftheir underlying principles, I shall now turn to consider their knower structures.These reveal a different picture (see Figure 5.1). Taking humanist culture first,I described how the humanities were portrayed as having been a ‘commonculture’ underpinned by the Classics. However, it was not Classics as skills, tech-niques and procedures that integrated the humanities into a culture, but ratherthe dispositions that a classical education was thought to guarantee. The idealhumanist intellectual was a gentleman amateur who pursued (usually) hisstudies ‘for the love of it’, viewing them as secondary to a clerisy role of culti-vating the cultured sensibility of the ‘English gentleman’ among studentsselected on the basis of fitting the character of the university (Maton 2004). Thehumanities were said to humanize; underpinning such claims was an image ofwhat it meant to be human – the sensibilities, character and personal attributesof an ideal knower. The basis of specialization in humanist culture was thus notknowledge (indeed, disciplinary specialization was strongly devalorized infavour of the all-round ‘generalist’) but the habitus of an ideal knower, and aclassical education served as shorthand for these dispositions. This culturalfocus was, moreover, a veneer for a tacit social hierarchy. To be educated in theClassics was (in the main) to have enjoyed a particular social and educationaltrajectory – typically male, higher social class, private school, ‘Oxbridge’ –against which other knowers were (tacitly) measured. In other words, humanistculture exhibited what I shall term a hierarchical knower structure: a systematicallyprincipled and hierarchical organisation of knowers based on the image of anideal knower which develops through the integration of new knowers at lowerlevels and across an expanding range of different (innate and/or social) dispo-sitions.4

This can be portrayed as a pyramid of knowers (Figure 5.1) with, in the caseof humanist culture, the ideal of the ‘English gentleman’ at its pinnacle. Thebasis of the recontextualizing principle of humanist culture and its ruler (inboth senses of measuring and dominating) was thus an idealized knower.

We can here further distinguish between those hierarchical knower structureswith (by way of analogy to Bernstein’s terms) stronger ‘knower-grammars’,where the biological and/or social bases of the ideal knower(s) are articulatedrelatively explicitly (such as the Great Chain of Being underpinning papal andmonarchical hierarchies and, more recently, various standpoint theories), andthose with weaker ‘knower-grammars’, where the biological/social bases ofknowers are more tacit and, as Bourdieu would put it (Bourdieu and Passeron1990), ‘misrecognized’ as cultural (as was the case for the ‘English gentle-man’).5

In terms of knowers, scientific culture was portrayed differently. Where thehumanist intellectual’s ‘ability is a personal matter, which on the whole he doesnot owe to his advanced training’, scientific knowledge was widely said to be‘fairly independent of the personal merits of its possessor’ (Gellner 1964: 75–6).Snow, for example, compared the class-bound basis of humanist culture to thedemocratic and meritocratic nature of scientific culture. He claimed science was


92

blind to colour, race, creed; it cut ‘across other mental patterns, such as thoseof religion or politics or class’ (1959: 9). In short, the basis ofspecialization in science was knowledge of scientific procedures, regardless ofbiological or social background. Science was thus portrayed as a horizontal knowerstructure: a series of strongly bounded knowers, each with its own specializedmodes of being and acting, with non-comparable habituses or embodied dispo-sitions based on different biological and /or social backgrounds and histories.

In terms of their dispositions, scientists could represent a series of segmentedknowers (Figure 5.1), each strongly bounded from one another in terms oftheir (non-scientific) ‘gaze’ and capable of being based on very different, evenopposed, assumptions.

Exploring knower structures highlights something not immediately obviousfrom studying knowledge structures alone: as illustrated in Figure 5.1, it is notonly hierarchical knowledge structures that are characterized by a hierarchy. Thelocation of the ‘hierarchical’ in an intellectual field could be described as thesite or basis of its recontextualizing principle. Hierarchical knower structuresthereby also possess a systematic principle for selecting and arranging actorsand discourses into a hierarchy. That is to say, within intellectual fields actorsand discourses are selected and recontextualized into positions within the fieldon the basis of a principle emanating from the knowledge structure, knowerstructure or, as I shall discuss, neither or both. In the case of the portrayal of thetwo cultures, the basis of their recontextualizing principles lay in the knowledgestructure for science and in the knower structure for humanities. The differ-ence between intellectual fields may thus be less whether they are hierarchical ornot and more where their hierarchizing principle lies.


Humanist culture

(horizontal)

(horizontal)

(hierarchical)

(hierarchical)

Knowledgestructures

Knowerstructures

Scientific culture

Figure 5.1 The two cultures as knowledge structures and knower structures

Analysing knowledge–knower structuresHaving described differences between the two cultures in terms of their know-ledge structures and knower structures, we are now in a position to analyse thestructuring principles underlying fields of production. If we understand the dis-cursive practices of intellectual fields as knowledge–knower structures that special-ize actors and discourses in different ways, then the principles underlying thesepractices can be addressed in terms of their legitimation codes of specialization.This notion is based on the simple idea that actors and discourses are not onlypositioned in both a structure of knowledge and in a structure of knowers, butalso establish different forms of relations to these two structures. One canthereby analytically distinguish between an epistemic relation (ER) to the know-ledge structure and a social relation (SR) to the knower structure. Each of theserelations can exhibit relatively stronger (+) or weaker (–) classification andframing. Varying their strengths for each relation independently generatesfour principal codes: ER+/–, SR+/–, where ‘ER+’, for example, condenses‘ER(+C, +F)’. In other words, actors may emphasize the knowledge structure,the knower structure, neither or both as the basis of distinctiveness, authorityand status; conversely, their identity, relations and consciousness are shaped indifferent ways by these two structures. These legitimation codes representdifferent ‘settings’ of the epistemic device, the means whereby intellectual andeducational fields are maintained, reproduced, transformed and changed(Moore and Maton 2001). Whoever controls the epistemic device possesses themeans to set the shape of the field in their favour, making what characterizestheir own practices (in terms of legitimation codes) the basis of status andachievement in the field. This brief and somewhat formal definition can befleshed out by considering the different ways in which the two cultures estab-lished relations to their knowledge–knower structures.

Perhaps the most controversial claim Snow made in his lecture was thatscience was the basis of a true ‘common culture’: ‘the scientific culture really isa culture . . . Without thinking about it, they respond alike. That is what aculture means’ (1959: 9, 10). The basis of this culture was scientists’ ‘sense ofloyalty to an abstraction called “knowledge”’ (Mackerness 1960: 15), commit-ment to ‘truth’ (Bronowski, 1961) and allegiance to their discipline(Pakenham, 1963), which specialized their identity and claims to insight. Inother words, for science the epistemic relation to its knowledge structure wascentral to the field; this structure strongly classifies and frames actors and dis-courses within the field (ER+), while the social relation to its knower structurewas less significant (SR–): what is defined as a knowledge code (see Figure 5.2).

In the case of humanist culture, knowledge itself mattered a lot less; posses-sion of procedures and skills was relatively unimportant in defining identity andachievement, so the epistemic relation to its knowledge structure was weaklyclassified and framed (ER–). Instead, the basis of specialization was possessingthe right kind of dispositions or character. In other words, the field stronglyclassifies and frames knowers (SR+); for the humanities, the social relation to itsknower structure was the key to the field – a knower code. Comparing the twocultures in Figure 5.1 shows it is that which is hierarchical (the pyramids) that


94

strongly classifies and frames actors and discourses within the intellectual field(bold type in Figure 5.2): the epistemic relation to the knowledge structure forscientific culture and the social relation to the knower structure for humanistculture.

Having described the two cultures in terms of their knowledge structures andknower structures and analysed the role they play in specializing insight andidentity in terms of legitimation codes, we can now return to the two questionsraised earlier: the basis of difference between the two cultures and reasons forthe shift of power between them. First, the debate can be redescribed as astruggle for control of the epistemic device between intellectual fields charac-terized by contrasting legitimation codes. These different codes characterize thekind of resources actors bring to the struggle. This is clearly illustrated by thetwo most famous protagonists in the debate: C.P. Snow and F.R. Leavis. Snowrepeatedly emphasized:

On these issues [of relations between the two cultures] our personalities meannothing: but the issues themselves mean a great deal . . . The important thing isto take the personalities, so far as we are able, out of the discussion.(1964: 56, 59)

In contrast, Leavis was concerned with Snow as a legitimate knower:

It is not any challenge he thinks of himself as uttering, but the challenge he is,that demands our attention.(1962: 10–11)

For humanists, as Leavis put it, a ‘judgement is personal or it is nothing; youcannot take over someone else’s’ (1962: 28). This represents a struggle between‘what you know’ (knowledge code) and ‘who you are’ (knower code) asmeasures of status, identity and insight.


Humanist culture

Epistemicrelation

Socialrelation

Legitima-tion code

Scientific culture

Figure 5.2 Legitimation codes of specialization for the two cultures

–C, –F

–C, –F

knower code(ER–, SR+)

knowledge code(ER+, SR–)

+C, +F

+C, +F

Note: Classification (C) refers to relative strength of boundaries between categories orcontexts; framing (F) refers to relative strength of control within these categories or contexts;ER refers to epistemic relation and SR to social relation; ‘+/–’ indicates relativelystronger/weaker. The notation for legitimation codes condenses, for example, ‘ER (+/–C,+/–F)’ to become ‘ER+/–’.

Given this code clash it is little wonder that between the two cultures was saidto lie, as Snow put it, ‘a gulf of mutual incomprehension’. Leavis could bespeaking for both sides when he exclaimed: ‘He doesn’t know what he means,and doesn’t know he doesn’t know’ (1962: 10). Moreover, the rise of scienceand the proclaimed crisis in humanities were intimately interrelated: risingstatus for science threatened to change the basis of the distribution of resourcesand status within the field and relegate humanists to second-class citizens. Ifscientists controlled the epistemic device, then the field would tilt in theirfavour by making a knowledge code the basis of achievement.

Secondly, the difference in codes also suggests reasons for why this shift inpower seemed imminent. One reason lies in the different relationships thecodes establish between their knowledge formations and horizontal discourse(or everyday knowledge). As discussed, science was portrayed as specialized byits language rather than its speakers: who was speaking was said to be less impor-tant than what they were talking about and how. The mathematization ofscience from the seventeenth century onwards had made this language pro-gressively different to commonsense understanding, making discursive distinctionfrom the contents and form of horizontal discourse the basis of the specializ-ation of science. The scientist B.C. Brookes, for example, claimed ‘it will neverbe possible’ to translate between the two and that ‘the learning of science is thelearning of a first, not a foreign, language’ that needed ‘lengthy and ruthlessindoctrination’ (1959a: 502–21, 1959b: 783–4). Measured in terms of its know-ledge code, science was thereby becoming ever more specialized in relation tohorizontal discourse. In contrast, the knower code basis of identity and status inthe humanities made dispositional distinction the basis of status; i.e. distinctionbetween the dispositions of humanist knowers and those of the laity, rather thanthe possession of specialized knowledge and skills.

In these terms the position of humanists was being undermined on twofronts. First, expansion was bringing more varied knowers into higher educationpresenting challenges to its hierarchical knower structure (and so its basis ofspecialization) under wider social conditions where belief in the integratingknower was waning; by the 1960s the ‘English gentleman’ was becoming viewedas outdated in what was being portrayed as an emergent ‘meritocratic’ society.Second, when judged by the discursive distinction of science’s knowledge code,the humanities were becoming less special. The extension of literacy under edu-cational expansion was giving birth to ‘the articulate society’ where everyonefelt entitled to speak and in which the ‘clerk is a nobody not merely because heis not a scientist, but also because in the developed societies everyone is now aclerk’ (Gellner 1964: 78). The humanities did not involve learning specializedprocedures – there ‘is no enormous discontinuity, a yawning gap, bridgeableonly by prolonged training’; instead one could pick up a discipline ‘simply bysoaking in the ambience’ (Gellner 1964: 70) – and so humanists were vulnera-ble to being viewed as speaking little more than a jargon-ridden form ofeveryday language. In short, what threatened humanist culture was the entry ofnew knowers into a field of higher education increasingly dominated by aknowledge code.

To recap, thus far I have addressed fields of knowledge production and the


96

first question I raised at the outset of the chapter of their generative principles.Alongside Bernstein’s ‘knowledge structures’ I introduced the notion of‘knower structures’ to more fully describe differences between intellectualfields. I argued that analysing these knowledge–knower structures in terms oflegitimation codes of specialization provides a means of conceptualizing theirunderlying structuring principles. I also briefly illustrated the kinds of insightsthis can provide into the form taken by different intellectual fields. The secondquestion raised concerned relations between the concepts of knowledge struc-tures and educational knowledge codes, or how Bernstein’s languages ofdescription of fields of production and fields of reproduction can be integrated.It is, therefore, to educational fields of reproduction I now turn.

Knowledge–knower structures in fields of reproduction

In his paper outlining the concepts of educational knowledge codes, Bernstein(1971, 1977) identifies two principal modalities as dominating educationalsystems: a collection code of relatively stronger classification and strongerframing, and an integrated code of weaker classification and weaker framing.These code modalities, he argues, help shape educational identity and con-sciousness in different ways. A collection code emphasizes educational know-ledge, producing what he calls a ‘clear-cut and bounded’ educational identitybased on one’s academic subject (+C, +F). Specialization is thus based on thepossession of knowledge; it ‘makes of educational knowledge something notordinary or mundane, but something esoteric, which gives a special significanceto those who possess it’ (1977: 99). In contrast, under integrated codes the roleof educational knowledge is weakened and one’s educational identity is (on thisbasis) less certain and must be negotiated constantly (–C, –F). This analysis ofeducational knowledge codes can be understood as homologous to thatprovided by knowledge structures such that one can say Bernstein is analysingthe way educational knowledge structures (such as a curriculum) specialize actorsand discursive practices.6 In these terms, Bernstein’s analysis is coding theepistemic relation of educational knowledge (ER+ and ER– for collection andintegrated codes, respectively).

For every educational knowledge structure there is also an educationalknower structure. So, in addition to Bernstein’s analysis, we can also code therole in specialization of the social relation to the educational knower structure. Thisstrength depends on the particular empirical case being examined, but forsimplicity of illustration one can say it is likely that under collection codes thedispositions of knowers play a lesser role (–C, –F). When emphasizing thepossession and transmission of their academic subject knowledge as the basis ofprofessional identity and practice, teachers are likely to comparatively downplaythe significance of their (and their pupils’) biological and/or social disposi-tions. In contrast, under integrated codes there is more space for knowers’dispositions to play a greater role in identity and consciousness (+C, +F); forexample, more emphasis may be given to the capacity to develop the disposi-tions of the ‘whole child’. These classification and framing strengths, which


invert those normally associated with collection and integrated codes, refer tothe social relation to educational knower structures (in these cases, SR– andSR+). Bringing the above together to consider educational knowledge–knower struc-tures, the examples represent a knowledge code (ER+, SR–) and a knower code(ER–, SR+), respectively.

Thus far in this chapter I have focused on instances where coding orient-ations for the epistemic relation to (educational) knowledge structures and thesocial relation to (educational) knower structures are inverted: ER+, SR–(knowledge code) and ER–, SR+ (knower code). However, this is not necessar-ily always the case. The strengths of these two relations may vary independentlyof each other; knowledge structures and knower structures can independentlyplay a stronger or weaker role in specializing actors and discourses. Varyingtheir relative strengths generates four principal legitimation codes of special-ization, as represented by Figure 5.3. As I shall show when using the concepts,this is not a set of dichotomized or binary ideal types: strengths for relations arerelative and represent a continuum; the four legitimation codes are akin tonaming directions created by points on a compass to help orientate oneselfwithin the terrain.

Of the four legitimation codes highlighted in Figure 5.3, I have alreadydiscussed:

• a knowledge code (ER+, SR–), where possession of knowledge (procedures,skills, techniques) is emphasized as the basis of specialization; and

• a knower code (ER–, SR+), where the dispositions or ‘gaze’ of knowers areemphasized, whether these are described as innate or natural (such asnotions of genius), inculcated (such as an artistic or literary sensibility


epistemic relation

socialrelation

knowledge

ER+

ER–

SR– SR+

elite

relativist knower

Figure 5.3 Legitimation codes of specialization

98

cultivated through prolonged immersion in great works) or resulting fromthe knower’s social position (such as standpoint theory based on class, race,gender, sexuality, religion, age, and so forth).

In addition one can also highlight two further coding orientations:

• a relativist code (ER–, SR–), where legitimate identity and insight is ostensi-bly determined by neither knowledge nor dispositions – thoroughgoingrelativism; and

• an élite code (ER+, SR+), where legitimacy is based not only on possessingspecialist knowledge but also being the right kind of knower.7

The élite code is exemplified in intellectual fields by science during the earlyEnlightenment period, when it was not enough to follow scientific procedures tobe considered a legitimate scientist, but one also had to be a gentleman (I discussan élite code in educational fields, below). In short, to think in terms of educa-tional knowledge–knower structures is to ask what makes actors, discourses andpractices special or legitimate: knowledge (knowledge code), dispositions(knower code), neither (relativist code) or both (élite code)? To illustrate howthese concepts can be used to investigate educational fields I shall briefly discusssome empirical research that is addressing the problematic position of music inthe English school curriculum.

School music: an élite code qualificationIn the current English school system there are a number of Key Stages (hence-forth ‘KS’) at which children are tested:

• KS1: school years 1–2 (ages 5–7)• KS2: years 3–6 (ages 7–11)• KS3: years 7–9 (ages 11–14)• KS4: years 10–11 (ages 14–16)

Pupils study a compulsory curriculum of ten academic subjects for KS1–3. Atthis point they can choose, from a wider range of available subjects, which onesthey wish to study for GCSE qualifications (comprising a combination of course-work and examination) completed by the end of year 11. Music is popularamong pupils up to the end of KS3 (Lamont et al. 2003), but there is very lowuptake for GCSE qualifications: approximately 7 per cent of pupils choose totake GCSE Music, compared to 38 per cent for History, 38 per cent for Art andDesign, and 15 per cent for Drama. This unpopularity has not gone unnoticed;in July 2004 a ‘Music Manifesto’ was launched by the British government,aiming to champion the status of the subject and encourage young people toremain involved in music making.

However, the question of why music is so comparatively unpopular remainsunanswered. Most studies of music focus on the learning and playing of musicalinstruments in formal and informal settings outside school, and music in the cur-riculum is typically described as simply ‘out of touch’ or perceived as irrelevant


by most children (e.g. Green 2001; Sloboda 2001). Studies of school musicitself, including the few studies addressing the issue of low uptake at GCSE level,typically offer speculation or ad hoc, piecemeal and largely descriptive accountsof best practice in teaching (e.g. Bray 2000; Harland et al. 2000; see Lamont2004; QCA 2004). This question of why the GCSE qualification in music is sounpopular forms the starting point for a collaborative, interdisciplinary project(with Alexandra Lamont, a music psychologist at Keele University, UK) usingthe concepts of legitimation codes to investigate the ways achievement and edu-cational identities in music are constructed. Our developing hypothesis is thatGCSE Music represents an élite code and that this code plays a role in its lowtake-up rate. To illustrate how these concepts can be used to analyse educationalfields I shall selectively report on the early stages of this research, focusing onthree pilot studies that address:

1. definitions of achievement in National Curriculum documents and syllabi;2. school-pupils’ perceptions of self-ability in, the significance of, and the basis

of achievement in a range of academic subjects including music; and 3. perceptions of university students of significance of and success in various

school subjects.8

1. Curriculum documentsThe first study addresses levels of achievement expected of pupils at differentKey Stages expressed in National Curriculum attainment targets and pro-grammes of study (for KS1–3), and in the GCSE syllabi of major examinationboards. The documents were analysed in terms of whether they emphasized theassessment of: skills, procedures, techniques and knowledge or dispositions ofthe learner, such as aptitude, attitude and personal expression. This analysissuggests the official requirements for music embody different legitimationcodes for different stages of the curriculum. In KS1–2 the National Curriculumdefines achievement in terms of the capacity of pupils to express themselvesrather than demonstrate skills or knowledge. At the end of KS2, for example,pupils are expected to be able to ‘develop their own compositions . . . withincreasing personal involvement, independence and creativity’ (DfES/QCA1999: 18): a knower code. At KS3 (ages 11–14) attainment targets downplayaptitude, attitude and personal engagement in favour of the demonstration ofmusical skills and knowledge; pupils should show an ‘increasing ability to dis-criminate, think critically and make connections between different areas ofknowledge’ (DfES/QCA 1999: 20): a knowledge code. At GCSE level the codechanges again. Examination syllabi for GCSE Music require pupils to demon-strate both their capacity for personal expression and ability with technical skills.The syllabus of the examination board Edexcel, for example, includes a solomusical performance assessed for being both ‘accurate and fluent’ and ‘anexpressive performance that is generally stylish’, with equal emphasis given to‘Accuracy’ and ‘Interpretation’ (Edexcel 2002: 21, 22): an élite code. Thissuggests one possible reason for low uptake may be a shift in legitimation codeunderlying prescribed definitions of achievement in music: from knowledgecode at KS3 to élite code at GCSE.


100

2. Perceptions of pupilsHaving analysed the National Curriculum, the next pilot study focused onwhether these definitions of achievement are reflected in the perceptions ofpupils. A questionnaire was completed by 912 pupils aged 8–14 at four com-prehensive schools of average size and achievement rating. The survey includedthree main questions about music, the core curriculum subjects of English,mathematics and science (which are compulsory subjects for study in KS4), andhistory (for comparison). For each subject pupils were asked to rate the impor-tance of being good at the subject, rate their self-ability, and describe the basisof success at the subject. I shall focus on the third question here: ‘What do youthink makes someone good at [the subject]?’. Respondents were offered aforced choice of one of four options, representing our first attempt at captur-ing relativist, knowledge, knower and élite codes, respectively:

(A) Anyone can do it, nothing special is needed;(B) You need to learn special skills or knowledge;(C) You need to have ‘natural ability’ or a ‘feel’ for it;(D) Only people with ‘natural ability’ can learn the special skills needed.

Analysis of the data for all pupils across all years suggests the basis of success isviewed differently for science and the humanities: modal responses were know-ledge option B for science, maths and music, and relativist option A for English(marginally, over B) and history. (As I discuss below, the latter two may resultfrom our wording of options C and D.) However, this global picture concealssignificant differences in results for different subjects and different pupil ages.One such result of interest here is that among pupils who have chosen theirGCSE subjects in year 9, Music was far more often characterized as embodyingan élite code than other subjects: 19 per cent chose option D for music,compared to a maximum of 3.6 per cent for the other subjects. This figurealmost doubles to 35 per cent among those pupils who chose to study music atGCSE. I shall return to consider the implications of these results shortly.

3. Perceptions of university entrants on school subjectsThe third pilot study explored, through surveys and focus groups, the percep-tions on school subjects of older students who have already made a number ofsubject choices and are starting their university studies. The survey I shalldiscuss comprised 93 new entrants, first-year students at a middle-rankingEnglish university. This included questions about the same three issues asoutlined above for the same four subjects, as well as psychology (in which all thestudents were taking at least a module). For this study we redesigned thequestion of the basis of success in academic subjects for three main reasons.First, our previous dispositional option C offered only ‘natural ability’ or ‘feel’,neglecting the notion of cultivated sensibilities or refined judgement, such asare often emphasized in literary and art criticism (which may account for theprevious low response rate for options including ‘natural ability’). Secondly, thephrasing of the élite option D made ‘natural ability’ the basis for access to‘special skills’ rather than bringing together both dispositions and knowledge.


Thirdly, the pupil survey was our first attempt at using coding concepts in quan-titative research; our choice of a forced-choice design began from the four codemodalities rather than from their basis in the strengths of epistemic and socialrelations. Such a categorical scale design suits ideal typical groupings, whereasthe theory emphasizes the relative strengths of the two relations in determiningcoding orientations, requiring a more continuous scaling approach. To addressthese issues, ‘taste, judgement or feel’ was added as an option alongside ‘naturaltalent’ (separately because though both are dispositional they are oftenopposed in well-known debates over, for example, ‘nature versus nurture’), andthe forced-choice design was replaced with a sliding scale of importance (seeFigure 5.4). The new design thus asked respondents to rate the significance ofa subject’s knowledge structure (‘skills’) and of its knower structure (‘talent’and ‘taste’).

The theory’s emphasis on relative strengths was also reflected by the analysis.The ratings were coded numerically as 1–4 and mean scores calculated acrossall subjects for the ‘skills’ scale and for the ‘talent’ and ‘taste’ scales taken together,to give baseline scores of significance for the epistemic and social relations,respectively. The scores of each subject for these two relations were thencompared to these two overall mean scores. From this analysis (see Figure 5.5):

• natural science (and, to a lesser extent, psychology) scored higher for‘skills’ (stronger epistemic relation) and lower for ‘talent / taste’ (weakersocial relation) – a knowledge code;

• for English these results were reversed, with the epistemic relation beingweaker and the social relation being stronger – a knower code;

• maths was average for both, a result requiring further investigation; • history scored lower for both skills and ‘talent / taste’ – a relativist code; and • music scored higher for both – an élite code.

In other words, comparative to other subjects one requires both skills and tasteor talent to succeed in music.


In your opinion, how important are these things for being good at [the subject]?

Not at all Not very Quite Very

Skills, techniques and specialist knowledge [ ] [ ] [ ] [ ]

Natural-born talent [ ] [ ] [ ] [ ]

Taste, judgement or a developed ‘feel’ for it [ ] [ ] [ ] [ ]

Figure 5.4 University students’ survey; question of basis of achievement

102

We also conducted a series of focus groups with members of the same pop-ulation. The above coding orientations were reflected in how participantsdiscussed achievement. Group discussions of science and English literature, forexample, illustrate their respective knowledge code and knower code. Partici-pants tended to state that for English the knower is the source of a legitimategaze which generates insights, whereas in science it is educational knowledgethat forms the basis of insight and identity; for example:

Moderator: What does it take to be good at English?Participant 1: I learnt to have my own opinion and back it up with my own

evidence but then use evidence from other people that have the sameopinion as me, so you’re still using other opinions but you’re findingthem after you’ve made your own.

Moderator: Is that different from science or maths? P2: Yeah, definitely. You can’t really say ‘Well, my theory of evolution is . . .’

It’s not like you can make up your own theory.P1: You’re given theories and you choose one rather than having your own

opinion and then finding someone who agrees with you.

The élite coding orientation of music was reflected in how participants wouldoften shift between talent or natural ability, developing skills and such issues asportraying emotion when discussing achievement in the subject:


Figure 5.5 University students’ perceptions of basis of achievement

NaturalScience

■■ Skills, techniques and specialist knowledge

■ Talent, taste, judgement or a developed ‘feel’

Maths Psychology EnglishLiterature

History Music

–

P3: It’s more talent-based, you have to have a natural ability.P4: You can’t just throw anyone in there and teach them, they have to have

that ability before they start. Everyone can learn the basics but to get tothe top . . .

P3: Music takes lots of practice . . . you have to practise every day to get betterat it.

P5: You can never say you’ve done all the work for it. You can always do a bitbetter. Whereas in science if you learn it there is a point where you’velearned everything that you need to know.

P3: Even someone with natural talent that’s very good at music still has topractise.

P5: It’s talent and skills and hard work. P3: You need to be able to portray emotion too.

Implications of music’s élite codeIf Music is portrayed as embodying an élite code in curriculum documents andthe perceptions of school-pupils and university entrants, the question is howthat coding orientation might relate to its unpopularity at GCSE level. Theabove discussion summarizes only part of the analysis of the data, whichincludes age differences, social variables such as gender, differences between‘taste’ and ‘talent’, rating of self-ability and the significance of different results.The research is also ongoing (for example, the redesigned survey is being usedwithin an intervention study in English secondary schools). In addition, furtherstudies are required for a fuller picture, including analysis of: the social distri-bution of legitimation codes among different groups of pupils; constructions ofachievement within the intellectual field of music education research; the for-mulation of curriculum in the field of recontextualization; and the ways schoolmusic is taught at different Key Stages. However, the results I have discussed arethemselves suggestive in several ways. First, the shift to élite code at GCSE levelis not simply a code shift (as happens between KS2 and KS3, from knower codeto knowledge code) but to one that is doubly demanding: pupils must bothdemonstrate possession of musical knowledge and skills, and express musicaldispositions. In other words, not only are the rules of the game changed, but itbecomes harder to play – an élite code has two hierarchies (one of which, theknower structure, it may not be possible to climb). Second, this may make musican unenticing option, particularly if one considers its perceived significance.When asked to rate the importance of being good at a subject, music was theleast important subject for both school-pupils and university entrants. Its élitecode thereby does not seem to be reflected by status; as one focus groupdescribed it:

P6: I don’t think if you were going to apply to be a doctor they’d say, ‘Haveyou got your grade 9 piano’ or whatever.

P7: I think if I told people I was doing a music degree everyone would be like‘What’s the point? Waste of time!’ kind of thing.

P8: Yeah, everyone thinks doing music at university is learning to play ‘ThreeBlind Mice’ on the recorder.


104

Lastly, if music’s élite code is not widely distributed socially and the keys to thecode not made visible in pedagogic practices, then school qualifications arelikely to remain restricted.

The central point here, however, is less the basis of music’s position and morethe issue of how analysing knowledge–knower structures in terms of legitima-tion codes offers fruitful ways forward for empirical research into educationalfields. It reveals not only contexts exhibiting stronger or weaker classificationand framing, but also those with both; such contexts may appear contradictoryor confusing if one considers educational knowledge structures on their own.Élite schools, for example, may operate with selection criteria based not only onqualifications but also on issues of character and disposition. Integratingknower structures into the analysis may show that such contexts exhibit an élitelegitimation code. By being anchored on the concepts of classification andframing, the strong external language of description of legitimation codetheory also enables both qualitative and quantitative analysis of the underlyingprinciples structuring curriculum guidelines, teaching practices, pupils’ per-ceptions, school organization, and so forth in a manner enabling systematiccomparisons within and between these contexts, something currently lackingfrom existing research on music in the curriculum.

Conclusion

In this chapter I have argued that one can analytically distinguish two messagesystems within the discursive practices of intellectual and educational fields. Thefirst is that addressed by Bernstein’s conceptualization of fields of production interms of knowledge structures, and fields of reproduction in terms of educa-tional knowledge codes. These (educational) knowledge structures announcehow knowledge should be created, distributed, recontextualized and evaluated.The second is to be found in their knower structures. I introduced the notionsof hierarchical knower structures and horizontal knower structures withstronger and weaker ‘knower-grammars’; these proclaim how legitimate knowersare created, distributed, recontextualized and evaluated. This is, though, an ana-lytical distinction. To understand fully intellectual and educational fields onemust, I argued, bring these together to think in terms of knowledge–knower struc-tures. Together they define the basis of specialization of actors and discourseswithin fields and so help shape relation, identity and consciousness.

These concepts enable two questions raised by Bernstein’s ideas to beaddressed. First, I showed how analysing knowledge–knower structures in termsof legitimation codes and the epistemic device provides a means of exploring thestructuring principles underlying intellectual fields. This also highlights newissues of interest; for example, it recasts the question of hierarchies in intellec-tual fields from ‘whether’ to ‘where’ (in their knowledge structure, knower struc-ture, neither or both) – horizontal knowledge structures may be characterized byhierarchical knower structures. Doing so further clarifies the recontextualizingprinciples of intellectual fields – this principle may be not only discursively basedbut also dispositionally based, with implications for the structure and develop-


ment of the field. The notion of knowledge–knower structures also expands thereach of the analysis, subsuming and integrating (rather than displacing) theexisting conceptualizations of knowledge structures and educational knowledgecodes. Integrating the analysis of knower structures with that of knowledge struc-tures within the concept of legitimation code not only enables their differentinsights to be brought together but also enables us to generatively conceptualizenew possibilities, such as relativist and élite codes. Secondly, having consideredintellectual fields (with the example of the ‘two cultures’ debate) I illustratedhow these concepts can also be applied to educational fields by discussing studiesinto school music. Preliminary results from these studies suggest that the unpop-ularity of school qualifications in music may be related to its élite code. Theconcepts thereby provide a means of bringing analyses of fields of intellectualproduction and of educational reproduction within the same framework. Theuse of these concepts is at a relatively early stage; as more empirical questions areaddressed, their tacit potential and limitations will become increasingly evident,necessitating further development of the theory and opening up the nextchapter of the story.

References

Bernstein, B. (1971), ‘On the classification and framing of educational knowledge’, inM.F.D. Young (ed.), Knowledge and Control. London: Collier Macmillan.

Bernstein, B. (1977), Class, Codes and Control, Volume III: Towards a theory of educationaltransmissions. London: Routledge & Kegan Paul.

Bernstein, B. (1990), Class, Codes and Control, Volume IV: The structuring of pedagogic dis-course. London: Routledge.

Bernstein, B. (1996), Pedagogy, Symbolic Control and Identity: Theory, research, critique.London: Taylor & Francis.

Bernstein, B. (1999), ‘Vertical and horizontal discourse: An essay’, in British Journal of Soci-ology of Education 20(2), pp. 157–73.

Bourdieu, P. and Passeron, J-C. (1990), Reproduction in Education, Society and Culture.London: Sage.

Bray, D. (2000), ‘An examination of GCSE music uptake rates’, in British Journal of MusicEducation 17(1), pp. 79–89.

Bronowski, J. (1961), Science and Human Values. London: Hutchinson.Brookes, B.C. (1959a), ‘The difficulty of interpreting science’, in The Listener LXII,

pp. 519–21.Brookes, B.C. (1959b), Letter to the editor, in The Listener LXII, pp. 783–4.DfES/QCA [Department for Education and Science/Qualifications & Curriculum

Authority] (1999), The National Curriculum for England: Music. London: HMSO.Doherty, C. (2004), ‘Student subsidy of the internationalized curriculum: Knowing,

voicing and producing the Other’. Paper presented at Reclaiming Knowledge: Registersof discourse in the community and school conference, University of Sydney, December.

Edexcel (2002), Specification: Edexcel Advanced Subsidiary GCE in Music (8501) and AdvancedGCE in Music (9501). London: Edexcel Foundation.

Gellner, E. (1964), ‘The crisis in the humanities and the mainstream of philosophy’, inPlumb, J.H. (ed.), Crisis in the Humanities. Harmondsworth: Penguin.

Green, L. (2001), How Popular Musicians Learn: A way ahead for music education. Burling-ton: Ashgate.


106

Harland, J. et al. (2000), Arts Education in Secondary Schools: Effects and Effectiveness. Slough:National Foundation for Educational Research.

Hickox, M. and Moore, R. (1995), ‘Liberal-humanist education: The vocationalist chal-lenge’, in Curriculum Studies 3(1), pp. 45–59.

Hood, S. and Maton, K. (2005), ‘The languages of disciplinarity: Knowledge, knowersand recontextualisation’. Paper presented at Discourses of Hope: International SystemicFunctional Congress 32, University of Sydney, July.

Lamont, A. (2004), What are the possible reasons for the low take-up of music GCSE?Music in Key Stage 4: QCA Music Development Group working paper.

Lamont, A., Hargreaves, D.J., Marshall, N.A. and Tarrant, M. (2003), ‘Young people’smusic in and out of school’, in British Journal of Music Education 20(3), pp. 1–13.

Leavis, F.R. (1962), Two Cultures: The significance of C. P. Snow. London: Chatto & Windus.Lee, H.D.P. (1955), ‘The position of Classics in schools’, in Universities Quarterly 9(2),

pp. 135–44.Mackerness, E.D. (1960), ‘Ignorant armies’, in The Universities Review 33(1), pp. 14–17.Maton, K. (2000a), ‘Recovering pedagogic discourse: A Bernsteinian approach to the

sociology of educational knowledge’, in Linguistics and Education 11(1), pp. 79–98.Maton, K. (2000b), ‘Languages of legitimation: The structuring significance for intellec-

tual fields of strategic knowledge claims’, in British Journal of Sociology of Education21(2), pp. 147–67.

Maton, K. (2002), ‘Popes, kings and cultural studies: Placing the commitment to non-disciplinarity in historical context’, in S. Herbrechter (ed.), Cultural Studies: Interdis-ciplinarity and translation. Amsterdam: Rodopi.

Maton, K. (2004), ‘The wrong kind of knower: Education, expansion and the epistemicdevice’, in Muller, J. Davies, B. and Morais, A. (eds), Reading Bernstein, ResearchingBernstein. London: RoutledgeFalmer.

Maton, K. (2006), ‘On knowledge structures and knower structures’, in Moore, R., Arnot,M., Beck, J. and Daniels, H. (eds), Knowledge, Power and Educational Reform: Applyingthe sociology of Basil Bernstein. London: Routledge.

Moore, R. (2004), Education and Society: Issues and explanations in the sociology of education.Cambridge: Polity Press.

Moore, R. and Maton, K. (2001), ‘Founding the sociology of knowledge: Basil Bernstein,intellectual fields and the epistemic device’, in Morais, A. Neves, I., Davies, B. andDaniels, H. (eds), Towards a Sociology of Pedagogy: The contribution of Basil Bernstein toresearch. New York: Peter Lang.

Moore, R. and Muller, J. (2002), ‘The growth of knowledge and the discursive gap’, inBritish Journal of Sociology of Education 23(4), pp. 627–38.

Morris, M. (1959), ‘The two cultures and the scientific revolution’, in Marxism Today3(12), pp. 374–80.

Muller, J. (2004), ‘On the shoulders of giants: A digression on knowledge, curriculum,and finally, the teacher’. Paper presented at the Third International Basil BernsteinSymposium, University of Cambridge, July.

Pakenham, T. (1963), ‘The study of the university teacher: The Gulbenkian EducationalDiscussion 1962’, in Universities Quarterly 17(2), pp. 149–64.

Plumb, J.H. (ed.) (1964a), Crisis in the Humanities. Harmondsworth: Penguin.Plumb, J.H. (1964b), ‘The historian’s dilemma’, in J.H. Plumb (ed.), Crisis in the Human-

ities. Harmondsworth: Penguin. QCA [Qualifications and Curriculum Authority] (2004), Music in Key Stage 4: Developing

qualifications within the vision of a holistic and progressive music education for all. QCAMusic Development Working Paper.

Singh, P. (2002), ‘Pedagogising knowledge: Bernstein’s theory of the pedagogic device’,in British Journal of Sociology of Education 23(4), pp. 571–82.

Sloboda, J.A. (2001), ‘Conference Keynote: Emotion, Functionality and the EverydayExperience of Music’, in Music Education Research 3(2), pp. 243–53.


Snow, C.P. (1959), The Two Cultures and the Scientific Revolution. Cambridge: CambridgeUniversity Press.

Snow, C.P. (1964), The Two Cultures and a Second Look. Cambridge: Cambridge UniversityPress.

Wheelahan, L. (2005), ‘Comparing realist and rationalist defences of propositionalknowledge in the curriculum’. Paper presented at Engaging Realist Alternatives: IACRConference, Sydney, July.

Notes

1 This paper expands on Maton (2006), in particular elaborating on grammars ofknower structures and developing the account of knowledge–knower structures ineducational fields.

2 See Doherty (2004), Hood and Maton (2005), Lamont (2004), Maton (2004) andWheelahan (2005) for examples of educational studies using legitimation codes.

3 This is not the same as asking how specific kinds of knowledge structures andcurriculum/pedagogic structures are related; for a perceptive account exploring thisissue, see Muller (2004).

4 Integration of new knowers may be through resocialization (such as that attemptedby the creation of new campus universities as resocializing institutions in 1960sEnglish higher education; Maton 2004) or through a mixture of indoctrination andcoercion (such as underlay medieval monarchical and papal hierarchies; Maton2002). Educational expansion has typically accommodated new knowers throughoffering a choice of resocialization as the condition of entry into higher statusinstitutional and disciplinary positions, or relegation into lower levels of these statushierarchies (cf. Hickox and Moore 1995). There may be more than one idealizedknower and pyramid of knowers in an intellectual field characterized by a hier-archical knower structure.

5 This has effects for the possibility and means of transmission/acquisition of the legit-imate ‘gaze’ conferred by knower status. In the case of stronger ‘knower-grammars’acquisition may be explicitly restricted to those already possessing knower status andtransmission may focus on attempting to raise to consciousness what is proclaimedto be a pre-existing gaze; with weaker ‘knower-grammars’, the possibility of success-ful acquisition of the legitimate gaze is claimed to be more widely available and trans-mission may take the form of attempting to resocialize potential knowers,restructuring their habituses.

6 I should emphasize that Bernstein’s concepts of knowledge structures refer to intel-lectual fields of production and not educational fields – they are not describing cur-riculum or pedagogic structures. I use ‘educational knowledge structures’ by way ofanalogy as a step towards offering a means of integrating analyses of intellectual andeducational fields.

7 One can expand the language of description to provide a more subtle theorizationby considering differing strengths of classification and framing. This generatessixteen modalities: ER(+/–C, +/–F), SR(+/–C, +/–F). Here I restrict discussion ofcoding orientations to where classification and framing strengths are both stronger orweaker for conceptual economy and because Bernstein’s theorization and applica-tions of the concepts suggest they are the most commonly found orientations.

8 My principal focus is illustrating the application of the concepts rather than themusic question per se. I thus discuss only selected aspects of the studies. We shallreport the results of this ongoing research and their implications for school musicmore fully in future publications.

9 It is tempting to account for the low uptake rate of music in terms of value in theoccupational marketplace alone; however, this would not easily account for theuptake rate for drama being double that of music.


108

Acknowledgements

I am extremely grateful to Alexandra Lamont, with whom I am conducting the musicresearch; Jim Martin and Fran Christie, for inviting me to give a keynote paper at theReclaiming Knowledge conference; Geoff Whitty and Brian Davies for enlightening discus-sions of my work as a whole; and Joe Muller and Rob Moore for critical encouragementin my ideas.


As part of society, the individual naturally transcends himself, both when hethinks and when he acts.

(Emile Durkheim)

‘Truth’ is the reigning sacred object of the scholarly community, as ‘art’ is for theliterary/artistic communities; these are simultaneously their highest cognitive andmoral categories, the locus of highest value, by which all else is judged.

(Randall Collins)

Introduction

This chapter addresses from the perspective of aesthetic concerns a set of issuesmore usually approached in terms of epistemology. Essentially, the concern iswith the production of judgements – the evaluation of the relative merits of artobjects and the discrimination of some as being ‘better’ than others. It must bestressed that this is an exercise in sociology, not in aesthetics or criticism. Noclaims are made about any particular art object relative to any other. To say thatit is sociological is to indicate that the focus is upon the sociality of the produc-tion of aesthetic judgements: that the production of aesthetic judgement issomething that people do within a special type of socio-historical context (a fieldor arena) that can be described in terms of its structural features, its generativeprinciples (its ‘powers’) and understood at the level of the individual in termsof the mediation of habitus. This type of context is conventionally termed a‘canon’ and its distinctive feature is that it is extensive in space and time.

In his later writings on hierarchical and horizontal knowledge structures, BasilBernstein (Bernstein 2000) began to develop a new conceptual languagethrough which to theorize the structurings of intellectual fields and the con-ditions of their productivity. The basic difference between these two types ofknowledge structure (each of which can vary in strength within itself as well as dif-fering between each other) is the capacity of hierarchical knowledge structuresto produce knowledge at increasingly higher levels of abstraction, generality andintegration (they are strong in grammaticality – see Muller, this volume) and,hence, are able to produce progression in knowledge. By contrast, horizontalknowledge structures specialize knowledge to knowers by constructing a high levelof segmentation between social categories of ‘knowers’ and grounding know-ledge in their distinctive and incommensurable sets of experience – in particu-lar, experiences of unequal power relations in society.

Horizontal knowledge structures legitimate themselves in terms of who knows

6 Hierarchical knowledge structures and thecanon: a preference for judgements

Rob Moore

110

rather than what is known. They authorize themselves through the ‘voice’ ofthose whose experiences they claim to represent (Moore and Muller 1999). AsKarl Maton has demonstrated (2000), these two types of knowledge structureoperate with different ‘languages of legitimation’ associated with different ori-entations of the ‘epistemic device’ (Maton, this volume; Moore and Maton2000). In an early formulation before he had developed the concept of ‘know-ledge structures’, Bernstein wrote as follows:

In a subject where theories and methods are weak, intellectual shifts are likely toarise out of conflict between approaches rather than conflict between explanations,for, by definition, most explanations will be weak and often non-comparable,because they are approach-specific. The weakness of the explanation is likely tobe attributed to the approach, which is analysed in terms of ideological stance.Once the ideological stance is exposed, then all the work may be written off. Everynew approach becomes a social movement or sect which immediately defines thenature of the subject by re-defining what is to be admitted, and what is beyond thepale, so that with every new approach the subject almost starts from scratch.(Bernstein 1977: 167–8)

This is from a ‘brief account’ of the sociology of education from the early 1970s.It is of interest because the distinction between debates at the level of‘approaches’ or ‘explanations’ points forward to the concepts of horizontal andhierarchical knowledge structures, and also because, some thirty years later,under the influence of post-modernist, post-structuralist and feminist stand-point approaches, intellectual fields within the social sciences and humanitiesdisplay these features of horizontal segmentation (low grammaticality) even morestrongly than when Bernstein was originally writing! The intention of this paperis to indicate ways in which Bernstein’s later concepts (and also the work ofothers) might enable solutions to problems in the sociology of knowledgethrough a new theory of the structuring of intellectual fields and, in so doing,suggest the conditions for intellectual productivity and progress in knowledgein the social sciences and humanities. Specifically, this chapter will address the‘sociality of judgement’ through a sociological defence of the concept of thecanon.

The canon under fire

For some time, the idea of a canon, and its associated concepts, have been chal-lenged from a number of quarters (by various currents of post-modernism, forinstance). The essence of this ‘critical’ challenge is that the basic principle of acanon – that it enshrines those things (art objects, ideas, etc.) that are of intrin-sic, superior worth – is fundamentally false. Canons are seen as arbitrary con-structions reflecting no more than the tastes and fashions of dominant socialgroups or, at worse, as ideological forces that legitimate and reproduce theposition and power of dominant groups. Whereas in the traditional manner acanon is taken as inclusive and universal, in the critical approach it is seen as


exclusive and partial. Critical approaches tend to assume two forms: reduction-ism and relationalism.

Reductionism has its origins in Marxist ideology analysis and standpointtheory. It operates by identifying the social base of a work of art or movement,the standpoint that it reflects and the interest it serves. In the traditional Marxistform artistic movements might be classified as ‘bourgeois art’; more recently,feminist examples of this approach might describe the literary canon as the‘male literary canon’.

Relationalism is best represented in the work of Pierre Bourdieu (e.g. 1993).In this case, items are distributed in a ‘field’ (e.g. the cultural field) and arerelated to each other in terms of their synchronistic positioning. Fields arestructured between high and low (that which is deemed to have distinction andconfers ‘cultural capital’, and that which is vulgar). Aesthetic value is generatedthrough the relationships between positions rather than through the art object initself. To attribute intrinsic aesthetic value to an art object is ‘essentialism’ thatinvolves ‘misrecognition’ and entails ‘symbolic violence’. Values are ultimatelyarbitrary in that they reflect no more than the social distribution of taste and thestructure of the field; no more than the system of power relations and economicinterests of society (Moore 2004a).

Although relationalism is conceptually more complex and sophisticated thanreductionism, ultimately, they both say the same things: cultural and intellectualfields are basically arbitrary and have no intrinsic values that can be consideredindependently of external social relations, interests and power. Once the socialbase has been revealed and the interest it serves and the standpoint it reflectsexposed, there is no more to say – nothing in terms of truth or beauty in theirown right.

In recent years these debates have been fiercely expressed in the so-calledculture and science ‘wars’ (Graff 1992; Brown 2001). Richard Hoggart has asked:

Why are the arguments so angry? Why are so many people so violently disinclinedto admit any differences in the value of different works of art; or between humanchoices as to activities? This is the most revealing of our multiple cultural hang-ups. It involves many people – the conclusion is inescapable – almost entirelyrejecting ‘great’ works of art in any form (people who dismiss George Eliot as‘merely a reflection of nineteenth century bourgeois values’ and clearly have notread her work). Such people as these are well-versed in one or the other art, butunwilling any longer to make value-judgements between them. By extension, theyare uneasy about talk of art’s possible relation to ‘meaning’. They avoid anyvertical judgements, in favour of the endlessly horizontal. By the Nineties a seniorofficial with Radio 3 [the BBC’s classical music station] could announce: ‘Thereis no art; only culture’. A spiritual mate adds: ‘Each man is his own culture’.(Hoggart 1995: 57–8)

Basically, the problem is that of relativism – cognitive, moral and aesthetic.The argument in this paper is that relativism is the consequence of a mistakenpreoccupation with absolutes in relation to values, and with a failure to properlyunderstand the character of the sociality of knowledge.

111HIERARCHICAL KNOWLEDGE STRUCTURES • MOORE

112

It is interesting that Hoggart also employs a distinction between the ‘hori-zontal’ and the ‘vertical’. For Bernstein, the origin of these ideas is inDurkheim. In the later part of the twentieth century Durkheimian scholars(notably Basil Bernstein and Randall Collins) began to develop the conceptualapparatus whereby the structures, principles, powers and possibilities of theseforms of sociality could be theorized and explored systematically. Others (e.g.John Beck, Karl Maton, Rob Moore, Johan Muller, Michael Young, individuallyand in various combinations: see the bibliography to this chapter) extended therange of this effort. This chapter draws upon that language and indicates waysin which longstanding philosophical issues might be addressed in terms of soci-ological understandings of how symbolic production ‘works’ in terms of thesociality of judgement.

The mind itself – the endlessly horizontal

Because the field of debate is so broad, the focus will be upon one particularrecent text in order to provide a central reference point. The emeritus professorof English Literature at Oxford University, John Carey, published a bookprovocatively entitled: What Good are the Arts? (Carey 2005). This book has beenextensively reviewed and, as intended, no doubt, caused considerable contro-versy. Carey explicitly endorses the view cited above, that ‘Each man is his ownculture’. His central argument is: ‘Anything can be a work of art. What makes ita work of art is that someone thinks of it as a work of art’ (Carey 2005: 29). Inmore substance, there is a distinction between those who

. . . assume the existence of a separate category of things called works of art . . .which are intrinsically more valuable than things which are not works of art, andwhich accordingly deserve universal respect and admiration. These assumptions,we can now see, belong to the late 18th century, and are no longer valid. Thequestion ‘Is this a work of art? – asked in anger or indignation or mere puzzle-ment – can now receive only the answer ‘Yes, if you think it is; no, if not.’ If thisseems to plunge us into the abyss of relativism, then I can only say that the abyssof relativism is where we have always been in reality – if it is an abyss. (Carey 2005:30)

Carey’s position is allied with post-modern relativism. For Carey, and others,relativism simply is not a problem. However, Carey does have a problem becausehis book contains two arguments, the second of which is a defence of literatureon the grounds that it is ‘superior to the other arts, and can do things theycannot do’ (ibid., p. 173). Carey is not unaware of the fact that this argumentseems at odds with the relativism that precedes it (loc. cit.), but he does not inany satisfactory manner resolve this tension. The reason for this is that it is notjust that Carey presents two contradictory arguments, but that he draws upon orimplies two quite different theories of knowledge in doing so. In the firstinstance there is the subjectivist relativism that he explicitly espouses, but in hissecond argument there are the elements of a radically different approach. In


Bernsteinian terms, Carey number one argues in ‘horizontal’ mode (whoknows), whereas Carey two argues in ‘vertical’ mode (what is known). The firsttask, though, is to demonstrate the wider implications of Carey’s position for theissues being addressed.

The argument will begin by drawing attention to how Carey believes the artswould have to be if they were indeed to be special: ‘a separate category of thingscalled works of art . . . which are intrinsically more valuable than things whichare not works of art’. He grounds this, as he sees it, elitist and mistakenapproach to the arts, in the ideas of Kant:

It is easy to identify the dictates of Kant and his followers in the notions about artthat are still in circulation today. That art is somehow sacred, that it is ‘deeper’ or‘higher’ than science and reveals ‘truths’ beyond science’s scope, that it refinesour sensibilities and makes us better people, that it is produced by geniuses whomust not be expected to obey the same moral codes as the rest of us, that it shouldnot arouse sexual desire, or it will become pornography, which is bad – these andother superstitions belong to the Kantian inheritance. So does belief in thespecial nature of artworks. For Kantians, the question ‘What is a work of art?’makes sense and is answerable. Works of art belong to a separate category ofthings, recognized and attested by certain highly gifted individuals who view themin a state of pure contemplation, and their status as works of art is absolute,universal and eternal.(Carey 2005: 14)

Compare Carey’s statement with this example of a post-modernist account ofscience:

The ascendancy of scientific method as the means of establishing knowledge hasresulted in a consistent failure to examine science as a social practice and as ahistorical and cultural product. Science has instead been seen as transcendentaland decontextualised. Knowledge, as well as the knowing subject, therefore,becomes context free. Rationality is cast as universal and transcendental, operat-ing across all historical and social contexts and practices but independent of allof them. The result is an individualistic epistemology where the solitary individualconfronts an independent reality of objects. (Usher and Edwards 1994: 36)

Carey himself does not set out to present a critique of science. Indeed, at anumber of points, his argument is precisely that aesthetics cannot do whatscience does (and in this respect Carey does not appear to be an out-and-outcognitive relativist). The point is that in both cases these writers present thesame basic logic of argument for rejecting truth claims in the areas of aesthet-ics and science. To put it succinctly, essentialism is to aesthetics what founda-tionalism is to epistemology. In both cases, it is being argued that truth claimsor value judgements can only be forwarded if they are ‘universal and transcen-dental, operating across all historical and social contexts and practices butindependent of all of them’, and are ‘absolute, universal and eternal’. And,similarly, there is the (basically Cartesian) model of the ‘privileged knower’ as‘certain highly gifted individuals who view them [artworks] in a state of pure


114

contemplation’ and where ‘the solitary individual confronts an independentreality of objects’. In both cases these writers present absolutist or infallibilistdefinitions of truth whether aesthetic or epistemological and atomistic, con-templative models of the self. They both provide strong demolitions ofabsolutist claims. The problem is that they then conclude that because truthclaims cannot be absolute – ‘universal and transcendental, operating across allhistorical and social contexts and practices but independent of all of them’ –then we cannot have truth claims at all. In this crucial respect, both sets ofwriters are simply wrong. The fact that we cannot produce truths that areabsolutely infallible in the areas of epistemology, aesthetics or morals does notmean that we cannot intelligibly produce judgements that some things or prac-tices are better than others.

The All-or-Nothing FallacyIt is important to note that Carey advances his position on the basis of a partic-ular set of alternatives: ‘The champion of high art would have to mean not justthat his experience were more valuable to him, for that would not prove thesuperiority of high art, only his preference for it. He would have to mean thatthe experiences he derived from high art were in some absolute and intrinsicsense more valuable than anything the other person could get from low art.’ Weare allowed only the choice between a preference or an absolute. This is a form ofwhat J.C. Alexander has called the ‘epistemological dilemma’ – i.e. that the onlychoices available to us are those of relativism and absolutism (1995). For Carey,if we cannot provide absolute (‘divinely decreed’) support for our judgementsthen we cannot have any judgments at all, but merely preferences and tastes.Niiniluoto has made an observation similar to Alexander when describing‘dogmatic scepticism’: ‘This mode of thinking could be called the All-or-Nothing-Fallacy. Its different forms propose a strong or absolute standard for somecategory . . . and interpret the failure or impossibility of satisfying this standardas a proof that the category is empty and should be rejected’ (2002: 81). This isprecisely Carey’s approach. Carey’s thinking is an example of a wider mode ofthought that has been identified and critiqued by a number of commentators(Moore 2004b: ch. 6).

Preferences and judgementsIn a key statement, Carey displays not only the manner in which his approachparallels a number of other kinds of relativism in the social sciences and culturalstudies, but also indicates the kernel of the problem at the centre of such per-spectives – essentially, a confusion over the difference between a preference anda judgement. Carey declares himself as follows:

I have suggested that those who proclaim the superiority of high art are saying, ineffect, to those who get their pleasure from low art, ‘What I feel is more valuablethan what you feel.’ We can see now that such a claim is nonsense psychologically,because other people’s feeling cannot be accessed. But even if they could, wouldit be meaningful to assert that your experiences were more valuable thansomeone else’s? The champion of high art would have to mean not just that his


experience were more valuable to him, for that would not prove the superiority ofhigh art, only his preference for it. He would have to mean that the experienceshe derived from high art were in some absolute and intrinsic sense more valuablethan anything the other person could get from low art. How could such a claimmake sense? What could ‘valuable’ mean in such a claim? It could have meaningonly in a world of divinely decreed absolutes – a world in which God decideswhich kinds of feelings are valuable and which are not – and this, as I have said,is not the world in which I am conducting my argument. (Carey 2005: 25)

There are four symptomatic points to take from this that apply equally acrossthe range of relativisms.

• The first is its subjectivism: claims are specialized to the experience or feelingsof the knowing subject.

• The second is that, as a consequence, knowledge relations are rewritten asrelations between social groups – groups of ‘knowers’ or, indeed ‘feelers’.

• The third is that statements of judgement are no more than statements ofgroup tastes or preferences (Carey devotes three pages, 117 to 120, toBourdieu’s classic study, Distinction – albeit with some reservations).

• The fourth is that the only alternative to taste is divinely decreed absolutes.

In these respects, Carey’s position in aesthetics reproduces the basic proposi-tions that underpin post-modern scepticism, feminist standpoint theory and awide range of other forms of contemporary relativist perspectivism. Knowledgeis no more than what different knowers know. This results in a kind of socialembarrassment in which it is taken that to judge some things, values or practicesto be better than others is to imply that some people are better than others.Because we cannot detach judgements about knowledge from judgementsabout knowers we cannot judge the one without judging the other (or ‘Other’in the post-structuralist, semiotic version of this). This answers Hoggart’squestion as to why the arguments are so angry – because people are taking thempersonally. What is to one an aesthetic judgement is to another a cultural snub.

Carey’s view is that we are limited only to preferences or absolutes in the areaof judgement, and because we cannot have absolutes we are left with nothingbut preferences – assertions of taste alone. However, it is a commonsense obser-vation that there is no major problem in distinguishing between the exercisingof a preference and the production of a judgement. The claim by any individ-ual that they prefer to read P.D. James’ murder mysteries, but judge the novelsof Henry James to be of superior literary significance is perfectly intelligible,as would be the statement that ‘I prefer to listen to Chuck Berry, but judgeBeethoven to be musically superior’. Preferences and judgements are differentthings that operate in different ways. In the main, preferences are acquired in anad hoc and contingent kind of way whereas judgements are actively producedaccording to publicly shared rules of various kinds. For any particular individual,


116

it is simply a matter of life-history whether they prefer beer to wine, and, if beer,traditional English ales to Continental lagers, or Chinese cuisine to Indian orFrench to Italian. The world of preferences is the world of relativism, of taste,because in many areas of life there is no good reason why anyone should do anyone thing rather than any other. But, as Carey makes very clear in his secondargument in defence of literature, it is not the same in the world of judgements.

Carey acknowledges that his attempt to demonstrate the ‘superiority’ ofLiterature as a form of art sits strangely with his argument that art is simply whatanyone wants it to be. He presents his intention in this way:

In the rest of this book I intend to make out the case for valuing literature, takingexamples for the most part, but not exclusively, from English literature – a branchof knowledge which, in recent years, has been progressively devalued in schoolsand universities, and regarded as rather shamefully parochial and old-worldcompared to, say, media studies or cultural history. In opposition to this, I shallalso try to show why literature is superior to the other arts, and can do things theycannot do. Just in case anyone should seize on these aims as inconsistent with therelativist cast of the first part of my book, let me emphasize that all the judgementsmade in this part, including the judgement of what ‘literature’ is, are inevitablysubjective. (Carey 2005: 173)

The crucial point at the end of this is the reference to the ‘inevitably subjective’nature of Carey’s judgements. He immediately goes on to say that his definitionof literature is ‘writing that I want to remember – not for its content alone, asone might want to remember a computer manual, but for itself: those particu-lar words in that particular manner’ (loc. cit.). But even in this, Carey cannothelp but introduce a distinction between an ordering of words that is intrinsicallyof value and others (such as those in a computer manual) that are merely instru-mental in their purpose, and whose end is extrinsic.

At this early stage of the argument, Carey could retort that somewhere therecould be someone for whom the computer manual does in fact produce‘artlike’ experiences, and, hence, for them it is an artwork – they prefer readingcomputer manuals to reading Keats’ odes and, indeed, experience the samethings as do poetically inclined ode readers when reading Keats (though,according to Carey, we could never actually know this). Though it would be con-sistent with his first argument, this is not the path that Carey follows in hissecond. Instead, he hopes that his argument will ‘persuade some or all of myreaders’ (ibid., p. 174, my emphasis). The point is that Carey’s readers shouldnot be persuaded simply by reading a list of Carey’s preferences because, as hehimself notes, the simple fact that someone prefers one thing to another is initself neither here nor there. A statement of preference is a statement about theperson preferring (i.e. ‘I am the kind of person who likes this’), not about thething preferred (i.e. ‘Given that this is an instance of a thing of this kind, it isbetter than that other instance of its kind, but not as good as this one of itskind’; in the way that hunters in a hunter-gather society might say, ‘This is a par-ticular spear and, given how we understand spears to be the kind of thing theyare, we judge that it is a better spear than this one, but not as good as that one’).


Carey’s preferences in life might be relevant to that small circle of people whobuy him birthday presents, but not to the majority of his readers who don’tactually know him – our interest is in his judgements. The ‘subjectivity’ in pro-ducing his argument in support of literature is not simply, ‘I just happen to preferit’, but, ‘I judge it to support my argument’. If Carey’s wish is to ‘persuade’ hisreaders, then he does so by producing reasoned judgements, not merely by tellingus about his preferences (i.e. telling us things about him, Carey, as opposed toproviding us with good reasons why we might come to share his judgements). Andthis is precisely Carey’s own argument for the superiority of literature.

The mind outside itself – the possibility of verticality

A central dimension of Carey’s argument throughout the book is that we cannotcompare artworks because we cannot compare the experiences that people haveof them because we cannot access the minds of others. Carey suggests, first, thatto compare artworks is invidious because it is to compare people and, secondly,that it is impossible anyway because the minds of others are closed to us. Thelatter is a peculiar argument from a teacher, critic and public intellectual. Atleast a certain aspect of Carey’s consciousness is available to us – that which isintentionally expressed in his book in order to engage our consciousness. Wemust assume that Professor Carey, as a teacher at Oxford University, madeaspects of his consciousness available in an intelligible and comprehensible wayto his students. Perhaps we are seeing again a problem of dogmatic scepticism– of course we cannot access Carey’s consciousness absolutely; not even he, headmits, can do that. But we don’t have to. We only have to access each other’sconsciousness well enough, from the most intimate to the most fleetingencounter, to maintain normal human social interaction. A problem withCarey’s extreme subjectivism is its neglect of the structured character of inter-subjectivity and its variety.

Carey quotes the metaphysical poet John Donne and goes on to say:

To take a proper, diagnostic look at your own mind, Donne reasons, you wouldneed to get outside your own mind – and that cannot be done. The instrument you mustuse to probe your mind is already bent, for it is your mind. (Carey 2005: 182, myemphasis)

This is the very heart of the problem. Emile Durkheim took the opposite viewto that of Donne and Carey. He said when defining a ‘thing’:

A thing is any object of knowledge which is not naturally controlled by the intel-lect, which cannot be adequately grasped by a simple process of mental activity. Itcan only be understood by the mind on condition that the mind goes outside itselfby means of observation and experiments, which move progressively from themore external and immediately accessible characteristics to the less visible andmore deep-lying. (Durkheim 1956: 58–9, my emphasis)


118

At one point, Carey quotes Durkheim approvingly:

Philosophers have often speculated that, beyond the bounds of human under-standing, there is a kind of universal and impersonal understanding in whichindividual minds seek to participate by mystical means; well, this kind of under-standing exists, and it exists not in any transcendental world but in this world itself. . . (Durkheim, cited in Carey 2005: 252)

Carey appears unaware, when invoking Durkheim, that the ‘kind of universaland impersonal understanding in which individual minds seek to participate bymystical means’, that Durkheim wishes to relocate in ‘this world itself’, is pre-cisely the Kantian ‘. . . mysterious realm of truth, which he [Kant] called the“supersensible substrate of nature”, where all such absolutes and universalresided’ (Carey 2005: 9).

Carey’s reductive subjectivist relativism could not be further removed fromDurkheim’s emergent historical materialism and the manner in which he trans-lates Kant’s ‘mysterious realm of truth’ into ‘this world itself’. Durkheim did thisby arguing for the sociality of knowledge and for the social origins of Kant’s tran-scendental categories (Moore 2004b: ch. 5). It is in this manner that forDurkheim the mind can go ‘outside itself’, and it is in Carey’s second argumentin defence of literature that we can develop a sense of how this is so.

In support of literatureHow, then, does Carey advance the cause of literature, and what are the broaderimplications of his second argument? The essence of his case is that:

. . . literature gives you ideas to think with. It stocks your mind. It does not indoc-trinate, because diversity, counter-argument, reappraisal and qualification are itsessence. But it supplies the materials for thought. Also, because it is the only artcapable of criticism, it encourages questioning and self-questioning.

Its function as a mind-developing agency gives it especial relevance in ourpresent culture. (Carey 2005: 208–9)

The language employed here is quite different from that of Carey’s firstargument. In the extract preceding the above, he refers to literature as ‘abranch of knowledge’ and in the above this is explicated in terms of ‘ideas tothink with’, ‘counter-argument, reappraisal and qualification’, ‘materials forthought’, ‘criticism . . . questioning and self-questioning’, ‘a mind-developingagency’. Furthermore:

Literature is not just the only art that can criticize itself, it is the only art, I wouldargue, that can criticize anything, because it is the only art capable of reasoning. Ofcourse, paintings can convey implicit criticism – Hogarth’s The Gate of Calais, say,or Ford Madox Brown’s Work. But they cannot make out a coherent critical case.They are locked in inarticulacy. Operas and films can criticize, but only becausethey steal words from literature, which allows them to enter the rational world. Whenliterature criticizes other arts its target is often their irrationality. (Carey 2005: 177,my emphases)


This passage contains the key terms: literature can criticize itself (it is reflexive),it is capable of reasoning, it can make out a coherent critical case, it is part of therational world and can target the irrationality of other arts.

Once again, Carey’s argument comes up against the problem of his priorassumption that truth claims can only ever be absolute, and he assumes that thisis the case in the natural sciences and that it is in this respect that they differradically from the humanities and social sciences:

. . . once belief in a God is removed, moral questions, like aesthetic questions,become endlessly disputable. Indeed, moral questions could be defined as ques-tions to which no answers are available. Consequently, agreement about them isnot to be expected. In this they differ from scientific or mathematical questions.Disagreement is, in other words, a necessary condition for the existence of ethicsas an area of discourse. (Carey 2005: 172)

But moral and aesthetic questions do not differ from scientific and mathemati-cal questions in this way (though they do differ in other ways) – they also are‘endlessly disputable’. It is the possibility of falsification and paradigm changethat drives knowledge in these areas. Carey appears to be wedded to a strangelyanachronistic, positivistic model of the natural sciences. He might well becorrect in claiming that ‘moral questions could be defined as questions to whichno answers are available’, but, given the absolutist way in which he defines an‘answer’, this is not the same as saying that these are questions about which nojudgements can be made.

Carey is much closer to what is really significant in these areas when he saysthat

. . . it is not being right or wrong that makes a scientist. It is respect for proof andfreedom from prejudice. (Carey 2005: 184)

What he says, here, about the scientist is as true for the critic. Certainly, whatcounts as ‘proof’ will differ in that the critic cannot draw upon empiricalevidence in quite the same way, but being able to do so does not provide thescientist with infallible truths that are absolute and certain and the critic mustprovide textual evidence to support his or her argument (as Carey, of course,does throughout his book and we take for granted that he does so in an honestand truthful and rigorous scholarly way rather than just inventing writers andquotations as it suits him). Although the mathematician, the scientist and thecritic will employ different models of ‘proof’, they do not employ differentmodels in the respecting of proof, and each must be equally free from prejudice(as far as that is ever humanly possible).

In any college bar anywhere in the world there will be groups of mathemati-cians, scientists and critics sitting around tables arguing about problems intheir respective fields. In key respects, they will be doing things differently, butin the most fundamental respect they will all be doing the same thing – pro-ducing, not necessarily agreements, but informed, rigorous judgementsgrounded in ‘respect for proof’. Crucially, ‘respect for truth’ is not simply a


120

slogan. It is an internalized value, acquired through a methodology of ‘educa-tion’ (the systematic formation of a habitus), embedded within and manifestedthrough shared, collective procedures, principles and criteria. And in doingthis they are doing something qualitatively different from the first thing they alldid when entering the bar; namely, ordering their drinks – the exercise simplyof their personal preferences. In the course of their discussions over theirdrinks, certain of these individuals might judge that they should ‘change theirminds’ on particular issues, but they could not in the same way be reasoned intochanging their preferences about what to drink. Preferences are simplypersonal, but judgements are intrinsically collective in character and presupposecollectivities of certain kinds.

In support of knowledgeCarey’s second argument can now be the starting point for the exploration ofan alternative theory of knowledge that draws upon a different understandingof the sociality of knowledge that sustains the variety of relativisms of whichCarey’s first argument is an instance. The key question is: how does the mind‘go outside itself’?

Consider Carey comparing aesthetics with ethical theory:

In aesthetics, likewise, there are no absolutes, we have to choose. Even in choosingto have no interest in the arts at all is a choice. But though preferences betweenarts, and decisions about what a work of art is, are personal choices, that does notmean they are unimportant. On the contrary, like ethical choices, they shape ourlives. Nor does it mean that they are unalterable. Just as we can be argued out ofor into moral convictions (as, for example, in cases of religious conversion), soour aesthetic preferences may change. This may be sudden and dramatic . . . Orit may be the result of gradual discovery and persuasion – a process we generally calleducation. (Carey 2005: 172, my emphases)

We see the fatal muddle concerning preferences and judgements and the wayin which that is associated with the preoccupation with absolutes. We have, onthe one hand, ‘preferences that are personal choices’, but on the other it isacknowledged that we ‘can be argued out of or into moral convictions’, butthere is no rigorous conceptual distinction between these two things in Carey’sbook. The reason being, that in the absence of absolutes in the areas of ethicsand aesthetics, then, for Carey, there is no difference.

What is important for Carey is that in both ethics and the aesthetic, ‘we haveto choose’. He is correct, but changing preferences is not the same thing asbeing ‘argued out of or into’ something. To change a preference is simply todecide to opt for one thing rather than another. Someone might decide thatthey are bored with Indian food and will start eating Chinese instead, but theyneed to tell their friends no more than ‘I just happen to feel like doing it’. If, bycontrast, someone declares that they have been persuaded (‘argued into’)changing their judgement about the relative literary merits of P.D. James andHenry James, and have come to accept that P.D. is the better writer, they wouldbe required to provide reasons – that is, to do all those things that Carey believes


make literature ‘superior’ and which he himself, of course, as a critic, does inhis book. Furthermore, there would be nothing in principle unintelligible inthis person also declaring that in the course of being persuaded in this judge-ment, they have also found that they now prefer to read Henry rather than P.D.and opt to do so. Preferences and judgements and opting and choosing aredifferent things gone about in different ways.

Carey’s first argument collapses judgements into preferences and choosing intoopting. His second argument does the opposite – judgements regulate and havepriority over preferences and we are persuaded rather than merely opting. Therecould well be an argument to the effect that, in the liberal humanist tradition,the purpose of education is to articulate preferences with judgements (and thiscould be implicit in what Carey says above), but this is a big question for anothertime.

To proceed, then, in the very different direction indicated by Carey’s secondargument and his recognition of a process of ‘gradual discovery and persuasion– a process we generally call education’. Implicit in this can only be an acknow-ledgement of the fact that in certain areas our preferences can be refined byjudgements (and in others operate independently), and that this process(‘education’) occurs when we interpolate our personal preferences into publicarenas of collective judgement (though this does not necessarily imply that inall cases we have to – we may choose to be left to our own devices as to ourchoices of football team, footwear, sexual orientation or alcoholic beverages).We interpolate our preferences and opinions in order to actively engage withexplanations and reasons in a field of ‘counter-argument, reappraisal and qual-ification’. Fields of judgement are not those of the purely personal subjectivityof preferences, of opting. They are fields of inter-subjectivity of a special kind.They are essentially interlocutory in character and their purpose is the formationof judgements according to publicly shared procedures and criteria.

Coalitions of the mindUnlike the exercise of preference, the formation of judgement is a particularkind of inter-subjectivity (of sociality) structured within specialized communitiesof interlocutory interpolation. Interlocutory in that they are grounded indialogue, and interpolationary in that these are dialogues concerning reasons andexplanations – ‘why?’ discourses: precisely those procedures that Carey identifiesas constituting the ‘superiority’ of literature among the arts. The problem thatthis presents for relativist thinkers such as Carey (in aesthetics) and Usher andEdwards (in epistemology) has been well summarized by Randall Collins:

That ideas are not rooted in individuals is hard to accept because it seems tooffend against a key epistemological point . . . It is assumed that objective truthitself depends on the existence of a pure observer or thinker, untrammelled byanything but insight into truth. The notion is that the social is necessarily adistortion, an alien intrusion in epistemology; if ideas are determined by socialinteraction, then they cannot be determined by truth. This objection comes sonaturally that it is hard to think except within this dichotomy: either there is truththat is independent of society, or truth is social and not objectively true. There are


122

two prejudices here. One is the assumption that constructing an idealised indi-vidual, outside of the social, provides a vantage point that social networks cannotprovide just as well. On the contrary: there is even more difficulty in connectingsuch a disembodied individual to the world than there is in connecting a socialgroup to the world, since a group is already to some degree extended in the worldof time and space.

The second prejudice or tacit assumption is that the criterion of truth exists infree-floating reality, along with the free-floating thinker-observer. But the veryconcept of truth has developed within social networks, and has changed with thehistory of intellectual communities. (Collins, 2000: 7–8)

The ‘prejudices’ described by Collins are those exhibited by Carey and Usherand Edwards, and are symptomatic of relativism in general – they are the preju-dices of ‘dogmatic scepticism’, of the ‘All-or-Nothing Fallacy’. The core problemis the assumption that ‘if ideas are determined by social interaction, then theycannot be determined by truth’. The irony is that this proposition, assumed dog-matically by relativists, is precisely the foundationalism that underpins the posi-tivism that they are out to refute, but which they themselves presuppose in theirsceptical conclusions (Moore 2006). In the Durkheimian tradition, Collins’argument is that truth is social in that it is the product of a distinctive form ofstructured and enduring sociality: ‘ . . . thinking consists in making “coalitions ofthe mind”, internalised from social networks, motivated by the energies of socialinteractions’ (Collins 2000: 7). These ‘coalitions of the mind’, as Collins’ workdemonstrates, as canons of various kinds, are extensive in time and space (Mooreand Maton 2002). The philosophical canons of Europe, Islam, India and China,for instance, have endured, and interacted, over many centuries, and intellectualnetworks today are global and transcultural. Collins opens his argument in amanner that indicates how things that for Carey and others like him are fatallyproblematical, can be approached in a different way:

Intellectuals are people who produce decontextualised ideas. These ideas aremeant to be true or significant apart from any locality, and apart from anyoneconcretely putting them into practice. A mathematical formula claims to be truein and of itself, whether or not it is useful, and apart from whoever believes [it].A work of literature, or of history, claims the same sort of status, insofar as it isconceived as art or scholarship: part of a realm that is higher, more valid, lessconstrained by particular circumstances of human actions than ordinary kinds ofthoughts and things. Philosophy has the peculiarity of periodically shifting its owngrounds, but always in the direction of claiming or at least seeking the standpointof greatest generality and importance. This continues to be the case when thecontent of philosophy is to assert that everything is transient, historically situated,of local value only; for the relativist statement itself is asserted as if it were valid.This is an old conundrum of the skeptical tradition, discussed at great length inHellenistic philosophy. Skeptics in attempting to avoid making assertions implic-itly stand on a meta-distinction among levels of assertion of varying force. Thisillustrates the sociological point admirably, for only the intellectual communityhas the kind of detachment from ordinary concerns in which statements of thissort are meaningful. (Collins 2000: 19)


This statement contains much of the kind of thing that Carey dislikes: ‘A workof literature, or of history, claims the same sort of status, insofar as it is conceivedas art or scholarship: part of a realm that is higher, more valid, less constrainedby particular circumstances of human actions than ordinary kinds of thoughtsand things.’ Collins also points to the unavoidably self-refuting logic of the rel-ativist position. But what is significant, here, is his view that ‘Intellectuals arepeople who produce decontextualized ideas’. And what is important is that it isthe ideas, not the intellectuals that are ‘decontextualised’. The key relationship isthat between ‘particular circumstances’ and the ‘higher realm’. It is the idea ofa ‘higher realm’ that Carey’s first argument rejects (but which is smuggled backwithin the second argument). How does Collins underpin this ‘higher realm’,and how might the way in which he does so connect with Carey’s secondargument?

Collins notes that ‘. . . if one refuses to admit anything beyond the local, onearrives at some version of scepticism or relativism; if one idealizes what happensin situations as the following of rules and uses these inferred rules as a tool forconstructing the rest of the world, one arrives at a type of idealism’ (20–1).Carey’s world of preferences as the only alternative to absolutes is the kind ofrelativistic localism described and entails the parochialism that his secondargument rejects and which, he argues, literature can take us beyond. Collins’purpose is to understand the sociality of knowledge in a way that is neitherrelativistic nor idealist.

Let us begin at the site of all action: the local situation. All events take place in ahere-and-now as concrete and particular. The perspective of micro-sociology,which analyzes the structures and dynamics of situations, is all too easily inter-preted as a focus on the individual actor or agent. But a situation is just the in-teraction of conscious human bodies, for a few hours, minutes, or evenmicro-seconds; the actor is both less than the whole situation and larger, as a unitin time which stretches across situations. The detached agent who makes eventshappen is as artificial a construction as the non-social observer, who representsthe idealized vantage point of classical epistemology. The self, the person, is moremacro than the situation (strictly speaking, the person is meso); and it is analyti-cally derivative because the self or agent is constructed by the dynamics of thesituation. (Collins 2000: 20)

The fact that all human embeddedness, consciousness and action is, in the firstinstance, local does not mean that it is nothing but local: ‘The local situation isthe starting point of the analysis, not the ending point. The micro-situation isnot the individual, but it penetrates the individual, and its consequences extendoutward through social networks to as macro a scale as one might wish’ (Collins,loc. cit.). It is through ‘extension’ (as with Bernstein’s ‘verticality’) that Collinsmoves from the particular, the local, to the ‘higher’ realm. However, he makesa crucial conceptual adjustment to the spatial metaphor:

To deny that anything exists other than the local is true in one sense, misleadingin another. It is true that nothing exists that is not thoroughly local; if it did not


124

exist locally, where possibly could it be found? But no local situation stands alone;situations surround one another in time and space. The macro-level of societyshould be conceived not as a vertical layer above the micro, as if it were in a dif-ferent place, but as the unfurling of the scroll of micro-situations. Micro-situationsare embedded in macro-patterns, which are just the ways that situations are linkedto one another; causality – agency, if you like – flows inward as well as outward.What happens here and now depends on what happened there and then. We canunderstand macro-patterns, without reifying them as if they were self-subsistingobjects, by seeing the macro as the dynamics of networks, the meshing chains oflocal encounters that I call interaction ritual chains. (Collins 2000: 21)

The sociality of judgement

Carey’s second argument is not, actually, that far removed from that of Collins(in fact, for him to intelligibly do what he does in his book presupposes a Collins-type situation). The crucial difference lies in Carey’s preoccupation withindividual subjectivism as opposed to Collins’ sociological concern with thesociality of inter-subjectivism. This difference can be illustrated by contrastingCollins above with Carey below:

Since every reader’s record of reading is different, this means that every readerbrings a new imagination to each book or poem. It also means that every readermakes new connections between texts, and puts together, in the course of time,personal networks of association. This is another way in which what we read seems tobe our creation. It seems to belong to us because we assemble our own literary canon,held together by our preferences. The networks of association we build up will notdepend on spotting allusions or echoes, though sometimes we may notice these,but on imaginative connections that may exist only for us. (Carey 2005: 242, myemphases)

As the added emphases illustrate, Carey’s view is unremittingly egocentric. Heis in a certain sense quite correct in what he says, but what he says is not allthat can be said. Carey’s description would apply only to a completely non-communicative reader who keeps his or her views entirely to themselves andassiduously avoids encountering or engaging with the views of others. Thecanon of ‘personal preference’ that exists ‘only for us’ applies only to thatdetached, asocial, contemplative self that Carey and the epistemological rela-tivists reject. In principle there is no reason why any particular individual mightnot opt to adopt this position (it is not uncommon to encounter committedindividual readers who resolutely proclaim that they do not want to ‘study’ lit-erature because it would destroy the pleasure they derive from it). It can also bethe case, however, that other readers enjoy talking to each other about what theyhave read (as the current enthusiasm for readers’ groups demonstrates). Thereis a distinctive form of sociality in which individuals come together precisely inorder to trade and exchange those ‘personal networks of association’ – in justthe way in which Carey in his book does with those who read his book. The


fundamental point in this is that these ‘networks of association’ are inter-subjective, not solipsistically or paradigmatically sealed and incommensurable.

The imaginative nexusThere are two important points here. The first is one that Carey himself makesand extensively illustrates in his book: writers operate intertextually within atradition, a canon. They write against the background of what Harold Bloom(1995) calls the ‘anxiety of influence’ – they want to be judged original, asspeaking in their own voice, and not merely repeating that which others havesaid before. This is the force that drives all modern fields of intellectual pro-duction: the quest for novelty and originality (it is, historically, their radicallydistinctive feature (Whitley 2000)). The second is that readers’ preferences flowinto each other. Carey’s ‘personal networks of association’ translate, commu-nicatively, into Collins’ ‘coalitions of minds’, into participation in enduring,structured forms of sociality: ‘The focus is on a particular kind of speech act: thecarrying out of a situation-transcending dialogue, linking past and future texts’(Collins 2000: 28, my emphasis).

The novelist Ian McEwan effectively conveys the sense of this ‘situation-transcending dialogue’ in the following:

Those who love literature rather take for granted the idea of a literary tradition.In part, it is a temporal map, a means of negotiating the centuries and the con-nections between writers. It helps to know that Shakespeare preceded Keats whopreceded Wilfred Owen because lines of influence might be traced. And, in part,a tradition implies a hierarchy, a canon; most conventionally it has Shakespearedominant, like a lonely figurine on top of a wedding cake, and all the otherwriters arranged on descending tiers. In recent years, the canon has beenattacked for being too male, too middle class, too Eurocentric; what remainsuntouched is the value of a canon itself: clearly, if it did not exist, it could not bechallenged.

But above all, a literary tradition implies an active historical sense of the past,living in and shaping the present. And reciprocally, a work of literature producednow infinitesimally shifts our understanding of what has gone before . . . Ideally,having read our contemporaries, we return to re-read the dead poets with freshunderstanding. In a living artistic tradition, the dead never quite lie down.(McEwan 2006)

As this suggests, what Carey calls ‘the imaginative nexus’ (Carey 2005: 243) isnot simply an aspect of individual consciousness, but in the Durkheimian sense,of a collective consciousness extended in time and space, and that embraces bothreader and writer – ‘the unfurling of the scroll of micro-situations’. Accordingto Collins, although all intellectual exchanges are grounded in the local:

Texts do not merely transcend the immediate particulars of the here-and-now andpush toward abstraction and generality. To be oriented towards the writings ofintellectuals is to be conscious of the community itself, stretching both backwardsand forwards in time. Intellectual events in the present – lectures, debates,


126

discussions – take place against an explicit backdrop of past texts, whetherbuilding upon them or critiquing them. Intellectuals are peculiarly conscious oftheir predecessors. And their own productions are directed toward unseen audi-ences. Even when they lecture to an immediate group, perhaps of personalstudents, disciples or colleagues, the message is implicitly part of an ongoingchain, which will be further repeated, discussed, or augmented in the future.(Collins 2000: 27)

Carey tells us that the ‘essence’ of literature is: ‘diversity, counter-argument,reappraisal and qualification . . . questioning and self-questioning’. But thesethings entail community, they entail, as Collins says, ‘a particular kind of speech act’– the interpolation of ‘why?’ within a collective arena.

The problem in Carey’s book is not with where he ends up, but with wherehe begins. The problem being that he cannot actually get to the one via theother – he needs to observe the ancient advice to the traveller asking directionsin a strange land: ‘Don’t start from here’. Carey, in his first argument, wishes toavoid snubbing people by making cultural judgements and retreats into acritical diffidence that says: in matters of literature, your feelings are just as goodas mine (the professor of English Literature at Oxford University). But it is notCarey’s feelings that interest us, it is his judgements. However, in his secondargument, Carey advances the view that ‘Like drugs, drink and antidepressants,literature is a mind-changer and an escape, but unlike them it develops andenlarges the mind as well as changing it’ (Carey 2005: 210, my emphasis). In orderto sustain this view, he needs to start from somewhere else.

Carey says of literature: ‘Once its words are lodged in your mind they areindistinguishable from the way you think’ (Carey 2005: 245) – they contributeto the habitus that links the individual to society and the present to the livingpast. As Collins stresses, it is the activity that is important and, on that basis, thelesson of Carey’s book, perhaps, is to do as Carey does, not as he says.

Conclusion

It was stated in the Introduction that this essay is a work of sociology, not ofcriticism or aesthetic theory. It is sociological in that it takes as its focus what Ihave termed the ‘sociality of judgement’ – that is, a particular form of structuredinter-subjectivity associated with a distinctive form of activity that occurs withina distinctive type of supra-individual social arena that is extended in time andspace: a canon. Bernstein’s theory of knowledge structures allows us to beginthe modelling of the forms, principles and possibilities of such fields of symbolicproduction.

The canon enables a particular type of activity and its principles are generative– it is the arena of endless dispute. Understood in this way, Bernstein’s approachavoids the false extremes of absolutism and relativism by enabling a new focuson the sociality of judgement. Judgements are less than absolutes in that theyacknowledge their fallibility. They are more than preferences in that theysubmit themselves to historically evolved rules of collective evaluation. It is the


knowledge not the knower that counts. It would be wrong, however, to restrict thisdistinctive mode of sociality to the realms of academic scholarship, criticism andscience where it assumes its most rigorous forms in what Bernstein calls ‘hierar-chical knowledge structures with strong grammars’. Its principles extend furtherinto the public sphere and constitute the discursive rules of civil society inmodern, secular liberal democracy (Ahier et al., 2003).

More formally, a canon is the type of thing that might be called a‘Durkheimian emergent social kind’. Durkheim describes such ‘things’ in thefollowing way:

Collective representations are the product of an immense cooperation thatextends not only in space but also through time; to make them, a multitude ofdifferent minds have associated, intermixed, and combined their ideas andfeelings; long generations have accumulated their experience and knowledge. Avery special intellectuality that is infinitely richer and more complex than that ofthe individual is distilled in them. (Durkheim 1995: 15)

It is this ‘very special intellectuality’ (an enduring structured form of socialitywith generative powers, habitus) that has been the focus of this chapter.

The major work of thinkers such as Basil Bernstein and Randall Collins(among others) has opened up new avenues for the sociology of knowledge andthe promise of breaking the deadlock that has vitiated theoretical advance inthat area for so long.

References

Ahier, J. Beck, J. and Moore, R. (2003), Graduate Citizens? Issues of citizenship and highereducation. London: RoutledgeFalmer.

Alexander, J.C. (1995), Fin de Siècle Social Theory: Relativism, Reduction and the Problem ofReason. London: Verso.

Beck, J. (1999), ‘Makeover or Takeover? The strange death of educational autonomy inneo-liberal England’, in British Journal of Sociology of Education 20 (2), pp. 223–38.

Bernstein, B. (1977), Towards a Theory of Educational Transmission: Class, Codes and Control,Vol. 3 (second edn). London: Routledge & Kegan Paul.

Bernstein, B. (2000), Pedagogy, Symbolic Control and Identity: Theory, Research, Critique,(revised edition). Oxford: Rowman & Littlefield.

Bloom, H. (1995), The Western Canon: the books and schools of the ages. London: Papermac.Bourdieu, P. (1993), The Field of Cultural Production: essays on art and literature, R. Johnson

(ed.). Cambridge: Polity.Brown, J. (2001), Who Rules in Science? An opinionated guide to the wars. Cambridge, MA:

Harvard University Press.Carey, J. (2005), What Good are the Arts? London: Faber & Faber.Collins, R. (2000), The Sociology of Philosophies: A Global Theory of Intellectual Change.

Cambridge, MA: Harvard University Press.Durkheim, E. (1956), Education and Sociology. Toronto: The Free Press.Durkheim, E. (1995), The Elementary Forms of Religious Life. New York: The Free Press.Graff, G. (1992), Beyond the Culture Wars: how teaching the conflicts can revitalize American

education. New York: W.W. Norton.Hoggart, R. (1995), The Way We Live Now. London: Chatto & Windus.


128

Maton, K. (2000), ‘Languages of Legitimation: the structuring significance for intellec-tual fields of strategic knowledge claims’, in British Journal of Sociology of Education 21(2), pp. 147–67.

Maton: this volume.McEwan, I. (2006), ‘A Parallel Tradition’, in The Guardian Saturday Review, 1 April 2006,

p. 4.Moore (2004a), ‘Cultural Capital: objective probability and the cultural arbitrary’, in

British Journal of Sociology of Education 25 (4), pp. 445–56.Moore, R. (2004b), Education and Society: issues and explanations in the sociology of education.

Cambridge: Polity.Moore, R. and Muller, J. (1999), ‘The Discourse of “Voice” and the Problem of Know-

ledge and Identity in the Sociology of Education’, in British Journal of Sociology ofEducation 20 (2), pp. 189–206.

Moore, R. and Maton, K. (2002), ‘Founding the sociology of knowledge: Basil Bernstein,intellectual fields and the epistemic device’, in Morais, A., Neves, I., Davies, B. andDaniels, H. (eds), Towards a Sociology of Pedagogy: the contribution of Basil Bernstein toresearch. New York, Peter Lang.

Muller. J (2000), Reclaiming Knowledge. London: RoutledgeFalmer.Niiniluoto, I. ( 2002), Critical Scientific Realism. Oxford: Oxford University Press.Usher, R. and Edwards, R. (1994), Postmodernism and Education. London, Routledge-

Falmer.Whitley, R. (2000), The Intellectual and Social Organisation of the Sciences. Oxford: Oxford

University Press.Young, M.F.D. (2000), ‘Rescuing the sociology of knowledge from the extremes of voice

discourse: towards a new theoretical basis for the sociology of the curriculum’, inBritish Journal of Sociology of Education 21 (4), pp. 523–36.


Section 2

Fields of Discourse – Disciplines of Discourse


Basil Bernstein (1999) has described as ‘horizontal discourse’ the forms anduses of language that realize the particular form of knowledge he characterizesas ‘common-sense’ in character. Such knowledge stands in contrast to special-ized and institutional forms of knowledge accessible only via specialist or‘vertical’ forms of discourse. It is common (sense) in being accessible to all, inapplying to all and in ‘arising out of common problems of living and dying’(Bernstein 1999: 159). Following from this, the most obvious case of the con-struction of common-sense knowledge must be in an individual’s very early life,where the child learns how to manage the routines of everyday life, and buildsup the skills and information that is common to all members of that community.It is therefore in the informal pedagogic contexts and processes of the home,family and peer group that the nature of horizontal discourse may perhaps bestbe exemplified. In this chapter, I shall use data from informal conversations inthe home between parent and child to provide such exemplification with athreefold aim. One goal will be to clarify the pedagogical and linguistic natureof horizontal discourse, a second will be to show how the discourse can changeover time in the process of construing the world, while a third will be todemonstrate its essential limitations for the developing child straining to gaininformation beyond everyday experience.

The data on which I shall base my discussion come from longitudinal studiesof the language development of my own sons, Hal and Stephen. Data from theelder child, Hal, was collected up to the age of thirty months and from hisbrother (41/2 years younger) from thirty months to five years. In each case, anaturalistic methodology was adopted whereby spontaneously occurring con-versations of up to 45 minutes’ duration were audio-taped at least once a week,supplemented on a more frequent basis by the pen and paper recording ofnovel utterances, brief conversational exchanges and contextual information.At no point were the children tested; nor were situations deliberately contrivedin the process of data collection, and the children remained unaware through-out that their language was under any scrutiny. The aim was to collect materialthat reflected as accurately as possible spontaneous and unselfconscious con-versations between family members. (See Painter 1984, 1999, for further detailsand discussion of data collection methods.) It should be noted that thechildren’s family is one in which an ‘elaborated’ coding orientation predomi-nates (Bernstein 1987) so that the forms of horizontal discourse described arethose most likely to harmonize with later forms of vertical discourse. At the sametime, the data described can support the very general principle that for anychild, as language develops and new forms of knowledge are confronted, both

7 Language for learning in early childhood

Clare Painter

132

pedagogy and language use will necessarily be subject to change. Moreover it ispossible that an elaborated coding orientation will put horizontal discourseunder earlier pressure in the child’s life, and thus amplify its nature and limita-tions.

To pursue these points, the remainder of this chapter will be organized intothree sections. The first will very briefly outline the systemic-functional orienta-tion to language and learning that underpins the case-study research. Thesecond will describe and exemplify Bernstein’s notion of horizontal discourseusing data from the first case study. The third section will use data from thesecond case study to focus on problems encountered by the child in the con-tinuing construal of everyday knowledge when using the linguistic resourcesand forms of discourse typical of the first two or three years of life. It will be sug-gested that in meeting these challenges, the child’s language comes underpressure to develop in ways that provide greater potential for accessing lesscommon-sense forms of knowledge. In the process, there are indications of thelimitations of horizontal discourse for building these new forms of knowledge.

Language and learning: the centrality of dialogue

The orientation towards language and learning taken here is best summarizedby M.A.K. Halliday’s claim that there are three aspects to learning one’s mothertongue: learning language, learning through language, and learning aboutlanguage. In a classic paper, Halliday (1980/2004) emphasizes that these are infact three different facets of the same process: as a child learns the language, sohe or she construes the social and informational knowledge that is realized inlanguage, including knowledge of language itself. To study the individual’sdevelopment of language, then, is necessarily to study the development ofknowledge (construction). The notion that learning language is at the sametime a matter of learning (other things) through language clearly resonates withBernstein’s statement that ‘horizontal discourse, in its acquisition, is the majorcultural relay’(1999, p. 160). However, Halliday prefers to speak of developmentor learning rather than acquisition, since to speak of language or its usage assomething that is ‘acquired’ can suggest that it is something finite, monolithicand unchanging rather than being – as Halliday would view it – an infinitevariable, dynamic resource, constructed and maintained interactively.

The interactive construction of language is supported by psychologicalresearch into the very beginnings of life and semiotic experience. According toneonatal researchers such as Trevarthen (1992), Bateson (1975), Meltzoff andMoore (1998), and others, it is emotional empathy between mother and babythat provides the basis for all later developments. This is evident not only ininfants’ displayed innate preference for attending to people over other stimuli(Field and Fox 1985; Messer 1994), but in their pre-linguistic communicativeexchanges. We now know that well before language – indeed as early as thesecond month of life – infants exchange attention with care-givers, using gaze,facial, vocal and bodily movements in emotionally charged encounters withother persons. In these encounters, ‘complex, well-organized repertoires of


action and facial expression are displayed that appear responsive to the formand timing of adult communication’ (Murray 1998: 127). In other words thereis a ‘precise interplay of address and reply in time’ (Trevarthen 1992: 108))being achieved between infant and care-giver (see further Bateson 1975;Trevarthen 1977, 1998). It appears then that dialogic exchange is at the heart ofhuman experience (and human learning).

It is not only the human world that attracts the child’s attention of course,and after a few months the infant directs reaching and grasping movementstowards inanimate objects in the environment. Halliday (2004) has suggestedthat the attention to objects and other external phenomena creates a kind ofpuzzle for the child in that what is ‘out there’ in the world beyond impinges(when seen and touched and reacted to) on what is ‘in here’ – the world of theinfant’s consciousness. But the infant’s consciousness is a world that has alreadybegun to be set up as a social domain of ‘you and me’. Certainly there isevidence from developmental psychology to suggest that there may for a timebe a competition between the infant’s fascination with the world of objects andevents (to which reaching movements are directed) and his or her engagementwith other persons (where attention is coordinated) (Trevarthen 1987: 191).This conflict between the claims of the material and social worlds is resolvedwhen the child is able to extend the dialogic exchange to include reference tothe world beyond the interpersonal dyad. This occurs from around nine monthsof age when a child can achieve joint attention to objects or happenings in theenvironment by creating an idiosyncratic vocal or gestural sign with which toaddress the other person. This is the ‘proto-language’ phase of language devel-opment documented in Halliday (1975/2004), Painter (1984, 2005) and Torr(1997). It can be exemplified by the following two examples from Hal (H) as ababy in interaction with his mother (M) and father (F). Here and throughoutthis chapter the age of the child is given in years, months and days following thetext example.

Example 1H: (reaching to biscuit tin) amamama’ma [gloss ‘I want that, give it to me’]M: Okay you can have one. (opening lid)H: amamama’ma(M hands over biscuit) (0;9;12)

Example 2H: (playing with hairbrush, touches bristles to hair) a’daH: (Looks up at F and brandishes brush) a’da!F: (smiles at H)H: (looks across at M and holds brush aloft) a’da!M: (smiles) Yes you’ve got a brush for your hair; is that fun? (1;1;1)

We can see here that through the use of a sign, some phenomenon external tothe dialogue – some phenomenon from the ‘objective’ domain of experience –is brought into the intersubjective context of mutual address. In this wayengagement with the world beyond the speaking interactants begins to be

133LANGUAGE FOR LEARNING IN EARLY CHILDHHOOD • PAINTER

134

mediated through other people. Thus even at this proto-language stage, objec-tive reality comes to be acted on, but also reflected on, or ‘known’ throughsymbolic interaction, not through private cognitive apprehension. This meansthat the orthodox psycholinguistic position that language is ‘a system of expres-sion’ designed ‘for taking the internal, personal, private mental meanings ofindividuals and making them external and public’ (Bloom 1993: 19) is simplynot tenable. Instead we must acknowledge that symbols arise in the context ofdialogue – not as an externalizing of the private inner, as Bloom suggests, but asa way of making sure that the inner is developed socially and that the worldoutside the child comes to be known through the mediation of symbols. This isan orientation entirely in harmony with Bernstein’s emphasis on the centralityof dialogue in the construction of identity and consciousness.

With an orientation of this kind, it is evident that the child develops proto-language and then language in the course of using it to make meanings withothers. And the fact that at the same time as the system of language is built up –and the child is learning how to form questions and answers, how to indicate pastand present time, how to make causal or contrastive connections, how to refer toself, addressee and third parties through the pronoun system, and so on – at thesame time as all the various features of the grammar are being constructed, thechild is experiencing these features, not as abstract paradigms, but as ways torealize meaning on particular occasions of use with particular other persons.

This can be readily demonstrated if we skip ahead in the developmental storyto the point where the child is first learning words, and the knowledge theyinstantiate, as seen in Example 3 below:

Example 3 (H is naming different animals in a picture book about a train ride)H: Bukkersefy; nother bukkersefy.M: Butterfly; butterfly.H: Butterfy. It’s got horns. M: No, they’re not horns, they- they-H: (turns page and points at horse) It’s got horns.M: It’s got ears.H: (points at rabbit) It’s got horns. M: No, it’s a rabbit! It’s got ears; big ears.H: (points at squirrel’s tail curved behind its head) It’s got horns.M: It hasn’t; it’s got a tail; it’s a squirrel. This one’s [=cow] got horns.H: It’s got horns.M: Yeah. (turns page)H: Doggy.M: (points at train and makes choo choo train noises)H: (points at lion) Lion. (makes growly noises)M: Lion. And where’s the lion?H: (sees cat and climbs off lap) Katy, Katy. M: No, leave her; you’re reading a book with Mummy.H: Hal stroke Katy; Hal stroke Katy; stroke Katy.M: Now you stroke her gently. (1;11;18)


It is clear that Hal is here learning language. But it is equally clear that in theprocess he is doing more. He is also construing knowledge about animals andtheir body parts, he is enacting (and thus learning to engage in) a particularsocial relationship with the mother and also gaining experience in the socialprocess (or ‘genre’) of bedtime book reading. He is learning all these thingsthrough linguistic interaction at the same time as he is learning the words andstructures of the language itself. A Hallidayan perspective on language devel-opment, then, is one that sees learning through language as an inextricable partof the process of learning the language itself.

The characteristics of horizontal discourse in the first two and a half years

Bernstein describes the kind of knowledge realized in horizontal discourse asbeing typically ‘oral, local, context dependent and specific, tacit, multi-layered,contradictory across but not within contexts’ and, most crucially, ‘segmentallyorganised’ (1999: 159). In this section a further small selection of texts fromHal’s early childhood will be used to illustrate some of these points and todiscuss the kind of pedagogy and the forms of language involved. Each textcited is representative of dozens of comparable ones that share the samefeatures. In presenting examples I will focus initially on the variable nature ofpedagogy in horizontal discourse of this kind, which is said by Bernstein to varyfrom ‘segment’ to ‘segment.’

For example, there are occasions when the pedagogy is quite explicit: that isto say both parties know what is to be learned, how it is to happen, who isteacher and who is taught, and what counts as success. This is most obviouswhen the child is being instructed in a specific skill, as exemplified in Examples4a, b and c.

Example 4a (M instructs Hal how to pull elastic for cat to chase)H: Katy, Katy. (pushes a bundle of elastic tape at her)M: No, she likes it when it moves; it’s got to move, you see.H: There you are Katy. M: No, she won’t take it like that. Come here and I’ll show you. You hold this

here; hold it there; and then you dangle this bit like that; like that; andthen you say ‘Here you are Katy’.

H: Here Katy! Here Katy! M: Just throw this and wheeeee! (tosses elastic) Chase it Katy! (H runs up to cat) M: No, she’s a bit frightened of you. (2;3;25)

Example 4b (M instructing H to buckle up sandal) M: Pull up; pull it up; that’s a good boy. No, got to pull it up some more.

That’s it. Pull it hard as you can. That’s the way. Now put this [=metalprong on buckle] in. Good boy! Want to put this [=end of strap] in? Let


136

me start it and then you can push it in. That’s the way! Oh, you’re a cleverlittle boy aren’t you? (2;3;25)

Example 4c (H treads heavily on M’s toe, while Grandmother (G) observes)M: Ow!!! (crossly) Oh Hal, be careful!G: It was an accident.H: Sorry Mummy.M: Darling. (kisses)(M and G talk)H: I tread on Penny too. M: You trod on Penny. Did you say sorry to Penny?H: No I didn’t.M: Did Penny shout at you?H: No; she say ‘Ow!’M: Well, next time say ‘Sorry Penny’. (2;6;16)

While the pedgagogy is explicit in relation to the physical or social skills beingtaught in each case, it should also be noted that there is tacit learning takingplace here. Neither party is conscious for example that such conversations arealso teaching the child that the appropriate way to learn new skills is throughtalk (rather than say simply by observation), which is something that we knowvaries from one social group to another and has consequence for the transitioninto school (see Heath, 1983). What is also tacitly conveyed of course is that apet cat is a possible playmate whose feelings have to be taken into account; thatmanaging your own dressing is a praiseworthy goal and that apologies defuseanger and deflect criticism.

As well as the tacit knowledge that is gained incidentally during instances ofexplicit pedagogy, there are many occasions where the instruction is itselfimplicit: that is to say, the learner is unaware of the goal of the lesson or theroute to its achievement. Here we can consider text 5, which is typical of thosethat took place before age 2 when Hal’s father returned home from work:

Example 5 M: Tell Daddy where we went today.H: (looks at M)M: Where did we go, Hal?H: Park.M: What did we do?H: Play football in park.M: Mm, we played football; it was fun, wasn’t it? (1;8;1)

Here, through modelling, there is implicit teaching of aspects of grammar inthe mother’s elaborated final response; for example of the past tense of play andthe use of the subject pronoun we. In addition, through the use of questioningand elaborated responses Hal’s mother is also implicitly teaching him how tostructure a recount of experience. Even more importantly, such an exchange


demonstrates to the child that language can be used to convey information tosomeone who didn’t share the experience, a realization that may take some timefor a young child to appreciate. (At first, as here, the child may only understandthat it is possible to re-share in language an experience that was shared in thefirst place, rather than reporting to a third party. See Halliday 1975; Painter1984, for further discussion.) Such conversations, even without the adult beingconscious of the fact, alert the child to the potential of language for true infor-mation-giving and also to the kind of information (and its attitudinal appraisal)that it is appropriate to share when a family member returns to the home.

On still other occasions of horizontal discourse, the pedagogy may be furthersubmerged. That is to say, neither party may even be conscious of the discourseas having any informative or instructive character. Text 6 exemplifies:

Example 6 (M and F have entered kitchen with H mid-morning to make coffee)M: (to F) Kettle’s boiling.H: I want my lamb chop Mummy. M: Um it’s a little bit early for lamb chops darling; would you like to have

half a banana? H: No no no no! A big banana!M: All right; you get one from up there. (indicating fruit bowl on counter)H: Can’t.M: Yes you can.H: (moves chair to reach) There! (2;6;0)

Even though the context here appears to the interactants to be one of actionrather than teaching and learning, the contributions by the mother incidentallyconvey cultural information about when it is appropriate to eat which kind offood, as well as the fact that acting independently is valued in their social group.

Another and very different context for tacit learning is provided by playfulinteractions. Text 7 below is one of very many examples in the corpus whereparent and child collaborate in a ‘misnaming’ game, initially introduced to thechild by the adult. It too looks at first to be a poor example of a pedagogic text,but is again significant for the tacit understandings gained or rehearsed:

Example 7(Looking at picture book which displays a picture of a tiger and of anelephant)M: Which one d’you like best? The tiger or the elephant?H: (points at tiger) That’s tiger. M: (confirming) That’s the tiger.H: (points at elephant) Tiger!M: (laughs)H: No.M: That’s not a tiger; that’s a gorilla.H: No!M: It’s a giraffe.


138

H: No!M: That’s a deer.H: No!M: That’s a zebra.H: No. M: What is it?H: Some more zebra? [i.e. request to continue misnaming]M: (laughing) That’s . . . what is it?H: Pussy cat!M: Pussy cat, is it? miaow!H: No.M: Is it a donkey? H: No.M: Is it a giraffe? . . . [further exchanges omitted as they cycle through six further wrong names]M: Is it an elephant?H: Yes!M: Yes! (2;1;27)

Since the child is well aware of the words tiger and elephant and to whichanimals they refer, this is not an explicitly pedagogic text in the sense of beinginstructional or informative in the same way as Example 3. And as far as theinteractants are concerned, the object is simply playful enjoyment. But tacitly,the child is learning here about the possibility of unhinging meaning andwording. This, it can be argued, constitutes the first step towards a more con-scious understanding that a name is a social symbol and that knowledge itselfis a representation and therefore can be misrepresented.

Summary: horizontal discourse up to age two and a halfFrom these very few examples cited from the much larger corpus we can get abetter sense of what Bernstein intends by the term horizontal discourse. In par-ticular, the following characteristics are prominent.

First of all, we can see that each fragment of conversation is, in a certainsense, self-contained. That is, the skills or knowledge gained in one episodearen’t required before the learner can tackle something new. It is not necessaryfor example to have mastered shoe buckling before learning to apologize; it isnot necessary to have mastered the difference between horns and antennaebefore being able to play a misnaming game in relation to elephants and tigers.This quality of the learning encounters is a key feature of what Bernstein refersto as the ‘segmental organization’ of horizontal discourse.

Two further features associated with this segmental organization are thecontext-specific nature of the knowledge being learned, and the fact thatlearning is a matter of repeating or accumulating successive instances. Learningis context-specific due to the isolation of segments from one another and thelack of explicit provision of more generalized information. For example, Hal isinstructed in what to say when treading on someone’s toe rather than beingprovided with an account of intentional and accidental behaviour as a basis for


the application of general moral rules. At the same time, the segmented,context-specific organization means that a key feature of horizontal discourse isthat learning takes place by continually accumulating or repeating instances,whether it be of buckling a shoe, practising a name or relating news.

A fourth feature of horizontal discourse that Bernstein draws attention to isthat the success of any segment is inherent in the specific occasion. Whether itis a meaning or a physical accomplishment that is being essayed, the child doesnot need to be told by someone else whether s/he has succeeded or failed.Whether or not Hal received praise, he knew whether the shoe got done up orwhether the apology worked or whether he got a banana, or whether he figuredout to his satisfaction the difference between ears and horns.

Fifthly, there is the question of the tacit nature of many of the understandingsconstrued through horizontal discourse. The data show clearly that the natureof the teaching and learning may be more or less explicit on different occasions.However, it is also clear that even where the pedagogic intention is explicitlydirected at a particular goal, there will be other tacit learning going on.Moreover, it is the combination of segmental organization and tacit learningthat makes it possible for contradictory or inconsistent knowledge to be heldrelatively unproblematically. This means, for example, it can be taken forgranted in one segment that one small animal can be a playmate and familymember whose affects are important, and in another segment that anotheranimal can be a commodity for consumption.

Finally, when it comes to a consideration of the language drawn on in hori-zontal discourse at this time, two additional and related features are important:one is the context-bound nature of the language, and the other is that themeanings all pertain to concrete experience. The characteristic of context-dependence relates to the fact that the language of a child under three willgenerally either accompany/request action in the immediate context (e.g. texts4a, 4b and 6), provide a commentary on the shared visual context (e.g. texts 3and 7), or reconstruct a recalled context (as in texts 4c and 5). The meaningsare ‘concrete’ in that they relate to the observable and tangible rather than thegeneral or abstract. As will be shown in the next section, these are limitationsthat begin to be transcended after about age 3.

All the general characteristics of horizontal discourse summarized above havebeen observed by Bernstein (1999), and are likely to be common features of anychild’s early language use. What is worth emphasizing though is that despite thesegmental nature of horizontal discourse, there are aspects which cut acrossthe boundaries of segments. I would suggest that the more tacit aspects –like the value of language to mediate physical learning (as seen in e.g. 4a and4b), the potential to unhinge meaning and form (apparent in the misnaminggame), and the value of being linguistically explicit and of conveying unsharedinformation to others – are actually relevant across different segments. And it isthese particular tacit aspects that may be specific features of elaborated codeusers. So even at this early stage, there are likely to be differences in the natureof horizontal discourse in different social groups.


140

Problems and resolutions: developments in the linguistic system and itsuse after age three

So far, I have exemplified the characteristics of horizontal discourse typical of achild’s use of language up to the age of about 3 years. Now I wish to considerthe limitations of this kind of discourse for construing knowledge. One way toaddress this is to look at examples of talk where the child appears to strugglewith meaning or to get things wrong. This throws light on the limitations of thelanguage and also the aspects of knowledge that prove problematic. In thissection, therefore, I shall identify moments from the second case study whereStephen has difficulty with construing common-sense knowledge, and outlinethe new developments in his linguistic system and its use that follow from these.(In these examples, because the child seeks help in construing his world, thepedagogy is more consistently of a relatively explicit kind, although of coursemany other conversations had a different character.) The developmentsdescribed involve on the one hand a move away from the necessity of context-dependence and specificity, and on the other a movement towards moreabstract meanings and a greater reliance on attention to textual rather thanpurely observational information sources. New linguistic resources and newpossibilities for discourse will be shown to equip the child to gain glimpses ofeducational knowledge, but the lack of technicality, and the unsystematic andincidental character of learning through conversation, limit the child’s access tosuch knowledge.

Problem 1: Limitations of ‘context-bound’ talk and observation for building semantictaxonomiesIn Examples 8a and 8b below, Stephen experienced difficulties tracking infor-mation that was detached from the observational context, and in organizingnames of everyday things into semantic hierarchies.

Example 8a (M and S’s brother have used the term pet in their talk together)S: (having overheard) What’s a pet? M: A pet is an animal who lives in your house; Katy’s our pet.(later same day)S: What’s a pet called? (2;11;15)

Example 8a above is still very much in the realm of common-sense knowledgerelevant to the domestic sphere, but the use of language by the parent is differ-ent. It is different from other texts where things are named in being no longercontext-dependent. Instead of pointing (linguistically and physically) at some-thing directly observable in the context by saying that’s a pet, Mum offersStephen a definition (a pet is an animal who lives in your house); in other words,both terms in the relationship X = Y are constructed in a linguistic form thatholds good regardless of the context of situation. Such a form can be termed‘self-contextualized’. The difference between the two forms is set out below inFigures 7.1 and 7.2:


Because Stephen was used to learning categories through contextually sup-ported language and not by processing definitions, he had difficulty with hismother’s statement and so he returned to the issue again later in the day, askingwhat’s a pet called? At this time call was Stephen’s only metalinguistic verb and hisapparently inappropriate use of it here was doubtless an attempt to use hiscurrent linguistic resources to make the new move to talking about meaningrather than talking directly about observable reality.

We see the same problem arising in Example 8b:

Example 8b (Traffic passes in the street where S and M are walking)S: A bus isn’t a truck, is it Mum?M: No.S: What is it?M: A bus is for carrying people.S: No, what is a bus called? I said, a bus, what is it called?M: It’s called a bus; that thing (points) is just a bus and it’s –S: (plaintively) Oh, why is it just a bus? (3;6;25)

In Example 8b, Stephen wanted a superordinate term in order to betterunderstand the relationship between buses and trucks, but his mother failedto recognize this. As his enquiry was about names and categories, Stephenattempted to reframe it by using his only metalinguistic verb (in what is itcalled?), but failed to make himself understood.

From grappling with these problems of categorization, Stephen was impelledto develop his own language further, and to learn in a less context-bound mode.He took up the metalinguistic verb mean and the use of self-contextualizeddefining and classifying clauses, together with reference to generic categories.In Examples 9 and 10 below, it is apparent that using this kind of language hasadvantages for extending knowledge both more quickly, and beyond here-and-


that is a pet

reference to material entity relator name/category

Figure 7.1 Context-bound language for categorizing

a pet is an animal who lives in your house

name/category relator category

Figure 7.2 ‘Self-contextualised’ language for categorizing

142

now experience: it shortcuts learning by discussing meanings rather than justtangible experiential reality. (In doing so, learning language and learningthrough language become also a matter of learning about language.)

Development: metalanguage

Example 9 (S looks up to window on hearing noise of gale blowing)S: It’s blowing and blowing out there. Can you hear it?M: Yes, I can.S: Hear it isn’t mean a talk is it? Means listen to it, isn’t it? Means listen to

it, isn’t it? It’s two words, two words. (3;4;5)

By focusing on synonymy and antonymy relations between the words hear andtalk, or hear and listen to, Stephen in Text 9 could address meaning outside theconfines of a specific context. At the same time, the meanings at stake are stillvery much of the world of lived experience and the everyday.

Development: generic reference and ‘self-contextualized’ defining and classifying clauses

In Example 10 below, Stephen shows a new ability to refer to generic categories(e.g. dogs), and can also manage defining and classifying clauses so as to refur-bish a taxonomy without needing to see and refer to any actual specific newinstances. (Here and elsewhere, uncertain transcription is indicated by (? ).)

Example 10 (M asks S if he knows a word in the book they are looking at)S: No.M: It’s an animal.S: Rabbit?M: No, it’s [the word] dog.S: Dog’s not an animal!M: Yes it is . . . [further talk omitted] What is it then?S: It’s, it’s just a dog.M: Yes, but dogs are animals.S: No, they aren’t.M: Well, what’s an animal then?S: Um (?a) giraffe’s an animal.M: Oh, I see, you think animal is only for zoo animals.S: Yeah.M: Dogs are animals too, they’re tame animals. And cats, cats are animals

too. Did you know that?H: (chipping in) And people, we’re animals.S: We’re not. (3;8;1)

This text, as it progresses, is self-contextualized; that is, the language itself doesnot depend on reference to the immediate observable context of situation. It


construes categories linguistically and then refers to those linguistic categoriesrather than to material reality. These linguistic developments are important forsolving the problem of building taxonomies through everyday talk, and enablea more explicit instruction in meaning by the conversational partner.

Problem 2: Observational understandings contradicted by textual information As Stephen continued to use his language for learning during his fourth year,he constantly discovered that his own understandings, gained from lived expe-rience, might be contradicted by the more knowledgeable adult. Interactionswhere this was the case oriented him to see text as the key source for learning.In excerpts 11a and 11b for example, the adult responses he received implicitlytaught him that a linguistic context provided by a conditional clause (if you bangyour head; if he was trying) carries more weight than the observable here and nowcontext of the talk:

Example 11a (M warns S that he will smash his head if he has accident on hard floor)S: No, ’cause it’s got bone in, see? (taps head)M: But bones can break.S: No, it’s hard.M: Yes, but if you bang your head on the hard floor, it can still break, the

bones can smash. (3;5;20)

Example 11b (In car at traffic lights discussing sports cars)F: And they [i.e. sports cars] go fast ’cause they’ve got a big engine.S: But that doesn’t go faster than us (pointing at car parallel to them) See?

We will go faster (as the light turns green)F: He’s not trying; if he was really trying he could go much faster than us.S: If he goes very fast he can – if he goes very fast he can beat us. (3;6;30)

In the last utterance, Stephen tries out his new understanding, mirroring theconditional clause structure modelled by the adult.

Development: privileging text over lived experienceThe fact that this kind of interaction takes the child in a new direction isapparent when we see Stephen puzzling over the status of his own observationalexperience and privileging textual information as a knowledge source. Text 12is an interesting example of this:

Example 12 (In car, S talks about seeing a spotted horse)S: You didn’t see it Mum.M: No, I didn’t. Did it have black spots?S: Yes, (?it didn’t have white spots). Really dogs have spots.M: Yes there are spotted dogs. But horses can have spots too.S: No, only dogs have spots.


144

M: But you just saw a horse. And giraffes have spots. And leopards havespots.

S: Yes. Tell me some more animals. (3;5;2)

Here Stephen appears uncertain about the status of an observed instance of aclass and expresses a desire to establish the category of spotted animals throughtalk.

Further unexpected examples of the tension between material and textualexperience took place just after Stephen’s fourth birthday, when he began to askwhether or not he could swim. This was despite the fact that he was regularly atthe beach and pool and should have known perfectly well that he could not.Excerpt 13 is from such a conversation:

Example 13 (M driving S to his pre-school)S: Mummy, can I s- Mummy can I swim by myself now?M: We’ll have to try, won’t we? You nearly can. (?You just have to) practise a

bit more.S: um I – I’m better tell Frank if he can swim or not –M: Ask Frank. (correcting)S: because, because you might be swimming when you’re five.M: (? . . . . . . ) It doesn’t really matter how old you are; it’s just whether

you’ve had some practice . . . [long further discussion on 4 or 5 different topics omitted]M: (now talking about new person at pre-school) Oh, is he a little boy?S: Yup, he’s three. I’m four. I’m bigger.M: You are, yes.S: Can I swim?M: No, not quite. You don’t swim just because you’re four.S: (?) Swim when you’re 5?M: You swim after you’ve practised – some people swim when they’re three,

some when they’re four, some when they’re five, some when they’re ten.S: Some when they’re six.M: Yeah.S: Or some when they’re seven. (4;0;8)

At the time, Stephen’s questions seemed very strange, but it is likely that he waspuzzling about his status as a swimmer because of the contradiction between hislived experience and verbal assurances he had received that he would be ableto manage swimming when he was older. The fact that he had been beingguided to orient more and more to text as a source of knowledge meant that hefelt the need to check up on this as soon as his fourth birthday arrived.

Development: reasoning from textAll the discursive features so far illustrated – the use of definitions, generic cat-egories, taxonomizing, orientating to text – come together in very occasionalexamples where Stephen shows the ability to use logical reasoning where he


draws an inference from a linguistically presented premise. Example 14 is sucha case:

Example 14 S: Do whales eat people?M: No.S: Can they kill people?M: I don’t think they usually do.S: Can they sometimes?M: There may be one kind of whale that can; but most whales are nice

creatures.S: They’re not creatures, Mum; they’re whales.M: Yes, creature is anything that’s alive.S: Are we creatures?M: Yeah.S: (laughing) No, we’re not! (4;4;10)

This text is a far cry from those produced by Hal in his second and third yearand described as fitting the descriptors of horizontal discourse. Here thelanguage is no longer context-specific and context-bound, but rather evidencesreasoning from linguistic propositions as Stephen draws the necessary (but tohim surprising) conclusion that people must be creatures given that a creatureis anything that’s alive. This text brings together many features that constituteone major strand in the development and use of language for learning after agethree. These features include generic categories, definitions, the use of self-con-textualizing language and a cognitive orientation to attending to text.

Problem 3: The inaccessibility of (proto-)grammatical metaphor and abstractionA second major strand of development after age three is the move towards lessconcrete meanings, which in turn builds on the less context-dependentlanguage that has already been discussed. In particular, the ability to usedefining clauses helped Stephen to begin understanding an early form of whatHalliday (1994) refers to as ‘grammatical metaphor’ (Taverniers 2003). Gram-matical metaphor most frequently occurs when a meaning that is congruentlyexpressed as a verb (as with laugh or he laughed) is instead expressed by a noun(as with laughter). When this kind of ‘nominalization’ happens, the action (here‘laugh’) takes on the quality of a thing, since nouns congruently name things orparticipants in processes. Occasionally coming up against such grammaticalmetaphor was a third source of problems for Stephen as he attempted to makesense of his world through talk.

As well as nominal expressions relating to time, such as week or year, a partic-ular expression that caused difficulty for Stephen was the term traffic jam. Wecan understand this by considering both the congruent and (grammatically)metaphorical ways this meaning might be expressed, as shown in Figures 7.3and 7.4 below:


146

Examples 15a, b and c below illustrate Stephen’s difficulty at age 31/2 inmaking sense of the term traffic jam.

Example 15a (S and M in car waiting to turn right at traffic lights)S: Is this a traffic jam?M: No, it’s (pause) just waiting at the lights.S: No, is this called a traffic jam? (3;5;8)

Example 15b (S and M in car stationary at red traffic light)S: Is this a traffic jam, Mum?M: Well, (pause)S: Is this a traffic jam?M: No, not really. (3;5;13)

Example 15c (S and M in car stationary at red traffic light)S: Is this a traffic jam?M: (weary of the question) Oh, it’s a little jam.S: No, it’s a big jam, a big traffic jam; there’s all cars (pause). Is it a traffic

jam? (3;5;14)

Development: steps towards abstraction These examples indicate the difficulty of interpreting an expression like trafficjam purely by reference to the perceptual context. Stephen attempts repeatedlyto test out whether a particular instance of experience constitutes an example


Example This traffic is jamming up close together

Meaning concrete things action manner

Grammaticalfunction Participant Process Manner

Figure 7.3 Grammatically congruent expression of meaning

Example This is a traffic jam

Meaningobservablecontext action and manner

Grammaticalfunction Participant = Participant

Figure 7.4 Defining clause with a (grammatically) metaphorical participant

of the category he is grappling with. But it proves an ineffective way of under-standing the concept. Indeed, it is only when he receives a definition of sorts(see Example 15d below) that he makes any progress:

Example 15d (M explains that now they are in a real traffic jam, because when the lightgoes green they still can’t go because there are too many cars. She then con-tinues explaining)M: A traffic jam is when you can’t go even when the light is green.(The car then stops again as the light ahead goes red)S: (? ) It means when it’s green it’s called a traffic jam, but this is not a traffic

jam cause it’s red; a green one is a traffic jam. (3;6;23)

The definition Stephen is offered involves another instance of grammaticalmetaphor because a complex of meaning involving different processes (whenyou can’t go, even when the light is green) is construed as a single participant in thegrammatical structure, as shown in Figure 7.5, below:

For a contemporary urban child, the phenomenon of a traffic jam doubtlessconstitutes common-sense knowledge, and it is certainly related to his everydayexperience of travel. At the same time, as far as the language is concerned, somedegree of grammatical metaphor1 and therefore a movement towards abstrac-tion is involved.

The construal of abstract concepts is not what we expect of horizontal dis-course, but it is a necessary development for the achievement of literacy andother forms of school(ed) knowledge. As Halliday explains:

The patterns of writing create systemic properties which are then named asabstract objects, like the beginning and end of a page or a line, spaces betweenwords, and letters, capital (or big) and small. (Halliday 1996: 342)

In the following text we can see Stephen attempting to come to grips with thenumber system rather than the writing system, where a similar kind of abstrac-tion is at play:


Example A traffic jam is when you can’t go, even when the light is green

Meaningaction andmanner = action of participant + conceded condition

Grammaticalfunction Participant = Participant

Figure 7.5 A definition involving grammatical metaphor

148

Example 16 (S overhears father (F) mention the word fifty)S: Is fifty a number? F: Yeah.S: How does it go?F: It comes after forty-nine.S: A hundred comes after forty-nine.F: A hundred comes after ninety-nine. (3;11;7)

Abstraction is involved because although one can point to a material inscriptionof 50 on the page and name it as an object with that’s 50, Stephen treats ‘anumber’ as an abstract category with no material referent when he says is fifty anumber. A form of grammatical metaphor is also evident in the use of come afterto signal a relationship in a series rather than a material action of movement.

These examples (15 a, b, c, d and 16) indicate that when children likeStephen come into contact with grammatically metaphorical expressions andespecially with the semiotic forms of educational knowledge, familiar forms oflanguage and the ways these have typically manifested in horizontal discourseprove inadequate, and the newer resources of self-contextualized talk are calledinto play. Through these means the child can then begin to reflect on theabstract forms of literacy and numeracy.

The problem of building systematic knowledge It is not only in dealing with non-congruent grammatical expressions and theabstractions of numeracy/literacy that the limitations of horizontal discoursebecome apparent. Stephen’s parents sometimes struggled to respond to hisquestions about everyday phenomena, such as the function of power points inthe house:

Example 17 (M warns S not to play with wall socket)M: It’s a switch for electricity.S: (?But what is it?)M: It’s for electricity.S: But I don’t know what it looks like.M: No, well, you can’t see it; the electricity makes things go.

In this case the invisibility of the named ‘thing’ (i.e. electricity) is problematicdue to its verbalization being a nominalized form, but also because a satisfactoryexplanation seems to call for a groundwork of concepts the child has not yetdeveloped. Stephen was able to use electricity in his everyday life, switching onlights and televisions etc., but not to understand it more ‘theoretically’ in termsof a definition.

Similarly, the following conversation created puzzlement for Stephen:


Example 18 (S and M are returning barefoot from the beach. M guides him off the scorch-ing bitumen) M: Walk on the white bit [i.e. the pale paving slabs]; the white is always so

much cooler than the black [i.e. bitumen].S: Why?M: Because when the sun shines down on the black, the black keeps the hot,

but when it shines on the white, the white – the white throws it off.S: (after a pause) You can’t feel it throw off.M: What?S: You can’t feel it throw the hot off.M: No, no you can’t.

The adult here grapples with the difficulty of explaining matters without usingnominalized or technical words, compromising with the black keeps the hot . . . thewhite throws it off, while the child tries to make sense of this by attempting toobserve the phenomenon that has been described. While he can feel for himselfthe difference in temperature, he can only glimpse the nature of explanation,which draws on uncommon-sense knowledge.

A further series of examples of the limitations of horizontal discourse foraccessing new knowledge can come from one area of consistent interest toStephen, that of relationships within the animal kingdom. He engaged in manyconversations discussing the eating habits of animals and humans which essen-tially involved information about food chains and food webs. In the attempt toextend knowledge into such a domain, the segmental nature of the discourseand the fragmented, unsystematic nature of the learning proved to be a con-straint. Texts 19–22 demonstrate well the nature of pedagogic conversations inStephen’s fifth year.

Example 19 below focuses on the eating habits of humans:

Example 19 (S, H and M are reading a picture book where a boa constrictor has squeezedits prey)S: And what’s this? (pointing at mangled fawn)M: That’s a deer.S: Oh! (upset) I wish a snake weren’t coming and eat it.H: You know the – you eat animals.S: No I don’t!H: You do. You eat – fish, don’t you? Salmon.S: Yeah, but just – (? . . . the little) fishes at the beach but – but – but I don’t

eat crocodiles, (M laughs) and that’s – that’s a animal.. . . [long discussion ensues about animals that people eat, including frogs]

S: Yeah but what – we don’t – how can we eat the skeleton?[conversation continues for several minutes about the bones of various foods:frog, chicken, salmon]S: (loudly, touching his back) This is backbone.M: Your backbone, yes . . . (4;0;2)


150

This snippet can be related to a number of others that took place over suc-ceeding months and which address questions of animal anatomy, particularlythe function and appearance of bones. Stephen initiated such conversationswith questions such as Mummy, how can your bones move? Mummy how can you movewhen something’s hard inside you? How can something move your bones? Why do yourgums hold your teeth a long time? and so on.

At the same time, Example 19 links to other conversations about animal pre-dation. For example, Text 20 provides more of the conversation about whalesthat was presented earlier as Example 14:

Example 20 (M and S are at home together)S: How do dolphins squirt water up their head? M: (pauses for thought)S: How do they?M: I think it’s whales that do that. S: Yeah, I meant (?that). M: I don’t really know how they do it, darling.S: Hey Mum, can dolphins eat boats? M: No.S: Why?M: Well –S: Why don’t they?M: They don’t want to eat boats; they eat fish. Are you thinking of that

movie?S: Yes.M: Oh, that was a whale, but they don’t really swallow boats.S: Do whales eat people?M: No.S: Can they kill people?M: I don’t think they usually do.S: Can they sometimes?M: There may be one kind of whale that can; but most whales are nice

creatures.S: They’re not creatures, Mum; they’re whales.M: Yes, creature is anything that’s alive.S: Are we creatures?M: Yeah.S: (laughing) No, we’re not! (4;4;5)

A few weeks later, a moment of play led to a further conversation that touchedon the same question of what animals eat and on their relations to each otherand to people (Example 21). Then, the following month, these issues arepursued at the zoo (Example 22). The questions Stephen asks in these conver-sations suggest that he is concerned to construct some hierarchy of predation aswell as to accumulate more knowledge about creatures that were not accessibleto his own perceptual explorations:


Example 21 (S leaps out at M from behind the curtain, roaring)S: Were you frightened?M: (teasingly) No, I’m not scared of lions. (grabs S) I just give them a

cuddle. (hugs S)S: (seriously) But it might scratch you. M: Mm, it might eat me up.S: Can lions eat people?M: Yes.S: Oh! Can sharks eat lions?M: Lions don’t go in the sea.S: Can lions eat sharks?M: Sharks never go in the jungle, so they can’t, they never meet.S: In the river, the lion might swim in the river.M: But sharks don’t go in the rivers, they live in salty water – in the sea.

Anyway, I don’t think lions swim in the rivers much; tigers do, but I don’tknow about lions.

S: That’s what I’m talking about – tigers.M: Oh, we’re talking about tigers now are we?S: Why don’t lions swim?M: I don’t know; maybe they do swim, but I don’t think they like the water.

(4;4;25)

Example 22 (The family is at the zoo watching a hippo)S: Does every animal eat hay?M: No. S: Horses do. They can’t eat people, can they? . . . [further talk omitted](In another part of the zoo, F lifts S to see the fox)S: Can foxes eat people? Can foxes eat people, Mummy?M: Oh no; they eat chickens sometimes. (4;5;16)

I would suggest that although these texts exemplify horizontal discourse insome respects, they also show its limitations as a pedagogical discourse when theknowledge at stake edges into the realm of the un-common-sense. It is true thatsome of the individual pieces of information contained in these texts are inprinciple accessible to observation and lived experience, and might thereforecount as ‘common’ knowledge. Moreover the discourse is segmental in that thecontexts in which these texts arise are varied – initiated by a movie narrative, apeepbo game, a visit to the zoo, looking at a picture, respectively – and in thefact that information is not systematically sequenced in its presentation norfinalized on any one occasion. Indeed Stephen’s continual return to similarthemes indicates his desire to engage in further ‘goes’ at meaning so as to betterintegrate his knowledge fragments. In certain respects, then, the talk has conti-nuities with what has gone before.

However, just as contact with the semiotic systems of school(ed) knowledge


152

(numbers and written words) creates pressures on the informal pedagogy of thehome and stretches the child’s language towards greater abstraction, so I wouldargue that the attempt to construe knowledge whose basis is less accessible toobservation does the same. When we consider the language of Texts 19–22, wecan observe that it exemplifies generalizations, is largely independent of theimmediate context, and may even use logical reasoning from a textual premise.The language has thus moved quite a long way from the context-specific anddependent language said to characterize horizontal discourse. Yet even withthese more developed linguistic resources, the knowledge is somewhat elusivesince it is being gained in a piecemeal fashion that does not allow all the frag-ments to be brought into a relationship with one another.

It is enlightening here to look at how at these topics get expressed in a moreovertly ‘vertical’ discourse. Any selection of texts from the World Wide Web onthe topics of food chains and food webs reveals both similarities and differencesbetween the conversational Examples 19–22. One such text explains a foodchain as shown in Example 23:

Example 23 FOOD CHAINS AND FOOD WEBS ‘In the living world, every form of life is food for another. Food chains andwebs show how food and energy are passed between species.’FOOD CHAIN‘A food chain is a food pathway that links different species in a community . . . ’

(The Balance of Nature, accessed 12/12/05)

Already it is obvious that the language here is self-contextualized, involves defi-nitions of terms, and uses generalizations, metaphor and abstraction (e.g. foodchain, food pathway). The text continues with explanations of ‘food web’,‘trophic level’, ‘producers’, ‘consumers’ and subcategories of these provided inrapid succession. The information in such a text thus involves explicit tax-onomies of concepts, systematicity, rapidity and comprehensiveness.

Accessing such a text depends on the kind of language that Stephen has beendeveloping in response to problems encountered in the process of construingcommon-sense knowledge, and in beginning to encounter more specializedforms of knowledge. Without the ability to learn from definitions, to managegeneric categories, to attend closely to text, and to handle abstraction, themeaning of such a text cannot be unpacked. Even so, the differences from thediscourse of the home are striking. First there is the speed at which newconcepts are provided and built on, something very different from the child’sconversational experience; secondly, there is the technicality that is a necessarypart of organizing the knowledge in a systematic way. Such texts may alsocombine technical language with a visual display of some of the relationshipsbeing systematized. This again achieves something different from the talk in thehome – it enables information that occurred at different points in the oral texts(but was never put together there) to be brought into explicit relationships andintegrated into a coherent whole.

Thus in the process of building up their everyday knowledge, children like


Stephen will engage in talk that is not restricted to the context-specific, context-dependent forms of horizontal discourse but will still not be able to build ahierarchical knowledge structure without a more radical departure from thesegmentally organized and unsystematic conversational pedagogy of the home.

Conclusion

My object in this chapter has been to bring into clearer focus the nature of thelanguage used for learning in early childhood, demonstrating the close relationbetween ‘learning language’ and ‘learning through language’, while simultan-eously clarifying the nature, limitations and potential of children’s experiencewith ‘horizontal discourse’. Evidence from the two case studies indicates thatBernstein’s characterization of this discourse is an apt one for the first few yearsof life and for much of the learning that takes place before formal educationbegins. However, this should not obscure trans-segmental aspects of the tacitlearning that occur from the start, nor the significance of developments thatgradually take place under the pressure of making sense of everyday experienceand in response to the child’s own inclination to extend his knowledge throughtalk – developments that change the character of the talk in significant ways.

I have provided examples of how one child grappled with a range ofproblems in meaning between the ages of 3 and 5 years. The first of these ledto an increasing ability for self-contextualized talk. The problem of buildingdeeper semantic taxonomies when limited to context-dependent classificationswas met by making use of everyday metalanguage (e.g. hear it means listen to it)and by using classifying and defining clauses which make no direct referenceout to the immediate context (e.g. what’s a pet?, (a) giraffe’s an animal), togetherwith the construal of generic categories (dogs have spots). All of these develop-ments enabled the discussion of meanings without reference to observedinstances of the phenomena under attention. This in turn shortcut the gradualprocess of building taxonomic relations and allowed for a single text to set upnew relations or revise an existing classification (M: Cats are animals too, did youknow that?). In these ways the existing roles of language for accompanying, mon-itoring and reconstructing specific contexts could be extended to generalizingacross contexts. At the same time, the possibility of creating a context linguisti-cally through a hypothetical scenario (if he goes very fast he can beat us; if you bangyour head . . . it can still break) was made salient in talk within Stephen’s family,perhaps due to the elaborated coding orientation favoured, and alerted him toanother problem: conflicts between material and semiotic experience. A privi-leging of the latter enables exploratory talk and logical reasoning, and thus hasthe potential to further facilitate the building of knowledge beyond observedand lived experience. Paying close attention to text is in fact important both forbuilding common-sense knowledge more quickly and for later learning throughthe written medium.

The other ‘meaning problems’ encountered by Stephen related to forms oflanguage or knowledge that are less common-sense. On the one hand therewere occasional encounters with nominalized language, both within the realm


154

of the everyday (traffic jam) and in the exploration of other semiotics (number).These led to brief forays into the abstract and metaphorical language that ismore characteristic of vertical discourse. On the other hand, there was thedifficulty of satisfying his curiosity about matters inaccessible to observation(what’s electricity? why is the white [ground] always cooler than the black? how can yourbones move?), as well as difficulty in integrating different pieces of less problem-atic information (e.g. that some animals eat plants, others eat other animals,humans eat animals, but not all animals eat humans, etc.) These problems werenot readily overcome because their resolution requires further development oflanguage through forms of pedagogy specifically designed to build knowledgein a systematic and coherently sequenced manner. While all the linguistic devel-opments outlined above will eventually help the child to engage in earnest withknowledge beyond the everyday, they can only be capitalized on in a limited wayin the fragmented pedagogic encounters of everyday life. The systematic organ-ization of knowledge and the need to master it in a particular sequence so thatone concept can build on and relate to another is typical of certain kinds of‘educational’ knowledge. It calls for a different form of discourse (e.g. Text 23)and for forms of pedagogy that will help the learner to access it. The efficacy ofhorizontal discourse for creating our common and often tacit understandingsshould not be assumed to make it well designed for the more conscious creationof the kind of knowledge to which Stephen appeared to be striving in hisfifth year.

References

The Balance of Nature: Food Chains and Food Webs, www.cas.psu.edu/DOCS/WEBCOURSE/WETLAND/WET1/balnat.html

Bateson, M.C. (1975), ‘Mother-infant exchanges: the epigenesis of the conversationalinteraction’, in D. Aaronson and R.W. Reiber (eds), Developmental psycholinguistics andcommunication disorders; Annals of the New York Academy of Sciences, 263, pp. 101–13.

Bernstein, B. (1987), ‘Elaborated and resticted codes: an overview 1958–85’, in U. Ammon,N. Dittmar and K.J. Mattheier (eds), Sociolinguistics: An International Handbook of theScience of Society Vol. 1. Berlin: De Gruyter.

Bernstein, B. (1999), ‘Vertical and horizontal discourse: an essay’, in British Journal of Soci-ology of Education 20.2, pp. 157–73.

Bloom, L. (1993), The Transition from infancy to Language: Acquiring the Power of Expression.Cambridge: Cambridge University Press.

Field, T.M. and Fox, N.A. (eds) (1985), Social Perception in Infants. Norwood, NJ: Ablex.Halliday, M.A.K. (1975/2004), Learning How to Mean. London: Arnold. Reprinted in his Collected Works Vol. 4. London: Continuum, 2004.Halliday, M.A.K. (1980/2004), ‘Three aspects of children’s language development:

learning language, learning through language and learning about language’, in Y.M.Goodman, K.K. Haussler and D.S. Strickland (eds), Oral and Written Language Devel-opment: Impact on Schools (Proceedings from the 1979–80 Impact Conferences). Inter-national Reading Association and National Council of Teachers of English, pp. 7–19.

Reprinted in his Collected Works Vol. 4: The Language of Early Childhood. London: Contin-uum, 2004, pp. 308–26.

Halliday, M.A.K. (1994), Introduction to Functional Grammar, 2nd edn. London: Arnold.


www.cas.psu.edu/DOCS/WEBCOURSE/WETLAND/WET1/balnat.html

www.cas.psu.edu/DOCS/WEBCOURSE/WETLAND/WET1/balnat.html

Halliday, M.A.K. (1996), ‘Literacy and linguistics: a functional perspective’, in R. Hasanand G. Williams (eds), Literacy in Society. London: Longman, pp. 339–76.

Halliday, M.A.K. (2004), ‘On grammar as the driving force from primary to higher-orderconsciousness’, in G. Williams and A. Lukin (eds), The Development of Language: Func-tional Perspectives on Species and Individuals. London: Continuum, pp. 15–44.

Heath, S.B. (1983), Ways with Words: Language, Life and Work in Communities and Classrooms.Camridge: Cambridge University Press.

Meltzoff, A.N. and Moore, M.K. (1998), ‘Infant intersubjectivity: imitation, identity andintention’, in S. Braten (ed.), Intersubjective Communication and Emotion in EarlyOntogeny. Cambridge: Cambridge University Press, pp. 47–62.

Messer, D. (1994), The Development of Communication: From Social Interaction to Language.Chichester: Wiley.

Murray, L. (1998), ‘Contributions of experimental and clinical perturbations of mother-infant communication to the understanding of infant intersubjectivity’, in S. Braten(ed.), Intersubjective Communication and Emotion in Early Ontogeny. Cambridge:Cambridge University Press, pp. 127–43.

Painter, C. (1984), Into the Mother Tongue. London: Pinter.Painter, C. (1999), Learning through Language in Early Childhood. London and New York:

Continuum.Painter, C. (2005), ‘The concept of protolanguage in language development’, Linguistics

and the Human Sciences 1.2, pp. 177–96.Taverniers, M. (2003), ‘Grammatical metaphor in SFL: a historiography of the introduc-

tion and initial study of the concept’, in A-M. Simon-Vandenbergen, M. Taverniers,L. Ravelli (eds), Grammatical Metaphor: Views from Systemic Functional Linguistics.Amsterdam and Philadelphia: John Benjamins, pp. 5–33.

Torr, J. (1997), From Child Tongue to Mother Tongue: A Case Study of Language Development inthe First Two and a Half Years. Nottingham: Dept of English Studies, University ofNottingham.

Trevarthen, C. (1977), ‘Descriptive analyses of infant communicative behavior’, inH.R. Schaffer (ed.), Studies in Mother-Infant Interaction. London: Academic, pp.227–70.

Trevarthen, C. (1987), ‘Sharing makes sense: intersubjectivity and the making of aninfant’s meaning’, in R. Steele and T. Threadgold (eds), Language Topics: Essays inHonour of Michael Halliday Vol. 1. Amsterdam and Philadelphia: John Benjamins,pp. 177–99.

Trevarthen, C. (1992), ‘An infant’s motives for speaking and thinking in the culture’, inA.H. Wood (ed.), The Dialogical Alternative: Towards a Theory of Language and Mind.Oslo: Scandinavian University Press, pp. 99–137.

Trevarthen, C. (1998), ‘The concept and foundations of infant intersubjectivity’, inS. Braten (ed.), Intersubjective Communication and Emotion in Early Ontogeny. Cam-bridge: Cambridge University Press, pp. 15–46.

Notes

1 This is a ‘proto’ grammatical metaphor in that it is only a single isolated routinizedexpression, rather than a more creative manipulation of the grammar in the courseof organizing a complex text field of knowledge.


Introduction

Subject English is just over a century old. The first university chairs in Englishwere created in the second half of the nineteenth century, while by the turn ofthe century, subject English (as distinct from discrete areas of knowledge, suchas ‘spelling’ or ‘reading’, Christie 2004a), had appeared in the programmes ofthe elementary schools. As the twentieth century unfolded, English was tobecome the most important subject in the school curriculum, its status rivalledonly by the claims of mathematics, though arguments about both the contentsof mathematics and its pedagogy have never rivalled those about subjectEnglish.2 In fact, subject English has always been a highly contested site in theschool curriculum, not least because discussion about education in the nationallanguage has been intimately bound up with discussion of matters to do with thenational psyche and identity, as well as with notions of the economic and socialgood of English-speaking countries. It was, for example, no accident that in theaftermath of the First World War, the English held their first inquiry into theteaching of English (the Newbolt Report, 1921).3 The war had exposed problemsto do with the literacy skills of the armed forces and there was an interest inimproving citizens’ capacities to read and write.

In the post-war period, secondary education was extended, especially in theEnglish-speaking world, and this led to a renewed significance for subjectEnglish. Before the end of the twentieth century, the English held two moremajor inquiries into the teaching of English, leading to the Bullock Report (1975)and the Kingman Report (1988). The work of the latter, developed in various waysby the Cox Working Group (1989), led to the adoption of the National EnglishCurriculum (1995). Developments in Australia were different, though by the1990s several federally co-ordinated initiatives led to major policy statements onlanguage and literacy. These included a federal Government White Paper, Aus-tralia’s Language. The Australian Language and Literacy Policy (1991), a nationallyendorsed statement, English: a Statement for Australian Schools (1994), and a laternational policy statement, Literacy for all, the Challenge for Australian Schools(1998).4 Thus, over the twentieth century, subject English assumed an increas-ingly important role in schooling, attracting considerable interest at govern-mental levels and also, by the end of the century, ushering in what have becomewell-established testing regimes in the English-speaking world, intended tomonitor and report on the progress children make in mastering English literacy.

Today, so significant is subject English, success in it is now an importantpassport to many avenues of privileged life and education. Yet, ironically, given

8 Building verticality in subject English1

Frances Christie and Mary Macken-Horarik

its increased importance, the nature of English is increasingly elusive, itsmastery not available to many students. A powerful invisible pedagogy oftenapplies, such that what is evaluated as success is tacitly understood, rather thanclearly articulated. In the contemporary school curriculum, there are in factmany ‘Englishes’, even if the singular form is retained in the name of thesubject. A review of its history (e.g. Christie 1993, 2003, 2004a) would suggestthat it has been of the nature of the subject that it has proliferated many forms,or ‘languages’, each possessed of its own preoccupations, its own questions, itsown value positions.

Bernstein (2000: 155–74) drew a distinction between horizontal and hier-archical knowledge structures, where each of these, he argued, constituted oneof the two possible modalities of vertical discourse. There is no complete fitbetween Bernstein’s use of the word ‘grammar’ and that in the systemic func-tional (SF) tradition. However, we shall suggest that in Bernstein’s terms schoolsubject English has horizontal knowledge structures with ‘weak grammars’,meaning that it is constructed in a series of specialized languages, segmentallyorganized, wherein what counts as achievement is adoption of a particularpedagogic position or ‘gaze’ (Bernstein ibid.) rather than any strongly definedtheoretical position which might confer a degree of commensurability acrossthe languages. Furthermore, with the passage of time, we shall suggest, whilevarious ‘languages’ have thus emerged, the content of subject English hassteadily become more invisible, the exception being some functionalapproaches (Macken-Horarik 1998, 2002) which have been notable for theirclaims to a visible content and a visible pedagogy in all subjects, includingEnglish. We shall also argue that, despite what has been historically an overalldrift towards invisibility of content, it is possible to identify certain apparentideologies that apply in English, and that privilege certain value positions overothers. Among other things, what is at issue for the learner is tacit acquisition ofcertain values and ‘truths’ of the knowledge of English, to do with expression ofsensibilities or what we might term ‘finer feelings’. For obvious reasons, suchtacit acquisition creates a situation that advantages some students rather morethan others.

Important questions for us concern how to address problems to do with thetacit acquisition of the values of English, and how to develop a model of thediscipline that is more theoretically robust and more transparent for students.Such a model would, among other things, be internally coherent, based onwell-theorized organizing principles and articulated in a (meta)language thatallows for progression up the years of schooling. Like Muller (this volume) weshall use the term verticality to refer to the characteristics of such a model. Weshall suggest that the model of English we are seeking will be marked bygreater verticality than presently applies. That is to say, it will show an enhancedcapacity for integration of the elements of its knowledge structure, providingan expanding sense of a coherent knowledge base as students move throughtheir schooling.

This chapter has two parts, in the first of which we take up the challengeissued by Bernstein in discussing horizontal knowledge structures and theiremergence, ‘to look at changes in the development of specialized languages

157VERTICALITY IN SUBJECT ENGLISH • CHRISTIE & MACKEN-HORARIK

158

across time’ and ‘to plot the increase in the number of languages’ found(Bernstein 2000: 166). Where Bernstein suggested this for sociology, we do itfor subject English. To do this, we explore the origins of different specializedlanguages (or models) of subject English and propose that the proliferation oflanguages/models has created difficulties for students not already wellpositioned to manage successfully its increasingly tacit requirements. This isparticularly evident in contexts where students are asked – often in examin-ations – to respond to open questions about unfamiliar texts, or to writeresponses to rather general questions, and where the basis of assessors’ judge-ments is left implicit. We argue that careful attention to the semantic qualitiesof successful students’ responses in these contexts enables us to probe thelinguistic dimensions of the relevant discourse in English, thereby exposingthe criteria for successful performance which otherwise remain unclear. In thesecond part of the chapter, we develop the argument by investigating thelinguistic features of two highly valued texts produced in formal examinationsby students in England and in Australia. Drawing on metalinguistic tools avail-able in SFL, we highlight the orientations to meaning that are rewarded at thekey transitional school stages that are represented in public examinations. Weidentify the linguistic choices which realize these successful orientations andconclude by arguing the potential of SFL to increase the ‘verticality quotient’(Muller 2004) of English.

The building of the languages of English teaching

It is possible to interpret the development of school English as the unfoldinggenesis of knowledge structures (models of curriculum) from its inception inthe nineteenth century till the early twenty-first century. The nineteenth-century concern for provision of a basic education for the poorer classes led tothe emergence of Basic Skills, and this gave way with the opening of the twenti-eth century to the Cultural Heritage model, whose preoccupations were with thecivilizing influence of ‘great literature’. In the period after the Second WorldWar, two different models emerged in the same period. On the one hand therewas a strong functional tradition we have termed Functional Language Studies,and on the other hand, a strong personalist tradition, referred to as PersonalGrowth. Where the former focused on socially conceived functions of languageand proposed overt teaching of knowledge about language, the latter focusedon personal growth in learning language, and its tendency was to resist teachingof knowledge about language. While Functional Language Studies were influentialin many places, Personal Growth tended to dominate in many versions of Englishcurriculum across the western world. In the last two decades of the century intothe early twenty-first century Cultural Studies became influential, presently con-stituting what is often the dominant model of subject English at the seniorlevels. Unlike Cultural Heritage, Cultural Studies eschews the literary canon, stress-ing the value of studying popular and multimedia texts, and encouragingstudents to make multiple readings of texts. Unlike Functional Studies this tradi-tion has no overt commitment to teaching knowledge about language, and in


fact it has some affinities with Personal Growth, in its emphasis on personalresponses to life and learning. Two further models of English emerged in thelate twentieth century: Multiliteracies and New Literacy Studies. The latter studiesemphasize the ‘situated’ nature of texts and meanings, asserting the irrelevanceof much traditional language study, while Multiliteracies stress the importance ofmultimodality and sociocultural diversity in contemporary communication, andlike Cultural Studies they propose the study of a wide variety of texts. In the fol-lowing sections, we turn our attention to a more detailed account of the genesisof each model of subject English.

Basic SkillsEnglish in the last years of the nineteenth century amalgamated what had beenhitherto some ‘discrete language skills’. The teaching of reading had had a longhistory (see Davies 1973, and Christie in press), while the teaching of writing,once an aspect of the teaching of rhetoric, had also had a reasonably longhistory, though by the nineteenth century it had survived rather better inScotland than in England. English school grammar, derived originally from thestudy of Latin, had been taught at least as far back as the sixteenth century,though it was in the eighteenth century that several accounts of Englishgrammar for students were written, the most influential being that of Murray(1795/1968; Christie 1993). What made the nineteenth century significant forsubject English was that this was the century in which school education wasregularized and the state became involved in providing elementary education.After the 1830s in England, and the 1850s in the Australian colonies, steps weretaken to provide state-supported elementary schooling (Christie 1976), andteacher training became an increasingly important enterprise.

Not surprisingly, it was English language, especially literacy, skills, thatloomed large in the thinking of those who devised the elementary school pro-grammes. Large numbers of graded school readers, spelling books, grammarbooks, dictionaries, and some composition books, appeared. Theories about themost appropriate methods for teaching the various English language skillsabounded, and textbooks appeared for teacher training. One, for example, wasthat by Fitch (1880), Her Majesty’s Inspector of Schools, which focused on thevalues of teaching the English language, grammar and grammatical analysis,paraphrasing and classification of words (Fitch ibid.: 256–85). Writing in 1880,Fitch noted the need for the English programme ‘to make some reference toEnglish Literature as a branch of school instruction’, also observing that ‘this isa comparatively new ingredient introduced of late into the school course’(Fitch, ibid.: 279–80).

The ideal pedagogic subject position associated with the nineteenth-centuryelementary school involved someone with basic literacy competence. Consider-able importance attached to learning to read (Gosden 1969; Goldstrom 1972;Cathcart Borer 1976; Christie 1976), for this was part of acquiring a little useful‘general knowledge’, including knowledge of the Bible, while it was also held tocreate sober, upright workers and citizens. The nineteenth-century elementaryschools brought a degree of ‘social governance and spiritual guidance’ to thechildren of the working class and the poor, and this regulatory function was


160

central to the creation of social order by emergent nation-states of the nine-teenth century (Hunter, 1994). The operation of what Bernstein later called‘the pedagogic device’ (e.g. Bernstein 1990) helped the young to forge impor-tant inner and outer disciplines related to Christian values, personal com-portment and civic responsibility; basic literacy skills were important to thedevelopment of these things.

The various English literacy skills were taught using a visible pedagogy, in thatchildren were drilled in such things as the alphabet, spelling and the parts ofspeech, and they were taught to read with graded readers. Composition writingcame much later, after painstaking drill in formation of letters, as well as writingsimple sentences. As many children did not stay at school for long, relatively fewlearned to write more than their names or perhaps simple letters. The visible,albeit conservative, pedagogy involved had some merit in that it explicitly iden-tified such things as spelling rules and taught children to read simple texts,while it also taught parts of speech and some sentence analysis for those wholearned to write. However, the language content taught was quite meagre, andthe linguistic practices it taught were unrelated to the rhetorical concerns ofsome earlier traditions of language study (Christie 1993).

Cultural HeritageIt was Arnold, poet and inspector of schools (1896; Sutherland 1973), who firstproposed teaching poetry to children because of its civilizing influence. Thiswould, he argued, contribute to achieving the order and harmony that amodern society required. He advocated having children commit poems tomemory, and regarded the reading of ‘great literature’ (by which he normallymeant poetry) as an important palliative to the ills that beset a world undergo-ing rapid change and social disorder. His general philosophy was part of abroader commitment to social regulation (Hunter, ibid.) that emerged withnineteenth-century elementary schooling, and Arnold contributed to theemergence of the Cultural Heritage tradition.

F.R. Leavis, influential at Cambridge after the First World War, articulatedmore fully the language of the Cultural Heritage, drawing on the work of Arnold,but also that of colleagues such as T.S. Eliot, I.A. Richards, and DenysThompson. The emergence of the so-called Literary Canon is most associatedwith their names, though Maybin (2000) suggests that its origins were mucholder (also see Peel et al. 2000). For Leavis, like Arnold, the function of literarystudies was to redress the significant social ills that he saw in the contemporaryworld. His works (e.g. with Thompson 1933; 1933, 1934, 1965) express a deepconcern for the poverty of the modern world, destroyed by ‘that great agent ofchange . . . the machine-applied power’ (Leavis and Thompson 1934/1964: 3),which had robbed society of culture and community. Culture had always beenin the keeping of a minority, Leavis believed, and the function of an Englishprogramme was to cultivate ‘taste and sensibility’ in people (Leavis andThompson, ibid.: 1).

Having proposed numbers of great works of literature for study (e.g.Chaucer, Shakespeare, Keats, Austen, George Eliot and T.S. Eliot), Leavisproposed teaching ‘close reading’, seeing literary texts as the expressions of


universal truths which the discerning reader could establish, thereby achieving‘culture’ and ‘sensibility’. However, even though Leavis and his colleaguesespoused close reading of literary texts, they did not engage substantively withthe language involved. Later in the century, students taught in the traditions ofthe New Critics, such as Northrop Frye, also addressed what were deemed greatliterary works, seeking to focus in rather decontextualized ways on such texts,where the intention was to develop taste and discrimination. Again, however,there was no direct engagement with the language used. The relatively visiblepedagogy associated with the learning of the Basic Skills (limited though its goalswere) was buried beneath the weight of the apparent imperative to enter into aculturally valued sensibility (Cultural Heritage) or a refined taste (New Criticism).As a result, the analytical skills that could be used for critical exploration ofvalues in literary texts were not foregrounded.

As the twentieth century proceeded, many literary critics rejected the élitistand ethnocentric exclusions of Leavis and others, while they challenged theidiosyncratic nature of his selection of literary texts, as well as his indifferenceto important post-colonial and feminist writers. Despite the fact that Leavis andhis colleagues fell into some disfavour among academic teachers of English, theinfluence of the Cultural Heritage in school English persisted well into the latetwentieth century, often sitting uneasily alongside what remained of the oldertradition of Basic Skills. This meant that the typical school English programmethat had emerged by the 1950s involved the study of ‘great literature’ and someamalgam of basic skills in school grammar, sentence parsing and analysis andwritten composition. The various elements of the English programme often didnot relate easily to each other, for in practice the ‘gazes’ required for eachdiffered. The skills taught were often unrelated to students’ reading or writingof texts and the study of the great works was rarely used to illuminate work ongrammar, or vice versa.

Personal GrowthThe Personal Growth model of English, like Functional Language Studies (discussedbelow) emerged in the 1960s. Both were born in the aftermath of the SecondWorld War, and the need felt at that time actively to rebuild schooling, thoughin practice the two offered very different models of the pedagogic subject.Personal Growth was a response to the appearance of various progressivist modelsof educational theory (after Dewey) and constructivist models (after confusedreadings of Piaget and Vygotsky) (see Muller 2000 and Christie 2004b), whoseeffect was to propose an ideology of the individual learner as embarked on somepersonal journey of discovery. Constructivism and related theories have hadmany names: ‘pupil-centredness’, ‘learner empowerment’, ‘activity-basedlearning’, ‘inquiry learning’, ‘process approaches to learning’, ‘problem-basedlearning’, to name a number. Stenhouse (1975) provided one influentialaccount of process models of curriculum, while the Plowden Report on Childrenand their Primary Schools (1967) proposed a similar ‘process’ approach to theprimary curriculum, reducing the role of the teacher and stressing that childrenshould learn in their ‘own’ ways.

Several English-teaching specialists met at Dartmouth to discuss the teaching


162

of English in 1966, and Dixon’s Growth through English (1967) was an outcome ofthe meeting. As the title suggests, the classroom was envisaged as a site in whichchildren could be encouraged to grow in independent ways, using their ‘own’language for self-expression and personal discovery. The curriculum theorists inthe constructivist tradition, and the English-teaching theorists after Dartmouth,shared a commitment to teacher facilitation rather than instruction, and theyheld an associated tendency to diminish the notion of any organized form ofknowledge, on the grounds that persons should fashion their ‘own’ knowledge.Correspondingly, principles of scope and sequence in the design of the Englishprogramme, which had been very strong, for example in the Basic Skills tradition,(and which were also a feature of much Cultural Heritage teaching) were activelyrejected, on the grounds that students should forge their own understandings,unfettered by external ordering of knowledge and experience. The skills ofspelling, grammar, sentence construction and reading foregrounded in the BasicSkills tradition were greatly reduced in significance. Grammar was to be taught‘at the point of need’ rather than formally introduced. (Christie 2004b, offers acritique of this view.) The values of ‘close reading’ of literary texts in the mannerof the Cultural Heritage were similarly downplayed, as children were encouragedto read widely and choose texts on the basis of personal, rather than institutional,interests. With the movement to Growth, the language of English became increas-ingly invisible, while the pedagogic position was that of one actively involved in‘self expression’. Subject English now embraced no less than ‘life itself’ (Brittonand Squires, in Dixon 1975: xviii), and children should discover their ‘own’truths, their own values and ‘competences’.

Bernstein (2000: 41–63) offered a very interesting critique of the emergenceof ‘competence’ in pedagogic models, arguing that though various, ofteninimical, scholarly traditions had contributed to it, all had much in common.They imagined ‘an in-built creativity, an in-built virtuous self-regulation’ inpedagogic subjects, such that all could develop in essentially benign ways,pursuing their own interests and acquiring similar competences. They reflectedwhat Bernstein termed ‘an announcement of a universal democracy of acquisi-tion’. Such a model of the pedagogic subject was a very unsatisfactory one,because it deflected attention from ‘the analysis of power and principles ofcontrol which selectively specialise modes of acquisition and realisations’(Bernstein, ibid.: 43). Thus, for the purposes of learning English, the PersonalGrowth model encourages the expression of sensitivities about life but the facilitywith language necessary to achieve this remains at best tacitly understood,because a theory prevails that persons are best left to ‘work things out for them-selves’. Those students well equipped by life experience and opportunity tointuit the desired skills are thus rewarded, while those unable to intuit these aredenied an opportunity to learn. Neither the content to be mastered nor thecriteria that apply for evaluation of students’ efforts are made clear.

In this context Bernstein’s observations about evaluation are significant. Indiscussing the operation of the pedagogic discourse (Bernstein 1990: 165–218)he observed that three sets of ‘rules’ were in play, all of them involved in theshaping of consciousness. Distributive rules concern the identification andspecialization of ‘different orders of meanings’; recontextualizing rules concern


the manner in which the specialist meanings are taken and relocated for peda-gogic purposes; evaluative rules involve evaluation of the pedagogic practices.Hence for school subject English, certain meanings – often to do with selfexpression and/or sensibility – are variously taken from literary and other sitesand recontextualized for pedagogic purposes. Success in control of these is cer-tainly evaluated, as we shall demonstrate by reference to students’ texts below,yet the terms of the task given the student do not on the whole make clear thenature of the pedagogic position that is rewarded, nor the linguistic practicesthat will embody the ‘right response’.

Functional Language StudiesThe tradition of Functional Language Studies owes most to the work of Halliday,who in the 1960s led two major projects in London devoted to the teaching ofEnglish. (See Hasan and Martin 1989 and Halliday and Hasan, in press). Atheory of Functional Language Studies was initially articulated by Halliday,McIntosh and Strevens (1964) to complement the developing languageprojects, though over the years the theory was significantly expanded andextended in the light of research. (See Christie and Unsworth 2005, for a reviewof developments in SFL language education from the 1960s to the present.)Many curriculum materials for schools as well as resources for teachers emergedfrom the work of the 1960s, while many more materials have appeared since, inthe light of developments especially in Australia. After the publication of thefirst edition of Halliday’s Introduction to Functional Grammar (1985), numbers ofbooks for use in schools appeared (e.g. Christie et al. 1990, 1992a, 1992b;Derewianka 1990), while a little later, several pedagogic versions of the func-tional grammar were published (e.g. Derewianka 1998; Christie and Soosai,2000, 2001; Droga and Humphrey 2002, 2003), and volumes to support educa-tional research and teacher education have also appeared (e.g. Unsworth, 2000;Christie 2005).

As the name suggests, Functional Language Studies adopts a functional ori-entation, proposing that language is a social semiotic, involved in the negoti-ation, ordering and structuring of experience, and the functional grammar isthe fundamental tool for analysis of texts. Like Personal Growth, FunctionalLanguage Studies have always espoused an interest in notions of language vari-eties, though the Functional tradition formalized this interest in studies ofboth dialect and register, and as a result it proposed a much more rigorousexamination of language than did the Personal Growth proponents. Inaddition, and unlike Personal Growth, Functional Language Studies recognizedgrammatical differences between speech and writing, and argued for thevalue of teaching explicit knowledge about language (‘KAL’), the better toequip students for informed use of the language system. In other words, inthis tradition learning about language does not just have an abstract analyti-cal importance, as it often did in the Basic Skills tradition for example. On thecontrary, in this tradition, language is itself the object of study for teachingand learning, and an intimate association of text and context is established,while the relationship of text and the grammatical resources in which it isrealized is a major focus of attention.


164

Functional Language Studies has been far more interventionist in its pedagogi-cal principles than Personal Growth, eschewing what Halliday once referred to asits policy of ‘benevolent inertia’, and adopting a model of the teacher as activelyguiding and directing students’ learning. Furthermore, Functional LanguageStudies has greatly increased the visibility of language in the curriculum, while ithas also pointed directions for pedagogic practice. After the late 1970s and early1980s, probably the best known of its pedagogic interventions has been genre-based pedagogy with which the name of Martin (1999) is associated, thoughmany of his colleagues have by now contributed to the ongoing development ofthis pedagogy (see Macken-Horarik 2000). The ideal pedagogic subject positionin this model is a linguistically informed one, capable of the analysis, produc-tion and critique of texts and contexts. A knowledge about language, it is held,enables students to interpret the linguistic bases of value positions and ideolo-gies embodied in texts. An understanding of the power of language supple-ments an understanding of the language of power (Macken-Horarik 1998;Martin and Rose 2003, 2006).

For all these reasons, we argue that use of the systemic functional grammarhas made important gains in developing a model of language studies which hasthe potential to inform the development of an enhanced discourse for subjectEnglish, one which can point directions for the design, scope and sequence ofthe language curriculum across the years of schooling. Among other things, thefunctional theory provides an understanding of the nature of the languagesystem, an awareness of the grammatical differences of speech and writing, anda sense of the developmental changes that need to occur as students mature,achieving greater mastery of language and the meanings it realizes. Below, byreference to two texts written by school students, we illustrate something of thepower of the functional grammar in unpacking what it is to be successful for thepurposes of school learning in subject English. To do this, we demonstrate someof the linguistic resources students deploy when they are evaluated as success-ful, and we suggest that an understanding of these resources should inform theteaching of English, providing knowledge about language that should consti-tute part of the overt objects of study for the future. First, however, we reviewthree remaining models of subject English.

Cultural StudiesLater in the twentieth century, the specialized languages of English expandedfurther with post-modernism and a powerful relativizing of the dominant nar-ratives of mainstream western literary traditions. Cultural Studies incorporatesand valorizes the texts of popular culture, elevating them to the status of literarytexts, and this model of subject English offers students a range of reading andsubject positions. In her work on Critical Practice, for example, Belsey (2002)explores the notion of subjectivity in textual study and foregrounds the impor-tance of interrogating texts from a variety of discursive reading positions. Here,the gaze implicitly constructed for learners is one of responding to and inter-rogating the subjective positions (values and ideologies) embodied in texts.Multiple readings of texts are invited, with an emphasis on critical analysis oftexts for their dominant discourses and sociocultural effects. Like Personal


Growth, Cultural Studies takes the tools for interrogating texts largely for granted.There is no detailed attention to the linguistic tools necessary to produce andsubstantiate particular readings, and as we noted above for Cultural Heritage, aconcern for the language in which meanings and values are realized is buried –in this case in the drive to tease out ideological positionings.

Despite the continued significance attaching to the Cultural Heritage modelalready referred to, it is apparent that, at least at senior secondary levels ofschooling, the Cultural Studies model is proving increasingly influential. In NewSouth Wales, for example, there is a curious disjuncture between the juniorsecondary English curriculum whose influences appear to be mainly CulturalHeritage, and the senior secondary English curriculum, whose expectations forthe ideal pedagogic subject position are quite other than those for the juniorschool, revealing the influence of Cultural Studies. Thus, the Higher School Cer-tificate (HSC) in English for 2005–7 provides for a major Area of Study (NSWBoard of Studies 2005), in which students are asked to study a disparate rangeof texts yoked together under the diffuse, abstract heading of Journeys. They can,for example, focus on Physical Journeys, Imaginative Journeys or Inner Journeys, andthey receive a Stimulus Booklet (involving some loosely connected short texts andimages about journeys) as well as a set of prescribed texts, both literary and non-literary. These include Shakespeare’s The Tempest and Coleridge’s poems such asThe Rime of the Ancient Mariner. In the case of Imaginative Journeys, for example,students must complete a written examination task in which they ‘compose anarrative which reflects the important ideas (they) have learnt from studyingImaginative Journeys’. Moreover, students must ‘demonstrate an understandingof the concept of journeys in the context of (their) studies’ and go on to writea response that ‘reflects the important ideas’ learnt; they must also use language‘appropriate to audience, purpose and context’ and show ‘use of appropriatenarrative techniques’. A study of texts produced in one privileged selective highschool in New South Wales (Christie and Derewianka, in progress) reveals con-siderable confusion in the Year 12 students (who are regarded as talented) inattempting to deal with the task; what they write bears no discernible relation-ship to the texts identified, and it is difficult to establish how notions of ‘appro-priate narrative techniques’ are understood, or indeed what are the ‘importantideas’ regarding journeys students might have learned. In addition, and com-pounding these problems, there is no evidence of any close attention to thelanguage of the texts studied, for the curriculum guidelines given to teachersand students offer no advice about these matters. Overall, the tools needed toanalyse and interpret the visual and verbal texts involved are not addressed, andCultural Studies thus carries forward the by now well-established tradition ofrendering invisible the linguistic skills students need to succeed in English. Thepedagogic subject position is in one sense not unlike that associated withPersonal Growth, since students are intended to respond in a ‘personally felt’manner to the ‘themes’ selected for study, while the selection of texts used,though including some valued in the Cultural Heritage tradition, also reflects thegeneral commitment of Cultural Studies to the view that a very catholic range oftexts should be selected.


166

MultiliteraciesMultiliteracies is a development of the late twentieth century, and in some form-ulations it bears a close relationship to Cultural Studies, in that it proposes aninterest in a wide variety of texts, verbal and visual, popular and otherwise,spoken and written, and its tendency is often to downplay the significance of aclear model of language. Its emergence is often associated with the manifesto ofthe New London Group, which appeared in the Harvard Educational Review in1996, and the ideas involved later led to a volume on Multiliteracies (Cope andKalantzis 2000). Some members of the New London Group have espoused a com-mitment to use of the functional grammar, while also arguing that the world ofcommunication requires more than a theory of language (e.g. Kress 2002, 2003)to explain and teach about it. In practice, among many proponents of Multilit-eracies the use of the functional grammar receives a very limited acknow-ledgement (e.g. Hawisher and Selfe 2000). Unsworth (2001) has offered themost comprehensive discussion of Multilieracies, which uses the functionalgrammar. Like others of his colleagues in the SF tradition (see Unsworth 2002)he refers to a range of ‘subject specific literacies’, a term intended to capturethe particular claims on literacy represented by the different subjects found inthe school curriculum. Together with colleagues (Unsworth et al. 2005) he hasrecently discussed the teaching of children’s literature and computer-basedlearning. Overall we suggest that for some proponents of Multiliteracies, this isbest understood as a ‘dialect’ of Cultural Studies, while for others, like Unsworth,it is part of the Functional Language Studies model. Hence, two somewhat differ-ent pedagogic subject positions are possible, the one closer to that of CulturalStudies, the other to that of Functional Language Studies.

New Literacy StudiesThe term New Literacy Studies (NLS) is one most associated with the work of Street(1984, 1993, 1997, 2001) who, working as an anthropologist, observed the uses ofliteracy, especially in non-western societies, and who argued that unless oneunderstood the range of social practices that surrounded a use of literacy, it wasimpossible fully to understand it or to teach it effectively. Taking the term ‘literacyevent’ from Heath (1983), he went on to argue that people bring to a literacyevent ‘concepts, social models regarding what the nature of the event is, thatmake it work and give it meaning’ (Street 2001:11). He drew a distinctionbetween what he termed conventional models of literacy that are ‘autonomous’(involving teaching basic literacy skills in an unproblematic manner, to be simplyrecycled throughout life), and ‘ideological’ models (which see literacy as part ofsocial practice and, as such, intimately bound up with values and meaning).According to Street and his colleagues, ‘autonomous’ models of literacy are unsat-isfactory because they imply a singular or unitary model of literacy, whereas‘ideological’ models recognize a diversity of literacy practices, which are‘situated’. NLS attracted others including Gee (1996) and Barton (1994), and ledto proposals that since ‘literacy practices are situated in broader social relations’it was necessary ‘to describe the social setting of literacy events, including the waysin which social institutions support particular literacies’ (Barton, ibid.: 35).Anyone working in the SFL tradition, at least, has no difficulty acknowledging the


significance attaching to social processes, though the importance attaching to‘literacy practices’ in the work of NLS theorists is nonetheless unfortunate, sinceit serves to deflect attention away from the language system, and its effect for ped-agogic purposes is to marginalize language study. Hence, like both Personal Growthand Cultural Studies, the NLS tradition has served to bury the study of languageitself, in this case because of the apparent imperative to focus on the social prac-tices that surround the uses of literacy. The pedagogic subject in this tradition isone preoccupied with the immediate literacy practice, such that relatively littleinterest attaches to any sense of the linguistic system with which one engages inorder to construct that practice. Muller (2001) offers a useful critique of NLS.

Table 8.1 presents the models of English synoptically, though as noted, it doesnot present Multiliteracies separately. It distinguishes among the knowledgedimensions of English (the objects of study in the different models), knower orsubject position dimensions (related to the ‘gaze’) and the linguistic practicesby which these are created. All these models are present in contemporaryschool English, though some in rather shadowy ways. The difficulty of readingthe requirements of the discipline is multiplied in this context, for teachers andstudents alike are often left uncertain of what is valued. Recognizing and real-izing ‘what counts’ in examination English becomes an increasingly dauntingtask (Macken-Horarik 2006). Bernstein pointed to the difficulties of acquisitionin horizontal knowledge structures such as English:

Because a horizontal knowledge structure consists of an array of languages, anyone transmission necessarily entails some selection, and some privileging within the setrecontextualized for the transmission of the horizontal knowledge structure. Thesocial basis of the principle of the recontextualizing constructs the perspective ofthe horizontal knowledge structure. Whose perspective is it? How is it generatedand legitimated? (Bernstein 2000: 164, emphasis added)

We now turn our attention to the semantic features of students’ written textsand what a functional linguistic analysis of these can teach us about the kinds ofknowledge, gaze and linguistic practices which are actually valued in examin-ation English. Using the evidence of our analyses, we argue both that it ispossible to make explicit a great deal of the linguistic resources needed tosucceed, and that in making it explicit we take a critical step towards enhancingthe verticality of subject English.

Unpacking the linguistic underpinnings of ‘finer feelings’

Despite the diversity and apparent openness of knowledge structures in subjectEnglish, careful attention both to the examination texts produced by studentsand their evaluation reveals which kinds of gaze, which forms of truth and whichlinguistic practices are actually privileged. In practice, we argue that value posi-tions associated with the expression of sensibilities – empathy, discernment,‘finer feelings’ – are particularly rewarded by teacher-examiners. However openthe invited response may be, the preferred response is that associated with


Analytical Categories Basic Skills Cultural Heritage Personal Growth FunctionalLanguage Studies

Cultural Studies New LiteracyStudies

Objects of study Discrete skills forbasic readingand a littlewriting(alphabet,spelling, parts ofspeech, gram-matical analysis,paraphrasing).

Accessing sensi-bility and culturethrough engage-ment with ‘thecanon’. Learningto read literarytexts as embodi-ments of greatand universaltruths.

Knowledgeabout spokenand writtenlanguage;analysis and pro-duction of textsin a range ofsocial contexts(through appli-cations of functionalgrammar).

Multiplereadings of different texts(spoken,written, literary,popular, multi-modal). Criticalanalysis of textsfor dominantdiscourses andsocio-culturaleffects.

Literacy eventsand their associ-ated literacypractices (bothspoken andwritten).Learning toread and writeas ‘situated’activities.

Processes oftalking, listening,reading andwriting, stimulatedby texts and expe-riences. Learningto read and writeas journeys ofpersonaldiscovery.

Subject positions or‘gazes’

A well-regulatedcitizenry, inculcated withthe basics; developing inner discipline,Christian valuesand personalresponsibility.

A culturallyvalued sensibility– the preserve ofélites who canunderstand (andinterpret) greatworks in thereceivedtradition.

A linguisticallyinformedanalysis, produc-tion andcritique of textsand contextspowerful inmainstreamculture.

Interrogation ofthe subjectivepositions(ideologies)embodied intexts; question-ing, critiquing,even subvertingtexts and their social conditions.

Investigation ofsocial contextsin which literacyis used and ofthe ideologicalpositionings andmeaningsinvolved.

A sensitive andself-consciousindividuality – indialogue with textsand with otherswho share theirpersonalresponses to textsand experience.

Table 8.1 Models of school English in contemporary schools

Analytical Categories Basic Skills Cultural Heritage Personal Growth FunctionalLanguage Studies

Cultural Studies New LiteracyStudies

Linguistic practices Use of languagethat buildsdiscretecompetences.* Texts such asspelling andgrammar books,dictionaries,poems, composi-tion books andbiblical passages.(Basic literacycompetences)

Use of languagethat interpretsand createsfictive worlds andliterary sensibilities.* Texts such asnovel, poetry anddrama; analyticaltexts such asinterpretiveessays.(Culturalliteracy)

Use of languagethat is analytical(metalinguistic)and well struc-tured.* Texts such asnarratives, expo-sitions, multi-literate text,analysis and textinterpretations.(Powerful metaliteracies)

Use of languagethat identifiesand challengesideologiesenshrined intexts.* Texts such asdeconstructiveessays, evalua-tions, satires,spoofs.(Multimodaltexts and multiliteratepractices)

Use of languagethat is localizedand particularto social events. * A variety oftexts, valued fortheir relevanceto immediatesetting.(Situatedliteracypractices)

Use of languagethat is close to‘self’.* Texts such asjournals, dia-logues, personalletters, imagina-tive reconstructionof literary texts,emails to friends.(Personal literacy)

Table 8.1 (continued)

170

Cultural Heritage English. The tendency to value texts which demonstratecontrol of the rhetoric of ‘finer feelings’ is clear, despite the apparent privileg-ing of ‘Cultural Studies’ in contemporary curricula, such as the New SouthWales senior syllabus for English, already referred to.

In order to illustrate some of this, we turn to a text written by a school studentin England as part of the national testing regime in that country. Text 8.1 wasused by the Qualifications and Curriculum Authority in preparing advice tomarkers for the English Test for Key Stage 3, which is the last of the Key Stagesidentified under the terms of the National Curriculum in England. The MarkScheme for Paper 1 for Key Stage 3 (Qualifications and Curriculum Authority 2002)gave a set of performance criteria which were illustrated using exemplars drawnfrom students’ written work produced in a pilot phase.

Text 8.1 was recognized as ‘in the range above level 7’ (the highest levelused), while it was given the highest mark available, namely a score of 33. (Thestudent’s spelling and punctuation are preserved.)

Text 8.1

As she eased her way past the leafy tendrils of the branches that beckoned to her; a senseof excited anticipation mixed with sad reflection filled her mind, accentuated by thetingling drops of rain that fell on to her head. Crystal drops. Each one glistened, as ifcrafted by a happy memory.

The rocks, weather worn and covered in ancient moss, were silouetted (sic) by thesmiling sun and they towered above Marie as she timidly stepped onto the path that ledits winding way to the heart of the gorge. Closing her eyes now, she felt giddy with hap-piness; this was the place – her place; untouched the rocks and moss knew her story; thegorge had hidden depths; stories of times long forgotten.

Marie didn’t know exactly where she was going, although she was certain of where sheneeded to be. Her thoughts became voices which echoed from the rocks, recalling the con-versations that had filled her childhood. She glanced down at her hands and, feelingthe wrinkles there, looked up at the rocky ways of the gorge in faint disgust. Grippingthe box tightly, she continued through the living arch of trees.

A tear rolled down her cheek; surprised to find it there, she brushed it aside.

As the birds of the air agreed with her, their song resonating with joy, she found it. ‘It’was a pool, quicksilver amongst the green; a treasure Marie’s mother had shown her,when she was a child. And there, beside the pool, was the rock her mother had alwayssat on. The air around it whispering and laughter ringing off the rocks. A glimmer ofa smile caressed her cheeks. She laid the box down. A fragile flower graced the bank ofthe pool. Marie picked it and laid it by the box.

‘Goodbye’, she murmured.

She left the box there, a treasure amongst treasures, and walked back into the gatheringgloom.


The task was relatively open-ended. Students were to write about ‘someonegoing for the first time into a secret place – for example, a cave or an aban-doned building’. The experience could be ‘real or imagined’ and students‘could describe the reactions of the person as he or she goes into and exploresthis place’; they should ‘build up an atmosphere of tension and suspense . . .(and) include an effective ending’ (English Test 1 Paper 1. Key Stage 3 (Levels 4–7),2002, QCA, page 9.)

Text 8.1 was judged ‘a confidently written response which engages theinterest of the reader through its selection of effective detail and control oflanguage and structure to create a sense of place and mood’. The writing, it wassaid, ‘moves smoothly through the description of the scene and events of child-hood, reflecting on the effects of the passage of time, both in terms of thechanges in scenery and in the writer’s own responses’. The student is a skilledwriter and uses language resources to build a sense of time past and of a nos-talgic regret for a happy childhood. Let us look closely at the semantic choicesmade in this text, and the linguistic realizations of these.

Semantically, the text creates a world of atmosphere and regret throughdense nominal groups which provide physical descriptions of leaves and rain androcks and elaborate on these in greater detail in embedded (or rank-shifted)clauses. These embedded clauses ‘plump out’ the nominal group throughdetailed elaborations of the information contained in its preceding head noun.Nominal groups are underlined below, head nouns in bold, and embeddedclauses indicated through double square brackets, in the following examples:

Example 1: as she eased past the leafy tendrils of the branches [[ that beckoned her]]

Example 2: accentuated by the tingling drops of rain [[ that fell on her head]]

Example 3: and there, beside the pool, was the rock [[ her mother had always sat on]].

A related feature of this student’s sentence structure is the use of enclosedclauses. Here the additional information is not part of what it is modifying (i.e.not embedded) but inserted into the middle of it. Halliday (1994: 227) callsthese ‘enclosed clauses’ and represents them by means of angled brackets, as in:

Example 4: the rocks << weather worn and covered in moss >> were silhouetted by thesmiling sun

Two other resources related to this student’s sentence structure are in evidencehere. They include the use of non-finite clauses, often in thematic positions,and the occasional use of elliptical clauses. Perera (1984: 235–6) observed thatnon-finite clauses emerged in children’s writing at about age 12, though not allchildren will have mastered this by that age. When placed first (i.e. in themeposition), non-finite clauses foreground the particulars of psychological andsensory experience, framing and deferring the message encoded in the mainclause. Examples of non-finite clauses which are either thematic or enclosedinclude:


172

Example 5: Closing her eyes now, she felt giddy with happiness (thematic)

Example 6: Gripping the box tightly she continued through the living arch of the trees(thematic)

Example 7: She glanced down at her hands and, << feeling the wrinkles there >>looked up at the rocky walls of the gorge in faint disgust (enclosed)

As for the elliptical expressions, these provide particular emphasis in partbecause they are not full sentences. They serve to distil significance by leavingout elements of the clause that can be taken as given (and supplied by the activereader).

Example 8: Crystal drops

Example 9: Stories of time long forgotten

An additional feature of the literate language of Text 8.1 is what Halliday callselaboration.

In elaboration one clause elaborates on the meaning of another by further spec-ifying or describing it. The secondary clause does not introduce a new elementinto the picture but rather provides a further characterization of one that isalready there, restating it, clarifying it, refining it, or adding a descriptive attrib-ute or comment. The thing that is elaborated may be the primary clause as awhole, or it may be just some part of it – one or more of its constituents. (Halliday1994: 225)

Using elaboration, the student is able to provide specific details about thenatural phenomena which her character, Marie, observes in her journey back toher childhood ‘mecca’.

Example 10: The rocks, weather worn and covered in ancient moss, were . . .

Example 11: This was the place, her place, untouched . . .

Elaboration and its related feature, apposition, is a defining feature of success-ful narratives and of highly valued interpretations of narrative (see Macken-Horarik, in press for 2006, for extended discussion of this resource in a widerrange of texts).

We have briefly touched on three linguistic features – atmospheric nominalgroups, thematic non-finite clauses, and elaboration – which are crucial to thecontrol of literate language. These features are developmentally a feature of latechildhood to adolescence. (Christie 2002; Derewianka 2003). They enablewriters to distil the meaning of experience and to expand on this in the waystypical of ‘highly literate’ language. These are related to the genesis of anotherlinguistic feature – that of metaphor (lexical and grammatical). Metaphor is the


gateway to alternative realms of experience and is crucial to success in English– in both reading and writing.

In lexical metaphor, rhetorical transference is created through the creativedoubling of one word or image with another. The reader is invited to construeexperience both literally and figuratively. This transcoding of experience, con-struing the sun as ‘smiling’, the drops of rain as ‘crafted by a happy memory’,and bird song as ‘agreement’ with the character’s happiness, enables a writer toconstrue experience symbolically, even if the following examples are jarring intheir self-conscious artifice:

Example 12: The rocks were silouetted (sic) by the smiling sun

Example 13: Each one (i.e. crystal drops) glistened, as if crafted by a happy memory

Example 14: As the birds of the air agreed with her, their song resonating with joy . . .

In the literary idiom adopted by this student, lexical metaphors serve not just anideational function (through their transcoding of experience); they also servean interpersonal function – provoking attitude in the reader through powerful,connotative realizations of attitude (see Martin and White 2005: 61–8, fordiscussion of the role of lexical metaphor within evaluation).

The other kind of metaphor important in English (and other disciplines) iswhat Halliday has called ‘grammatical metaphor’ (Halliday 1994). This involvesnot a doubling of meanings (literal and figurative) but of grammatical forms(congruent and non-congruent). Here processes such as ‘to anticipate’ or ‘toreflect’ (congruently realized in verb forms) are encoded in nominal form, asin:

Example 15: A sense of excited anticipation [[mixed with sad reflection]] filled hermind.

The effect of the nominalization in this latter example is to create an abstrac-tion – a sense of excited anticipation mixed with sad reflection. These non-congruentforms are good examples of grammatical metaphor – a resource which ‘untiestexts from situations and allows writers to re-construe activities as things andthus break the iconic connections between linguistic and material activity’(Martin and Rose 2003: 245). Using this resource enables writers to interpretexperience from a ‘meta’ point of view, to abstract away from material activitythrough linguistic activity.

What can we say about the gaze that is adopted in this text? Semantically, thestudent demonstrates an adult sensibility in his or her nostalgic regret for timespast. The narrator adopts a subject position something like this: happy timesoccur in childhood in company with one’s mother, and old age is about lookingback. The narrator’s voice is one of empathic engagement with the past, andthis is projected through thoughts and feelings.


174

Example 16: Marie didn’t know exactly where she was going, although she wascertain of where she needed to be. Her thoughts became voices whichechoed from the rocks, recalling conversations that had filled her child-hood.

Subject positions involving empathy and discernment are important ifstudents are to adopt and demonstrate the appropriate gaze in English(Macken-Horarik 2003). A key linguistic resource for producing this kind ofsensibility is appraisal, which is concerned with ‘the kinds of attitudes that arenegotiated in a text, the strength of feelings involved and the ways in whichvalues are sourced and readers aligned’ (Martin and Rose 2003: 22). In this text,as in others like it, the student encodes a particular reading of experiencethrough the combination of inscribed affect (via attitudinal lexis) and invokedappraisal (via connotative renderings of experience). In this text, invoked affectis either flagged via non-core vocabulary to describe the character’s physicalbehaviour, or provoked by means of lexical metaphor (see Martin and White2005, for a fuller treatment of this important domain of interpersonalmeaning). The following examples display the combination of invoked affectthrough underlining and inscribed affect through bold face.

Example 17: Each one glistened, as if crafted by a happy memory.

Example 18: Closing her eyes now, she felt giddy with happiness.

Example 19: Gripping the box tightly, she continued . . .

The sensibility embodied in this narrative and valued by the examiners was notexplicitly called for in the wording of the question. Furthermore, no fourteen-year-old might be expected to have had the necessary experience to produceText 8.1. The sensibility rewarded here is far more likely to come out ofexposure to fiction than out of lived experience of old age, loss and memory.The difficulty for many students (and their teachers), however, is that the QCAcomments, while laudatory, do not identify the language resources used toproduce the sensibility which is so highly valued.

However, if indeed such texts and their values are to be rewarded in subjectEnglish, we argue that we should teach children the language skills necessary toaccess and produce the vertical discourse involved. From the point of view ofFunctional Language Studies, this means that we should make the forms of knowl-edge – the gaze and the linguistic resources that are valued – explicitly availableto students, both by modelling and by analysis of examples (see Gray 1998,about a programme to do this for remote Indigenous children in Australia). Interms of the gaze rewarded in this situation, careful scrutiny reveals that what isrequired is an expression of the linguistic demands of Cultural Heritage English.

Having examined an instance of a ‘literary’ narrative produced in examina-tion conditions, we now turn our attention to the ubiquitous challenge, in bothAustralia and England, of the unseen text and the open question in response toit, to be answered in examination conditions.


Interpretation and the linguistic underpinnings of the symbolic readingof a literary text

Like their English counterparts, Australian educators have adopted teachingand testing policies which reward the creation of alternative worlds and of ‘finerfeelings’, though again within terms which obscure the language proficiencyactually needed. As we shall see, the linguistic features of vertical discourse shiftonly slightly when we move from narrative to interpretation of narrative.

Text 8.2 is taken from the NSW School Certificate Reference Test in English, 19955

– an examination which all sixteen-year-old students face at the end of Year 10.In this test, students were presented with a text adapted from a story called TheRed-Back Spider by Peter Skryznecki. The story concerns a young migrant boy andhis mother going to work clearing weeds in the garden of an Australian woman,Mrs Burnett, who appears to be a racist. Mrs Burnett shows no consideration forthe pair and refuses to let the young boy (through whose eyes events are seen)play with toy soldiers which he finds while sheltering from the hot sun underher house. She also ignores the plight of the boy’s mother when she discovers anest of dangerous red-back spiders in a tin while weeding. The story focuses onthe young boy’s growing awareness of Mrs Burnett’s persecutory attitude tothem and of his own mother’s emotional response to her rejection. It is a chal-lenging narrative, made more difficult by the response task students were given.

Although the story is called The Red-Back Spider, it ends with the words, ‘. . . I knewit was nothing to do with the spider’. What do you think the story is about? Howdoes the writer make it an effective story?

It is not hard to imagine the difficulty of this question for any students whoassumed (understandably) that the title of this story would be related to what itis about (i.e. about a spider). However, the wording of this question directsstudents to reconstrue the question as ‘What do you think the story is (really)about?’ This makes the whole process even more mysterious than it appears, orneeds to be, for students (see Macken-Horarik 2006, for more extensive discus-sion). Text 8.2 – like others in the higher range of achievement – ignores thered herring of the question and produces a response in line with a CulturalHeritage agenda and the privileging of the symbolic reading (Macken-Horarik2006). The text is presented below.

Text 8. 2

‘The Red Back Spider’ is a story about an old woman who finds it necessary to build awall between herself and her worker, about a woman who is so selfish that she will notallow a young child to play with toys which have remained unused and unwanted forsome time. The old woman lives alone, and feels like she needs to separate herself fromthe gardener and her child, ‘she closed the door slowly, deliberately, with a metallic click’.

The story’s ending revolves around the burning of the spider and the sharp words of theold woman, ‘Please return all the toys and do not play with them’. However the sadness


176

of the child’s mother runs deeper than her fear of the spider. She is upset that this oldwoman, like the spider, hides her precious things away, and runs to protect them at anysign of danger. The fact that her child finds enjoyment in playing under the house andwith blocks of wood in the dirt and weeds seems to show that they don’t have too manypossessions. The mother is distressed at the scenario that the only enjoyment her son isgetting in his school holidays has been cut off by the old woman.

The separation between the two lifestyles, that of the lonely, selfish old woman, and thehardworking mother is the essence of the story. An old woman who is clinging to a pastof child’s toys and a house, with no feelings for the people living in the present, asopposed to a tough, self-sufficient wife who struggles to provide for her family and missesher husband.

The writer uses descriptive metaphorical language, ‘whose brown-as-leather skin hungover her frame like a synthetic material’, to enable to reader to picture the exact scene intheir head. This brings the reader closer to the characters and the backyard of the house.The questions the child asks to themselves make the reader consider ‘was there somethingevil in it’s nature that it had to hide?’ This question seems as if it is being asked alsoabout the old woman.

The author shifts the focus of the story from the description of the house, to the historyof the family, to the physical characteristics, and onto the child’s games and the discov-ery of the spider. This movement of the story at a reasonably quick speed keeps the readerinterested. The story is effective because it brings the reader into the story, allows themto think about it, and moves at a reasonable pace.

This interpretation of The Red-Back Spider was awarded an A+ (15 marks out of20). It is typical of those in the higher range of achievement. Semantically, itembodies several choices which typify ‘verticality’ in the discourse of subjectEnglish. Foremost is control of symbolic abstraction. In this linguistic task,sensory experience (e.g. what happens in a story) is construed in non-sensibleterms (the narrative significance of characters and events). In the highly valuedliterary interpretation, the student draws on a key lexical metaphor in thestimulus narrative to distil the abstract meaning of events. This symbolic readingis typically represented in dense nominal groups containing generalizationsabout the overarching significance of the text. Note the combination of densenominal groups (underlined) and the sequence of three embedded clauses (indouble square brackets) in the following interpretation of the character of theold woman in the story:

Example 20: ‘The Red Back Spider’ is a story about an old woman [[ who finds it nec-essary to build a wall between herself and her worker, about a womanwho is so selfish [[ that she will not allow a young child to play with toys[[which have remained unused and unwanted for some time]]]]].

Nominalizations which distil the organizing oppositions of the narrative arepowerful resources in the service of abstraction, particularly when combined


with elaboration (indicated with an equals = sign). In the following example,the opposition between two lifestyles is identified and elaborated:

Example 21: The separation between the two lifestyles (=) that of the lonely, selfish oldwoman, and the hardworking mother is the essence of the story.

In this task, the student needs to discern key metaphors (such as the spider) anduse these to organize the response (contra the confusing wording of the exam-ination question). The semantic strategy underpinning this symbolic readinginvolves making an extended analogy between the spider and some other aspectof the narrative. In this interpretation, the student relates the spider to the oldwoman and interprets her actions through the symbolic mesh of this analogy:

Example 22: She (the mother) is upset that this old woman, like the spider, hides herprecious things away, and runs to protect them at any sign of danger.

All A-range responses in this corpus (approximately 10) focused sharply on thespider motif and created connections between this and other parts of the nar-rative. In short, they ignored the confusing lead given in the question (seeMacken-Horarik 2006, for discussion of some of these other responses).

What kind of gaze is created here? As in Text 8.1, the sensibility projected isone of empathy (feeling with the main character), realized primarily by expres-sions of affect, as in the following two examples:

Example 23: The mother is distressed that the only enjoyment her son is getting hasbeen cut off by the old woman.

Example 24: However the sadness of the child’s mother runs deeper than her fear of thespider.

However, in this text there is also a strong dose of judgement, especially negativejudgement of the old woman, Mrs Burnett. As in the earlier narrative text, aliterary sensibility is created through a combination of direct and indirectexpressions of attitude. Tokens of judgement (indirectly invoked through non-corelexis) are underlined while explicit (directly inscribed through attitudinal lexis)inscriptions of judgment are in bold.

Example 25: . . . about a woman who is so selfish . . .

Example 26: . . . an old woman who is clinging to a past of child’s toys and a house,with no feelings for the people living in the present, as opposed to atough, self-sufficient wife who struggles to provide for her family.

In other words, empathy is combined with discernment in this response, tuninginto the way the story narrativizes prejudice. Being able to read and re-enactboth emotional sensitivity to characters’ feelings and ethical judgement of theirbehaviour is crucial to the gaze in English, and often rewarded within Cultural


178

Heritage models of the subject (both in the disciplinary and the personal senseof the word ‘subject’).

Table 8.2 presents an overview of the most salient linguistic features of thetwo texts we have considered, along with relevant examples.


Table 8.2 Semantic features of Texts 1 and 2 with examples

Text 1 Narrative

Abstraction (via dense nominal groups and embedding)

Text 2 Literary interpretation

• The leafy tendrils of the branches [[thatbeckoned her]]

• the tingling drops of rain [[that fell onher head]]

• the rock [[her mother had always saton]]

• a woman [[who finds it necessary tobuild a wall between herself and herworker]]

• the separation between the two lifestyles,that of the lonely old woman and thehardworking mother is . . .

Metaphor (lexical and grammatical)

• The rocks were silhouetted by the smilingsun

• Each one glistened, as if crafted by ahappy memory

• The birds of the air agreed with her.• Her thoughts became voices which

echoed from the rocks

• However the sadness of the child’smother runs deeper than her fear of thespider

• She is upset that his old woman, like thespider, hides her precious things awayand runs to protect them at any sign ofdanger.

Elaboration and apposition

• The rocks, weather worn and covered inancient moss

• This was the place, her place, untouched. . .

• the gorge had hidden depths, stories oftimes long forgotten . . .

• This is a story about a woman [[whofinds it necessary to build a wall . . .]]about a woman who is . . .

• the separation between the two lifestyles,that of the lonely old woman and thehardworking mother is . . .

Appraisal (via various resources that build affect or judgement)

• Each one glistened as if crafted by ahappy memory

• Closing her eyes now, she felt giddy withhappiness

• Gripping the book tightly she continued. . .

• She is upset that this old woman hidesher precious things away

• The mother is distressed• About a woman who is selfish• An old woman with no feelings for the

people living . . .

Conclusion: building a metalanguage for enhancing verticality in subject English

Bernstein (2000: 155–74) was clear that horizontal knowledge structures,marked by ‘weak grammars’ as he suggested, tend to produce discourses thatare segmentally organized, and that ‘there is no necessary relation betweenwhat is learned in the different segments’ (ibid.: 159). Our discussion hasrevealed a range of different knowledge structures (models) and their assoc-iated gazes which have emerged over the history of subject English, such thattheir various languages are often incommensurate, even contradictory. Para-doxically, and partly as a result of the dilemmas caused by a history that has beenso segmented, as school subject English has evolved, its tendency has been todiminish the status of overt teaching of knowledge about language, while pro-moting various valued subject positions, though the tools for their expressionhave become increasingly invisible. This is especially clear in examinations ofstudents’ written responses to open questions and unseen texts in the later yearsof junior secondary English (age 16 and up).

We have argued that the metalanguage of SFL has proved useful in exploringthe linguistic character of texts valued in many contexts of instruction and eval-uation in English, and we have sought to demonstrate this by reference to twovalued texts from England and Australia. In making the features of these textsvisible through the SF metalanguage we are able to make subject English morevisible. English teachers have access to a metalanguage through which to artic-ulate requirements, and hence to make it more accessible to a greater numberof students. In its work within Functional Language Studies, functional linguisticshas created links across ‘segments’ of communication, and this has helpedcreate mechanisms for identifying and integrating the diverse specialized lan-guages of various subjects with horizontal knowledge structures (see Martin, thisvolume). As we have noted, SFL gives us tools for connecting meanings towordings. The wording choices of successful students in English give evidenceof a particular orientation to meaning (gaze) which is summarized in Table 8.2.The orientations which are highly valued include preferences for abstraction(realized in use of dense nominal groups, nominalization and embeddedclauses), for restatement or reformulation (realized in elaboration and apposi-tion), for lexical and grammatical metaphor, and for empathy (realized in directand indirect expressions of affect) and ethical discernment (realized in direct andindirect expressions of judgement).

The metalanguage of SFL can be used to highlight commonalities and dif-ferences across the specialized languages and methodologies of English, and wesuggest it can be used to outline a scope and sequence for progressing learningin English at both primary and secondary levels. The work of doing this hasalready begun (e.g. Christie and Derewianka, in progress; Christie, Macken-Horarik and Maton, in preparation), though more needs to be done to developa coherent, theoretically rigorous account of the English curriculum fromKindergarten to Year 12. At this stage, students are often forced to intuitrequirements as they progress up the years, without access to mechanisms fordeciding which responses to which texts are going to be valued. SFG provides us


180

with tools for translating across, and comparing, texts and contexts, showingwhat they share linguistically and how their covert fashions of meaning arerelated to one another. In this sense, functional linguistics offers a new set oflenses for creating a new gaze, for making the subject at least partly visible, andhence for strengthening the discipline.

References

Arnold, M. (1869), Culture and Anarchy. An Essay in Political and Social Criticism. London,Edinburgh and NY: Thomas Nelson and Sons.

‘Australia’s Language. The Australian Language and Literacy Policy’ (1991), Canberra: Aus-tralian Government Publishing Service.

Barton, D. (1994), Literacy. An Introduction to the Ecology of Written Language. Oxford UKand Cambridge USA: Blackwell.

Belsey, Catherine (2002), Critical Practice (second edn). London and NY: Routledge-Falmer.

Bernstein, B. (1990), The Structuring of Pedagogic Discourse. Class, Codes and Control, Vol. IV.London and NY: Routledge.

Bernstein, B. (2000), Pedagogy, Symbolic Control and Identity. Theory, Research, Critique (rev.edn). Lanham, Boulder, NY and Oxford: Rowman and Littlefield.

Board of Studies (1999), English Stage 6 Syllabus. Sydney: New South Wales. http://www.boardofstudies.nsw.edu.au (accessed on 30 December 2005)

Britton, J. and Squires, J. (1975), ‘Foreword’, in Dixon, J., 1975, pp. vii–xviii.‘Bullock Report, The’ (1975), A Language for Life. Report of the Committee of Inquiry appointed

by the Secretary of State for Education and Science under the Chairmanship of Sir AlanBullock. London: HMSO.

Cathcart Borer, M. (1976), Willingly to School. A History of Women’s Education. Guildford andLondon: Lutterworth Press.

Christie, F. (1976), ‘The teaching of English in elementary schools in New South Wales1848–1900: an inquiry into social conditions and pedagogical theories determiningthe teaching of English’. Unpub. MEd Thesis, University of Sydney.

Christie, F.G. (2002), ‘The development of abstraction in adolescence in subject English’,in M.J. Schleppegrell and M.C. Colombi (eds), Developing Advanced Literacy in Firstand Second Languages. Meaning with Power. Mahwah, NJ and London: Erlbaum, pp.45–66.

Christie, F. (1993), ‘The “received tradition” of English teaching: the decline of rhetoricand the corruption of grammar’, in B. Green (ed.), The Insistence of the Letter. LiteracyStudies and Curriculum Theorizing (Critical Perspectives on Literacy and Education).London and Washington: Falmer Press, pp. 75–106.

Christie, F. (2003), ‘English in Australia’, in RELC Journal 34, 1, pp. 100–19.Christie, F. (2004a), ‘The study of language and subject English’, in Australian Review of

Applied Linguistics 27, 1, pp. 15–29.Christie, F. (2004b), ‘Revisiting some old themes: the role of grammar in the teaching of

English’, in J.A. Foley (ed.), Language, Education and Discourse. Functional Approaches.London and NY: Continuum, pp. 145–73.

Christie, F. (2004c), ‘Authority and its role in the pedagogic relationship of teaching’, inL. Young and C. Harrison (eds), Systemic Functional Linguistics and Critical DiscourseAnalysis. Studies in Social Change. London and NY: Continuum, pp. 173–201.

Christie, F. (2005), Language Education in the Primary Years. Sydney: University of NewSouth Wales Press.

Christie, F., Gray, P., Gray, B., Macken, M., Martin, J. and Rothery, J. (1990), ExploringProcedures: Books 1–4 and Teachers Book. Sydney: Harcourt Brace Jovanovich.


http://www.boardofstudies.nsw.edu.au

http://www.boardofstudies.nsw.edu.au

Christie, F., Gray, P., Gray, B., Macken, M., Martin, J. and Rothery, J. (1992a), ExploringExplanations: Books 1–4 and Teachers Book. Sydney: Harcourt Brace Jovanovich.

Christie, F., Gray, P., Gray, B., Macken, M., Martin, J. and Rothery, J. (1992b), ExploringReports: Books 1–4 and Teachers Book. Sydney: Harcourt Brace Jovanovich.

Christie, F. and Soosai, A. (2000), Language and Meaning 1. Melbourne: Macmillan Edu-cation.

Christie, F. and Soosai, A. (2001), Language and Meaning 2. Melbourne: Macmillan Edu-cation.

Christie, F. and Unsworth, L. (2005), ‘Developing dimensions of an educational linguis-tics’, in R. Hasan, C. Matthiessen and J.J. Webster (eds), Continuing Discourseon Language. A Functional Perspective, Vol.1. London and Oakville: Equinox, pp.217–50.

Christie, F. (in press), ‘Literacy and current debates over reading’, in R.Whittaker,M. O’Donnell and A. McCabe (eds), Language and Literacy: Functional Perspectives.London and NY: Continuum.

Cope, B. and Kalantzis, M. (eds) (2000), Multiliteracies. Literacy Learning and the Design ofSocial Futures. South Yarra, Victoria: Macmillan.

Cox National Curriculum English Working Group, The (1989), English for Ages 5 to 16:Proposals of the Secretary of State for Education and Science and the Secretary of State for Wales.London: Department of Education and Science and the Welsh Office.

Davies, F. (1973), Teaching Reading in Early England. London: Pitman Publishing.Derewianka, B. (1990), Exploring How Texts Work. Sydney: PETA. Derewianka. B. (1998), A Grammar Companion for Primary Teachers. Sydney: PETA.Derewianka, B. (2003), ‘Grammatical metaphor in the transition to adolescence’, in

A.M. Simon-Vandenbergen, M. Taverniers and L. Ravelli (eds), GrammaticalMetaphor. Amsterdam and Philadelphia: Benjamins, pp. 185–220.

Dixon, J. (1967), Growth through English. London: National Association for the Teachingof English and Oxford University Press.

Dixon, J. (1975), Growth through English (third edn). London: National Association for theTeaching of English and Oxford University Press.

Droga, L. and Humphrey, S. (2002), Getting Started with Functional Grammar. Berry, NSW:Target Texts.

Droga, L. and Humphrey, S. (2003), Grammar and Meaning. An Introduction for PrimaryTeachers. Berry, NSW: Target Texts.

‘English: a Statement for Australian Schools’. A Joint Project of the States, Territories and the Com-monwealth of Australia initiated by the Australian Education Council (1994). Carlton,Victoria: Curriculum Corporation.

‘English Curriculum and Standards Framework’, Victoria. (http://www.vcaa.vic.edu.au/prep10/csf/ accessed 30 December 2005)

Fitch, J. (1880), Lectures on Teaching. Cambridge: CambridgeUniversity Press.Gee, J.P. (1996), Social Linguistics and Literacies: Ideology in Discourses (second edn).

London: Taylor and Francis. Goldstrom, J.M. (1972), Education: Elementary Education 1780–1900. (David and Charles

Sources for Social and Economic History). Devon: David and Charles.Gosden, P.H.J.H. (1969), How They were Taught. An Anthology of Contemporary Accounts of

Learning and Teaching in England 1800–1950. Oxford: Blackwell.Gray, B. (1998), ‘Accessing the Discourses of Schooling: English Language and Literacy

Development with Aboriginal Children in Mainstream Schools’. Unpublished PhDthesis, University of Melbourne.

Halliday, M.A.K. (1994), An Introduction to Functional Grammar (second edn). London:Edward Arnold.

Halliday, M.A.K., McIntosh, A. and Strevens, P. (1964), The Linguistic Sciences and LanguageTeaching. London: Longmans.

Halliday, M.A.K. and Hasan, R. (in press), ‘Retrospective on systemic functional linguis-


http://www.vcaa.vic.edu.au/prep10/csf/

http://www.vcaa.vic.edu.au/prep10/csf/

182

tics and literacy’, in Whittaker, R., O’Donnell, M. and McCabe, A. (eds), Language &Literacy: Functional Approaches. London and Washington: Continuum.

Hasan, R. and Martin, J.R. (1989), Language Development: Learning Language, LearningCulture. Meaning and Choice in Language: Studies for Michael Halliday. (Advances inDiscourse Processes Series, Vol. XXVII). Norwood, NJ: Ablex.

Hasan, R., Matthiessen, C. and Webster, J.J. (2005), Continuing Discourse on Language.A Functional Perspective, Vol. 1. London and Oakville: Equinox.

Hawashir, G.E. and Selfe, C.LK. (2000), Global Literacies and the World Wide Web (LiteraciesSeries). London and NY: Routledge.

Heath, S.B. (1983), Ways with Words. Cambridge: Cambridge University Press. Hunter, I. (1994), Rethinking the School. Subjectivity Bureaucracy Criticism. St Leonards, NSW:

Allen and Unwin.Kachru, B.J. (1990), ‘World Englishes and applied linguistics’, in M.A.K. Halliday, J.

Gibbons and H. Nicholas (eds), Learning Keeping and Using Language, Vol. 11. Am-sterdam and Philadelphia: Benjamins, pp. 203–29.

‘Kingman Report, The’ (1988), Report of the Committee of Inquiry into the Teaching ofEnglish Language. Committee chaired by Sir John Kingman. London: HMSO.

Kress, G. (2000), ‘Multimodality’, in Cope and Kalantzis, pp.182–202.Kress, G. (2003), Literacy in the New Media Age (Literacies Series). London and NY: Rout-

ledge.Leavis, F.R. (1933), For Continuity. Cambridge: The Minority Press. Leavis, F.R. (1934), ‘Introduction’, in Determinations. Critical Essays. London: Chattus and

Windus, pp. 1–9.Leavis, F.R. (1965), The Common Pursuit. London: Chatto and Windus. Leavis, F.R. and Thompson, D. (1933), Culture and Environment. The Training of Critical

Awareness (this edition 1964). London: Chatto and Windus. ‘Literacy for all, the Challenge for Australian Schools’ (1998). Canberra: Commonwealth

Department of Employment, Education and Training. Macken-Horarik, M. (1998), ‘Exploring the requirements of critical school literacy: a view

from two classrooms’, in F. Christie and R. Misson (eds), Literacy and Schooling.London and NY: Routledge, pp. 74–103.

Macken-Horarik, M. (2002), ‘“Something to shoot for”: a systemic functional approach toteaching genre in secondary school science’, in A.M. Johns (ed.), Genre in the Class-room. Multiple Perspectives. Mahwah, NJ: Erlbaum, pp. 17–42.

Macken-Horarik, M. (2003), ‘Appraisal and the Special Instructiveness of Narrative’, inMacken-Horarik, M. and Martin, J. R. (eds), Text 23, 2, pp. 285–312.

Macken-Horarik, M. (2004), ‘Interacting with the Multimodal Text: reflections on imageand verbiage in ArtExpress’, in Visual Communication 3 (1), pp. 5–26.

Macken-Horarik, M. (2006), ‘Recognizing and realizing “what counts” in examinationEnglish: Perspectives from systemic functional linguistics and code theory’, in Func-tions of Language 13, 1; pp. 1–35.

Martin, J.R. and Rose, D. (2003), Working with Discourse. Meaning beyond the Clause. Londonand NY: Continuum.

Martin, J.R. and Rose, D. (2006), Genre Relations. London and Oakville: Equinox.Martin, J. R. and White, P.R.R. (2005), The Language of Evaluation: Appraisal in English.

Hampshire: Palgrave Macmillan.Maybin, J. (2000), ‘The canon. Historical construction and contemporary challenges’, in

J. Davison and J. Moss (eds), Issues in English Teaching. (Issues in Subject Teaching Series).London and NY: Routledge, pp. 180–95.

Muller, J. (2000), Reclaiming Knowledge. Social Theory, Curriculum and Education Policy.London: RoutledgeFalmer.

Muller, J. (2001), ‘Intimations of boundlessness’, in A. Morais, I. Neves, B. Davies andH. Daniels (eds), Towards a Sociology of Pedagogy. The Contribution of Basil Bernstein toResearch. (History of Schools and Schooling Series, Vol. 23). NY: Peter Lang, pp. 129–52.


Muller, J. (2004), ‘On the shoulders of giants: a digression on knowledge, curriculum,and finally, the teacher’. A plenary paper given at the Conference on Reclaiming Knowl-edge: Registers of Discourse in the Community and School, held at the University of Sydney,13–15 December 2004.

Murray, L. (1795, 1968), English Grammar. Facsimile Reprint No.106. Menston: Scolar Press. ‘National English Curriculum, The’ (1995), London: Department for Education and

Science and HMSO.‘Newbolt Report, The’ (1921), Report on the Teaching of English in England: being the Report

of the Departmental Committee appointed by the President of the Board of Education to inquireinto the Position of English in the Educational System of England. London: HMSO.

New London Group (1996), ‘A pedagogy of multiliteracies: designing social futures’, inHarvard Educational Review 66, pp. 60–92.

New South Wales Board of Studies Senior English Syllabus. http://www.education.tas.gov.au/english/sylrev.htm, accessed 7 March 2006.

Peel, R., Patterson, A. and Gerlach, J. (2000), Questions of English. Ethics, Aesthetics, Rhetoric,and the Formation of the Subject in England, Australia and the United States. London andNY: RoutledgeFalmer.

‘Plowden Report, The’ (1967), Children and their Primary Schools. London: HMSO. Stenhouse, L. (1975), An Introduction to Curriculum Research and Development. London:

Heinemann Educational Books. Street, B. (1984), Literacy in Theory and Practice. Cambridge: Cambridge University Press.Street, B. (ed.) (1993), Cross-cultural Approaches to Literacy. Cambridge, NY and Oakleigh,

Victoria: Cambridge University Press. Street, B. (1997), ‘The implications of the “New Literacy Studies” for literacy education’,

in English in Education, Vol. 321, No. 3, pp. 45–59.Street, B. (ed.) (2001), Literacy and Development. London and NY: Routledge.Sutherland, G. (ed.) (1973), Arnold on Education. London and Melbourne: Penguin

Education.Unsworth, L. (ed.) (2000), Researching Language in Schools and Communities. Functional

Linguistic Perspectives. London and Washington: Cassell.Unsworth, L. (2001), Teaching Mutiliteracies across the Curriculum. Changing Contexts and

Images in Classroom Practice. Buckingham and Philadelphia: Open University Press.Unsworth, L., Thomas, A., Simpson, A. and Asha, J. (2005), Children’s Literature and

Computer Based Teaching. Maidenhead: Open University Press.

Notes

1 The authors acknowledge the assistance of Joan Rothery who contributed to an earlydraft of this paper which was delivered at a conference on ‘Reclaiming Knowledge’at the University of Sydney in December 2004.

2 English is a major international language, spoken by millions. Significant policystatements attach to the teaching of English in the many countries in which it is notthe mother tongue, though they are not our concern. This discussion concerns the‘Inner Circle’ countries (Kachru, 1990: pp. 203–29) of English speakers, though itdraws most on the experience of England and Australia.

3 Maybin, 2000, p. 184, says the Newbolt Report was the culmination of the work of theEnglish Association, formed in 1907, though this does not in itself diminish thesignificance of the First World War as a contributing factor.

4 As this book went to press, an Australian National Inquiry into the Teaching of Literacy,had produced a Report on Teaching Reading, released in January 2006.http://www.dest.gov.au/nitl/report.htm (accessed 8 March 2006)

5 The nature of this test has changed since 1995, though similar tasks are still used.


http://www.education.tas.gov.au/english/sylrev.htm

http://www.education.tas.gov.au/english/sylrev.htm

http://www.dest.gov.au/nitl/report.htm

Introduction

Bernstein (1999) distinguishes between what he refers to as vertical and hori-zontal discourses. He defines a vertical discourse as follows: ‘a vertical discoursetakes the form of a coherent, explicit, and systematically principled structure,hierarchically organised, as in the sciences, or it takes the form of a series ofspecialised modes of interrogation and specialised criteria for the productionand circulation of texts, as in the social sciences and humanities’ (Bernstein1999: 159). In contrast: ‘a horizontal discourse entails a set of strategies whichare local, segmentally organised, context specific and dependent, for maximis-ing encounters with persons and habits’ (Bernstein 1999: 159).

In addition to vertical and horizontal discourses, Bernstein also characterizesvertical discourses as having either hierarchical knowledge structures (thephysical sciences) or horizontal knowledge structures (the humanities andsocial sciences) (Bernstein 1999: 162). Bernstein represents a hierarchicalknowledge structure as a triangle:

And he states: ‘This form of knowledge attempts to create very general propo-sitions and theories, which integrate knowledge at lower levels, and in this wayshows underlying uniformities across an expanding of apparently differentphenomena. Hierarchical knowledge structures appear, by their users, to bemotivated towards greater and greater integrating propositions, operating atmore and more abstract levels. Thus, it could be said that hierarchical know-ledge structures are produced by an “integrating” code’ (Bernstein 1999: 162).

In contrast Bernstein (1999: 162) describes horizontal knowledge structuresas consisting of ‘a series of specialised languages with specialised modes ofinterrogation and criteria for the construction of specialised texts. Thus, anyone of the specialised disciplines within the form of a horizontal knowledgestructure found within the humanities can be visually portrayed as: L1 L2 L3 L4

L5 L6 L7 . . . Ln’ (Bernstein 1999: 162).To continue the quote:

9 Vertical and horizontal discourse and the social sciences

Peter Wignell

Thus, in the case of English literature, the languages would be the specialisedlanguages of criticism; in Philosophy, the various languages of this mode ofinquiry and in Sociology . . . the languages refer, for example, to functionalism,post-structuralism, post-modernism, Marxism, etc. (Bernstein 1999: 162)

Bernstein’s classification of the physical sciences as hierarchical knowledgestructures and the humanities as horizontal knowledge structures can be cali-brated quite readily with systemic functional work on the discourses of thephysical sciences and humanities. Hierarchical knowledge structures more orless correlate with discourses which establish an elaborate technical frameworkas a means of categorizing and interpreting the world. The physical sciences area prime example. Horizontal knowledge structures more or less correlate withdiscourses which use abstraction (see Martin, 1997) as their major resource forinterpretation, such as the humanities. The case of the social sciences, however,is not so clear.

Wignell (1997) proposes that the discourse of social science is different inkind from both the physical sciences and the humanities but, because of itsorigins and evolution, shares some features of both. In a very brief summary ofthe conclusions of a substantial body of research, the physical sciences use tech-nicality as a primary resource for naming, categorizing, analysing and inter-preting the world, whereas the humanities use abstraction as their principalresource. The social sciences are both abstract and technical, having evolvedover several hundred years as a discourse which had its origins in the abstracthumanities discourse of moral philosophy (how people should behave), butover time that abstract discourse evolved into one which is both technical andabstract, resulting in the social sciences (how humans do behave).

Wignell (1990, 1992, 1997, 1998) argues that the language of what came tobe the social sciences evolved as a hybrid of the languages of the humanities andof the physical sciences, with, over time, the language patterns of the physicalsciences taking a more and more prominent role. This suggests that the socialsciences are, to varying degrees, likely to be hierarchical rather than horizontalknowledge structures. Wignell’s (1997) work certainly implies this. This workanalysed the orthodoxy of three social science disciplines – economics, sociol-ogy and political science – as presented to undergraduate students. In analysingthe orthodox this work did not examine what happens to unorthodox voiceswithin the disciplines. It is in focusing on the orthodox that a conclusion thatthe social sciences are hierarchical knowledge structures might be favoured.

In an effort to resolve this issue, what I propose to do in this chapter is topresent a diachronic perspective on the social sciences and then refer back toBernstein’s analysis. What I hope to demonstrate by using a number of samplesfrom several canonical texts from the history of social science is the buildingover time of a hierarchical knowledge structure which then ruptured into anumber of smaller knowledge structures. That is, what are now known as thesocial sciences began as and evolved as a single hierarchical knowledge structurefor over two hundred years, and then experienced a Tower of Babel syndrometowards the end of the nineteenth century. What I hope to discover from myanalysis and discussion is what kind of knowledge structures resulted: did what

185VERTICAL AND HORIZONTAL DISCOURSE • WIGNELL

186

became the various social sciences then evolve into horizontal knowledge struc-tures, or continue as smaller-scale hierarchical knowledge structures?

I will be basing my discussion on examples from some key texts in the historyof social science. The texts are Thomas Hobbes’ Leviathan (1651), AdamSmith’s The Wealth of Nations (1776), David Ricardo’s Principles of Political Economyand Taxation (1817) and Karl Marx’s Capital, Volume One (1867).

The evolution of the social sciences

What became the social sciences was originally called civil philosophy byThomas Hobbes in the mid seventeenth century. Then, from the late eight-eenth century to around the end of the nineteenth century, the discipline thatevolved into the social sciences was called political economy. What I hope todemonstrate is that political economy grew as a hierarchical knowledgestructure until Marx. After Marx a process of fragmentation set in. Part of thediscipline of political economy continued on under another name as Marxism.The rest of it either broke into or spawned other major social science disciplinessuch as sociology, economics and political science.

What is recognizable as social science, or the application of the discourse ofscience to discussion of the social, appears to have begun with Thomas Hobbes.The extract below is from Hobbes’ Leviathan (1651). It is an example of a dis-course of the same kind as the example from Newton’s Treatise on Opticks(published in 1704; written 1675–87). Halliday (1988) treats Newton’s Treatise onOpticks as ‘registering the birth of scientific English’. Hobbes was writing in asimilar style a quarter of a century before Newton’s work was written.

The two examples are far from identical. The example from Newton is a descrip-tion of a set of experiments. Hobbes’ example is an explanation, written in a ‘thisis how things are’ manner without experiments as verification. This is a differencein method rather than a difference in kind. I argue that the two examples are closeenough in kind to be considered as belonging to the same discourse.

The cause of sense, is the Externall Body, or Object, which presseth the organproper to each sense, either immediately, as in the Tast and Touch, or mediately,as in Seeing, Hearing, and Smelling which pressure, by the mediation of Nerves,and other strings, and membranes of the body, continued inwards to the Brain,and Heart, causeth there a resistance, or counter pressure, or endeavour of theheart, to deliver it self: which endeavour because Outward, seemeth to be somematter without. And this seeming, or fancy, is that which men call sense; and con-sisteth, as to the Eye, in a light, or Colour figured; To the Eare, in a sound; To theNostrill, in an Odour; To the Tongue and Palat, in a Savour; And to the rest ofthe body, in Heat, Cold, Hardnesse, Softnesse, and such other qualities, as wediscern by feeling. All of which qualities called Sensible, are in the object thatcauseth them, but so many several motions of the matter, by which it presseth theorgans diversly.


Extract One: ‘The causes of sense’ – Hobbes (1651: 85)

I found moreover, that when Light goes out of air through several contingousrefracting Mediums as through Water and Glass, and thence goes out againinto Air, whether the refracting Superfices be parallel or inclin’d to oneanother, that Light as often as by contrary Refractions ’tis so corrected, that itemergeth in Lines parallel to those in which it was incident, continues everafter to be white. But if the emergent Rays be inclined to the incident, theWhiteness of the emerging Light will be by degrees in passing on from thePlace of Emergence, becoming tinged in its Edges with Colours. This I try’d byrefracting Light with Prisms of Glass placed within a Prismatick Vessel of Water.Now those Colours argue a diverging and separation of the heterogeneous Raysfrom one another by means of their unequal Refractions, as in what follows willmore fully appear. And, on the contrary, the permanent argues, that in likeIncidences of the Rays, and by consequence no inequality of their wholeRefractions. Whence I seem to gather the two following theorems.

What Hobbes did was apply the emerging discourse of the physical sciencesto discussions of the social, which, up until then, had been the domain of thehumanities, notably moral philosophy. Hobbes did not adopt the discourse ofscience and apply it to the social for its own sake: Hobbes was not writing atreatise on sense. His goal was similar to that of moral philosophers: to instructpeople on how they should behave. Hobbes’ main task in the first half ofLeviathan is to present an irrefutable argument for the need for sovereigngovernment. The example above comes from Chapter 1 of Leviathan, whichestablishes one of the fundamental premises of Hobbes’ argument: that humansare analogous to bodies in motion. In trying to arrive at first principles Hobbesneeds to establish the argument that the primary cause of motion in humans issensory input. In paraphrase the argument runs along the following lines:sensory input creates appetites and aversions in individuals. These are unlimitedand, in a state of nature, each individual has complete liberty to act on them.The inevitable result of this is a ‘warre of all against all’, or total anarchy. Thisstate is intolerable, therefore some form of sovereign government is needed tokeep people away from each other’s throats.

Incidentally, references to a ‘state of nature’ are a recurring theme in Enlight-enment discourse. This hypothetical state is meant to be the state humans livedin between the Flood and when they became ‘civilised’. Each writer who refersto it tends to construct it for his own purposes: for Thomas Hobbes it is a hellon earth, for John Locke it is a liberal property owner’s paradise, for AdamSmith it is an entrepreneur’s Eden, and for Karl Marx it is a prototypicalcommunist utopia.

The ‘science’ in Hobbes is there to further the argument. But why? We needto look at this in the context of the times. Hobbes was writing at a time of socialturmoil in England. He was trying to convince the people who could do some-thing about it to do something about it. I will not speculate on whether or notHobbes chose to use the discourse of the physical sciences simply because it


Extract Two: Newton 1704 (from Halliday 1988)

188

would further his cause best, or because he believed in it. Either way it was avoice that people of influence were likely to listen to.

The methods of the physical sciences were, in Hobbes’ time, in the process ofcompletely overturning prevailing notions about the nature of the universe.The combination of observation, experimentation and calculation hadprovided (or almost provided in Hobbes’ time) proof of a uniform, solar-centred cosmos governed by universal, mathematically derived physical laws.The emerging discourse of the physical sciences was, even then, a very powerfuldiscourse, not yet quite ‘naturalised’ or hegemonic (although it was shortlyafter, post Newton), but it was a voice with status that was listened to, especiallyby the emerging power brokers, the bourgeoisie. The physical sciences hadstatus and were gaining more: science was a discourse on the ascent. It wasascending not only because it could prove things but also, and just as impor-tantly, because it could, through its application to technology, produce things.Therefore, in using a discourse most likely to convince, Hobbes made thelogical choice. In economic terms he adopted the discourse with the most pur-chasing power.

Hobbes lived in the early (or late-early) stages of what is often referred to asthe scientific revolution. This revolution did not happen in isolation, but waspart of a broader sweep of social change. In terms of class, the bourgeoisie wasascending in economic and political power. In conjunction with this revolution,liberal, individualistic and utilitarian ideas were ascending in philosophy andcapitalism was in its early stages. Both material and semiotic relations wereshifting in favour of the bourgeoisie. The social class that derived its wealth andprestige from manufacture and trade had more to gain from the application ofscience to technology than the class that relied on ownership of land as itswealth and power base.

Science, however, was not only in the interests of the bourgeoisie. Especiallyin its applications, science had direct and indirect benefits for all of those whowere dominant in society at the time: the aristocracy, the church, the military,the emerging capitalists, even intellectuals and scientists themselves. Theseinterests can be summed up in six words: discover, conquer, plunder, convert,civilize and trade. Discover and conquer happen sequentially. The rest happenconcurrently. Improvements in shipbuilding, navigation and ballistics enabledrelatively large numbers of well-armed men to be moved around the world tomore places, and more quickly, than ever before. What emerged was a verypowerful discursive alliance between the established and the emerging becausethere was something in it for all of them.

Hobbes did not change completely the way the social was written about.Philosophers continued, after Hobbes, to write like philosophers, not likephysical scientists. Hobbes did, however, introduce a possibility, a new choice.He negotiated the potential for a new discourse to evolve. This choice was takenup 125 years after Leviathan by Adam Smith.

Adam Smith’s The Wealth of Nations (1776) marks the second beginning of thesocial sciences with the discipline of political economy. By the time Adam Smithwas writing in the late eighteenth century, the discourse of physical science waswell developed and virtually the only ‘recognised’ paradigm for explaining the


physical world. In the broader social context, the bourgeoisie had attained polit-ical dominance, the philosophical climate (in England) was liberal, individual-istic and utilitarian, and early capitalism (pre-large-scale industrialism) was thedominant economic practice.

Hobbes, therefore, cannot be given sole credit for influencing Smith. Hadthe physical sciences not been the dominant paradigm for interpreting theworld there would have been little point in drawing on Hobbes. Likewise, hadself-interested individualism and utilitarianism not been the dominant themesof British philosophy, they would not necessarily have been taken up by Smithas assumptions about human behaviour. There are, however, identifiablelegacies of Hobbes in Smith. As well as direct citation, the use of the ‘state ofnature’ and mechanistic, self-interested assumptions about human nature, ifnot originating with Hobbes, were given explicit voice by him. Smith’s reason-ing in the early chapters of The Wealth of Nations is very similar to Hobbes’reasoning in Leviathan until Smith begins to build field. In building field, Smithuses the same set of resources as are used in the physical sciences.

Adam Smith began the building of the hierarchical knowledge structure thatevolved into the social sciences of today, and he did it while trying to work outone of humanity’s fundamental questions: what are things really worth? Indoing this he used Hobbes as his point of departure through the followingquote:

The Value, or Worth of a man, is as in all other things, his Price; that is to say, somuch as would be given for the use of his power: and therefore is not absolute;but a thing dependant on the need and judgement of another.

Wealth, as Mr Hobbes says, is power. (Hobbes, T. 1651, cited in Smith, A. 1776:26)

Smith then immediately continues by linking the citation to Hobbes to his ownwork:

But the person who either acquires, or succeeds to a great fortune, does not nec-essarily acquire or succeed to any political power, either civil or military . . . Thepower which that possession immediately and directly conveys to him, is thepower of purchasing: a certain command over all the labour, or over all theproduce of labour, which is in the market . . . (Smith, A. 1776: 26)

I will take up this question of value and look at how it is treated in examples fromthe works of three political economists: Adam Smith, David Ricardo and KarlMarx. These examples will demonstrate the emergence of social science as asmall but growing hierarchical knowledge structure. What happened, in brief,was that Smith established a technical, theoretical framework, using much thesame resources as those used by the physical sciences. This framework was chal-lenged, critiqued, modified and added to by Ricardo. Both Smith and Ricardowere then challenged, critiqued, modified and added to by Marx. What is impor-tant in this critique and modification is that they all used the same language. Thecritique is from within the field and is used to alter and build the field.


190

In the first four chapters of The Wealth of Nations a number of extended lexicalmotifs are developed. Three of these are crucial. There is a motif of trade: barter,buy, sell, exchange, trade, money, price, market, commerce, goods, commodities. There isa motif of labour: toil, work, labour, division of labour. There is a motif of quantifi-cation: amount, number, quantity, value. These motifs are realized both congru-ently and metaphorically but, as they continue, the metaphorical forms are thepreferred realization and can thus be regarded as abstractions.

At the end of Chapter 4 Smith makes a major shift. Some of the abstractionsfrom these lexical motifs are distilled into technicality. This pivotal point marksboth the beginning of Western economic theory as a technical discipline, andthe new beginning of what evolved into the social sciences.

It is in this manner that money has become in all civilised countries the uni-versal instrument of commerce, by the intervention of which goods of all kindsare bought and sold, or exchanged for one another. What are the rules whichmen naturally observe in exchanging them either for money or for oneanother, I shall now proceed to examine.

These rules determine what may be called the relative or exchangeable valueof goods.

The word value, it is to be observed, has two different meanings, and some-times expresses the utility of some particular object, and sometimes the powerof purchasing other goods which the possession of that object conveys.

The one may be called ‘value in use’; the other, ‘value in exchange’.The things which have the greatest value in use have frequently little or no

value in exchange; and, on the contrary, those which have the greatest value inexchange have frequently little or no value in use.

In order to investigate the principles which regulate the exchangeable valueof commodities, I shall endeavour to show, First, what is the real measure of thisexchangeable value; or wherein consists the real price of all commodities.

Second, what are the different parts of which this real price is composed ormade up.

And, lastly, what are the different circumstances which sometimes raise someor all of these different parts of price above, and sometimes sink them belowtheir natural or ordinary rate; or, what are the causes which sometimes hinderthe market price, that is, the actual price of commodities, from coincidingexactly with what may be called their natural price.

In this extract Smith for the first time explicitly does what the physicalsciences do. He defines technical terms, begins to arrange them taxonomically,and says that he is about to explain how things work. This emerging technicalframework is to be deployed in the search for semiotic abstractions: principles,and institutional abstractions: rules. Or, in Bernstein’s terms, ‘very generalpropositions and theories’ (Bernstein 1999: 162).

The following taxonomy of value can be drawn up from the passage above:


Extract Three: Adam Smith 1776: 24

value

value in use value in exchange(exchangeable value)

real price market price(natural price) (actual price)

It is this initial technicality which is used for analysis in the rest of the chaptersin Book One of The Wealth of Nations.

The next chapter, Chapter 5, integrates the lexical motif of labour into thisnew technical framework by arguing that labour is the true measure of value:

First, is a justification for why labour is used as the measure of value.‘The value of any commodity, therefore, to the person who possesses it, andwho means not to use or consume it himself, but to exchange it for other com-modities, is equal to the quantity of labour which it enables him to purchase orcommand. Labour, therefore, is the real measure of the exchangeable value ofall commodities.’

‘What is bought with money or with goods is purchased by labour as muchas what we acquire by the toil of our own body. That money or those goodsindeed save us this toil. They contain the value of a certain quantity of labourwhich we exchange for what is supposed at the time to contain the value of anequal quantity. Labour was the first price, the original purchase-money that waspaid for all things.’

Second, the difficulty of seeing labour as the underlying measure of value.‘But though labour be the real measure of the exchangeable value of com-modities, it is not that by which their value is commonly estimated.’

Next, money as the conventional measure of value.‘Every commodity, besides, is more frequently exchanged for, and therebycompared with, other commodities than with labour. It is more natural, there-fore, to estimate its exchangeable value by the quantity of some other com-modity than by that of the labour which it can purchase.’

‘But when barter ceases, and money has become the common instrument ofcommerce, every particular commodity is more frequently exchanged formoney than for any other commodity.’


Figure 9.1 Taxonomy of value

Extract Four: Smith 1776: 26–9

192

Then, money itself is variable in value, therefore it is not an accurate measureof value.‘Gold and silver, however, like every other commodity, vary in their value, aresometimes cheaper and sometimes dearer, sometimes of easier and sometimesof more difficult purchase.’

‘Labour alone, therefore, never varying in its own value, is alone the ultimateand real standard by which the value of all commodities can at all times andplaces be estimated and compared. It is their real price; money is but theirnominal price only.’

Apparent variation in the value of labour is an illusion.‘But though equal quantities of labour are always of equal value to thelabourer, yet to the person who employs him they appear sometimes to be ofgreater and sometimes of smaller value. He purchases them sometimes with agreater and sometimes with a smaller quantity of goods, and to him the priceof labour seems to vary like that of all other things. In reality, however, it is thegoods which are cheap in the one case, and dear in the other.

In this popular sense, therefore, labour, like commodities, may be said tohave a real and a nominal price. Its real price may be said to consist in thequantity of the necessaries and conveniences of life which are given for it; itsnominal price, in the quantity of money. The labourer is rich or poor, is well orill rewarded, in proportion to the real, not to the nominal price of his labour.’

It should be noted that much of this argument is carried out using the techni-cal framework that has been established previously. For example, exchangeablevalue, real price and nominal price (synonymous with market price) are all used inthe argument. This technical framework is interwoven with the motif of meas-uring through the use of nominal group structure.

How market price is determined is then explained. At the same time new techni-cal terms (natural price, effectual demand and absolute demand) are introduced.

The market price of every particular commodity is regulated by the proportionbetween the quantity which is actually brought to the market, and the demandof those who are willing to pay the natural price of the commodity, or the wholevalue of the rent, labour, and profit, which must be paid in order to bring itthither. Such people may be called the effectual demanders, and their demandthe effectual demand; since it may be sufficient to effectuate the bringing ofthe commodity to market. It is different from the absolute demand.


Extract Five: Building technicality (from Smith, A. 1776: 49)

The actual price [at which any commodityis commonly sold for]

Value

is called its market price.

Token

(Smith, A. 1776: 49)

Compare the extract above with an example of science discourse from about thesame time.

When the equilibrium of this fluid in any body is not disturbed; that is, whenthere is in any body neither more nor less of it than its natural share, or thanthat quantity which it is capable of retaining by its own attraction, it does notdiscover itself to our senses by any effect. The action of the rubber upon anelectric disturbs this equilibrium, occasioning a deficiency in one place, and aredundancy of it in another.

Smith uses much the same grammatical resources as those used by Priestley. Theextract from Smith parallels the discourse of the physical sciences of its owntime.

I will now skip forward 41 years to David Ricardo’s Principles of Political Economyand Taxation (1817). By Ricardo’s time industrial capitalism had become thedominant mode of production, distribution and exchange in England. Enlight-enment, liberal and utilitarian philosophical trends were fully established andthe bourgeoisie were becoming dominant in politics and economics.

Ricardo takes Smith to task on what should be the measure of value. WhatSmith introduced at the end of a chapter and then developed, Ricardo nowpicks up as his point of departure.

It has been observed by Adam Smith that ‘the word Value has two differentmeanings, and sometimes expresses the utility (of some particular object), andsometimes the power of purchasing other goods which the possession of thatobject conveys. The one may be called value in use; the other value in exchange. Thethings’, he continues, ‘which have the greatest value in use, have frequently littleor no value in exchange; and, on the contrary, those which have the greatest valuein exchange, have little or no value in use.’(Ricardo, D. 1817: 5)

Ricardo then moves into dispute with Smith’s analysis. Ricardo uses an extensivequote from Smith as his point of departure. The quote outlines Smith’s positionthat the value of something is determined by labour:

‘The real price of everything,’ says Adam Smith, ‘what everything really costs tothe man who wants to acquire it, is the toil and trouble of acquiring it. What every-thing is really worth to the man who has acquired it, and who wants to dispose; ofit, or exchange it for something else, is the toil and trouble, which it can save tohimself, and which it can impose upon other people.’ ‘Labour was the first price– the original purchase-money that was paid for all things.’ (Ricardo, D. 1817:6–7)

Here Ricardo contends that Smith erred in equating ‘the quantity of labourbestowed on the production of any object’ and ‘the quantity which it can


Extract Six: Comparison extract. From Priestley’s The History and Present Stateof Electricity, with Original Experiments (1757) (Taken from Halliday 1988)

194

command in the market’. Ricardo argues these two are different and that, interms of what it can ‘command in the market’, labour varies in value.

From this point Ricardo presents his alternative analysis of exchangeable value.This is principally done by supporting the argument by alternative explanations.For example:

In the same country double the quantity of labour may be required to produce agiven quantity of food and necessaries at one time than may be necessary atanother and a distant time; yet the labourer’s reward may possibly be very littlediminished. If the labourer’s wages at the former period were a certain quantityof food and necessaries, he probably could not have subsisted if that quantity hadbeen reduced. Food and necessaries in this case will have risen 100 per cent, ifestimated by the quantity of labour necessary to their production, while they willscarcely have increased in value if measured by the quantity of labour for whichthey will exchange.. . .It cannot then be correct to say with Adam Smith, ‘that as labour may sometimespurchase a greater and sometimes a smaller quantity of goods, it is their valuewhich varies, not that of the labour which purchases them;’ and therefore, ‘thatlabour, alone never varying in its own value, is alone the ultimate and realstandard by which the value of all commodities can at all times and places be esti-mated and compared;’ — but it is correct to say, as Adam Smith had previouslysaid, ‘that the proportion between the quantities of labour necessary for acquir-ing different objects seems to be the only circumstance which can afford any rulefor exchanging them for one another;’ or in other words that it is the compara-tive quantity of commodities which labour will produce that determines theirpresent or past relative value, and not the comparative quantities of commoditieswhich are given to the labourer in exchange for his labour. (Ricardo, D. 1817:8–9)

The second paragraph is Ricardo’s conclusion to his refutation of Smith’sanalysis. In his refutation Ricardo argues and concludes by using much the sameresources as were used by Smith. In arguing with Smith, Ricardo uses the samelanguage as Smith. What he is doing is fine-tuning Smith’s taxonomies andexplanations. A combination of technicality and abstraction carries the experi-ential weight of the argument. Value is not in dispute, but how it is measured is.Ricardo concludes that money should be used to measure value.

Suppose money to be that commodity. If a salmon were worth $1 and a deer $2,one deer would be worth two salmon. But a deer might become of the value ofthree salmon, for more labour might be required to obtain the deer, or less to getthe salmon, or both these causes might operate at the same time. If we had thisinvariable standard, we might easily ascertain in what degree either of thesecauses operated. If salmon continued to sell for $1 whilst deer rose to $3, wemight conclude that more labour was required to obtain the deer. If deer contin-ued at the same price of $2 and salmon sold for 13s 4d, we might then be surethat less labour was required to obtain the salmon; and if deer rose to $2 10s and


salmon fell to 16s 8d, we should be convinced that both causes had operated inproducing the alteration of the relative value of these commodities. (Ricardo, D.1817: 16)

We now skip forward another 50 years to Karl Marx. When Marx was writingindustrial capitalism was in full swing as the dominant mode of production,distribution and exchange in England.

In Capital, Volume One (1867), right at the beginning, in Chapter One, Marxchallenges Ricardo and Smith. Marx both uses and builds on the technicality ofthe discipline. He begins by making explicitly technical a term used as a genericabstraction but never fixed by definition by Smith and Ricardo: a commodity. Sixpages of Chapter 1 are taken up by argumentation which arrives at a final defi-nition of a commodity. The argumentation derives from the field as alreadyestablished, and in its course reorganizes the taxonomic relations of part of thefield.

Marx first defines a commodity as:

A commodity is, in the first place, an object outside us, a thing that by its proper-ties satisfies human wants of some sort or another. (Marx, K. 1867: 43)

Marx then uses the theory of value to assign attributes to a commodity:

The utility of a thing makes it a use value. But this utility is not a thing of air. Beinglimited by the physical properties of the commodity, it has no existence outsidethat commodity. A commodity, such as iron or corn, or a diamond, is therefore,so far as it is a material thing, a use-value, something useful.(Marx, K. 1867: 44)

Marx then introduces quantity and quality and uses these to rework the existingtheory of value to arrive at a different definition of a commodity. Marx reworksthe taxonomy established by Smith and assumed by Ricardo:

value

use value exchange value

The arguments Marx presents in reworking this taxonomy are somewhatcomplex. In paraphrase, using the notions of quality and quantity, Marxreverses the proposition, so that at first sight it appears that use value is quanti-tative and that exchange value, also quantitative, is derived from the exchange ofdifferent amounts of use value, that ‘exchange-value, at first sight, presents itselfas a quantitative relation, as the proportion in which values in use of one sortare exchanged for those of another sort’.

Marx proceeds to eliminate use value as a measure of value. He does this byarguing that, in order to be exchanged for one another, commodities must have


Figure 9.2 Taxonomy of value

196

something in common which can be used as a measure. He argues that differ-ent uses represent different qualities, not quantities, therefore use cannotmeasure value. Still, in order to be exchanged, they must have something incommon. Since commodities all have different uses, are all different, but areexchanged for relative quantities of each other, they must have something incommon besides their use value. That ‘something’ is external to the qualities ofcommodities themselves (use value). This external must be ‘a total abstractionfrom their use-value’. This ‘something’ is their exchange value. Use value is seenas a qualitative relation, while exchange value is seen as a quantitative relationwhich does ‘not contain one atom of use-value’. Use value is thus excluded fromany estimation of how much of one commodity will exchange for another. Whatall commodities have in common is that they are the ‘products of labour’. Thisis not any particular kind of labour but ‘human labour in the abstract’.

The taxonomy in Figure 9.2 above is rearranged so that value is no longer ina relation of hyponymy to the co-hyponyms use value and exchange value. Usevalue is removed, and value and exchange value become synonymous.

The following quote illustrates this:

. . . there is nothing left but what is common to them (commodities) all; all arereduced to one and the same sort of labour, human labour in the abstract.

Let us now consider the residue of each of these products; it consists of thesame insubstantial reality in each, a mere congelation of homogenous humanlabour, of labour power expended without regard to its mode of expenditure. Allthat these things tell us is, that human labour power has been expended in theirproduction, that human labour is embodied in them. When looked at as crystalsof this social substance, common to all of them, they are – Values. (Marx, K. 1867:46)

Here Marx also ‘solves’ a previous problem: that of different kinds of labour anddifferent intensities of labour over time being ‘worth’ more or less than eachother. By excluding use value as a qualitative relation Marx also excludes the dif-ferent qualities of labour.

As labour is averaged, so commodities are averaged: ‘each individual com-modity, in this connexion, is to be considered as an average sample of its class’(Marx 1867: 47).

Thus:

as values, all commodities are only definite masses of congealed labour-time.

Circumstantial role Token Value

(Marx, K. 1867: 47)

This involves a distillation of several preceding pages into a redefinition ofcommodities. A distinction between products and commodities is then intro-duced. Things produced for one’s own use are not regarded as commoditiesbecause these things do not enter into any relationship with other things: they


are not exchanged. Taking this addition into account, a final definition of com-modities concludes this section of the chapter. This definition is accomplishedby enhancement in a clause complex:

x β αTo become a commodity a product must be transferred to

another, who it will serve as a use-valueby means of an exchange

Token Value

(Marx, K. 1867: 48)

This final definition of a commodity is then used as the basis for Marx’s analysisof labour and of exchange value in the next two sections of the chapter.

The term labour was a continuing motif in the works of Smith and Ricardo.From Smith can be inferred a distinction between labour as the most super-ordinate abstraction and work as a more subordinate abstraction through hisdifferent lexical choices when referring to labour and work. This distinction isnot formalized: the words are simply used in different contexts, not defined.Neither does Ricardo make a formal distinction. Marx undertakes a formalanalysis of labour using the technical framework he has reordered in theprevious section. That is, Marx builds the field by making technical things thatwere latent earlier in the discourse (commodities and labour).

In analysing labour Marx uses the same method as he used to analyse com-modities. He takes up a point introduced in the previous section, human labourin the abstract, labour power. He distinguishes concrete (specific) labour fromlabour in the abstract and extracts qualitative differences in labour from calcula-tions of exchange value (value) just as he extracted use value from the determina-tion of value. Concrete labour, it is argued is, like use-value, qualitative. Marx (49)distinguishes concrete/useful labour from abstract labour by defining it:

This qualitative difference of labour allows exchange to happen. Things are notexchanged for things that are qualitatively the same: ‘coats are not exchangedfor coats, one use-value is not exchanged for another of the same kind’ (Marx,K. 1867: 49).

The above is restated more technically two paragraphs later:


the labour, [[whose utility isthus represented by the value inuse of its product]] or [[whichmanifests itself by making itsproduct a use-value]]

Value

we call useful labour.

Assigner Token

198

In the use-value of each commodity there is contained useful labour, ie produc-tive activity of a definite kind and exercised with a definite aim. Use values cannotconfront each other as commodities unless the useful labour embodied in themis qualitatively different in each of them. (Marx, K. 1867: 49)

Useful labour is then excluded:

Productive activity, if we leave out of sight its special form, viz., the useful charac-ter of labour, is nothing but the expenditure of human labour-power. Tailoringand weaving, though qualitatively different productive activities, are in each aproductive expenditure of human brains, nerves, and muscles, and in this senseare human labour. They are but two different modes of expending human labour-power . . . It is the expenditure of simple labour-power, ie., of the labour powerwhich, on an average, apart from any special development, exists in the organismof every ordinary individual . . . Skilled labour counts only as simple labour inten-sified, or rather, as multiplied simple labour, a given quantity of skilled being con-sidered equal to a greater quantity of simple labour . . . A commodity may be theproduct of the most skilled labour, but its value, by equating it to the product ofsimple unskilled labour, represents a definite quantity of the latter labour alone.(Marx, K. 1867: 51)

Marx has theorized labour so that it can be used to measure value (which he hasalso redefined):

Coats and linen, however, are not merely values, but values of a definite magni-tude, and according to our assumption, the coat is worth twice as much as the tenyards of linen. Whence this difference in values? It is owing to the fact that thelinen contains only half as much labour as the coat, and consequently, that in theproduction of the latter, labour-power must have been expended during twice thetime necessary for the production of the former.

While, therefore, with reference to use-value, the labour contained in a com-modity counts only qualitatively, with reference to value it counts only quantita-tively, and must first be reduced to human labour pure and simple. (Marx, K.1867: 52)

So far Marx has specified the field specific meanings of two terms, commoditiesand labour, previously regularly used but undefined and untheorized, and hasboth redefined value and reordered the taxonomy of kinds of value.

Having defined a commodity and proposed a measure for how to measure itsvalue, Marx now uses the new developed technical framework to analyse therelativities of exchanging one commodity for another. To borrow an analogy fromSystemic Functional Linguistics, the relationship is akin to the relationshipbetween Token and Value in a relational identifying clause: one thing repre-sents another. This is embodied in the following:


A. Elementary or Accidental form of value

x commodity A = y commodity B, orx commodity A is worth y commodity B.20 yards of linen = 1 coat, or20 yards of linen are worth 1 coat.

Here two different kinds of commodities (in our example the linen and the coat),evidently play two different parts. The linen expresses its value in the coat; thecoat serves as the material in which that value is expressed. The former plays anactive, the latter a passive, part. The value of the linen is represented as relativevalue, or appears in relative form. The coat officiates as equivalent, or appears inequivalent form. (Marx, K. 1867: 55)

coat = Token, linen = Value

This relationship between two commodities is the basis of exchange-value. The nextform expands this to include all commodities:

B. Total or Expanded form of value

z Com A = U Com B or v Com C or W Com D or x Com E or &c(20 yards of linen = 1 coat or = 10 lbs tea or = 40 lbs coffee or = 1 quarter corn or = 2 ounces gold or = 1/2 ton iron or = &c

The value of a single commodity, linen, for example, is now expressed in terms ofnumberless other elements of the world of commodities. Every other commoditynow becomes a mirror for linen’s value. It is thus, that for the first time, this valueshows itself in its true light as a congelation of undifferentiated human labour.(Marx, K. 1867: 68)

Marx argues that the first form of value could result from accident but that thisexpanded form could not. It is put up as proof of labour as the measure of value.From this Marx reaches a conclusion which reverses the conclusions of Smithand Ricardo:

It becomes plain, that it is not the exchange of commodities which regulates themagnitude of their value; but, on the contrary, that it is the magnitude of theirvalue which controls their exchange proportions. (Marx, K. 1867: 69)

Marx points out practical problems with this form of value; potentially every com-modity stands in relation to every other commodity. Everything is a potentialToken and a potential Value.

This leads to the next form:


200

C. The General form of value

1 coat10 lbs of tea40 lbs of coffee1 quarter of iron = 20 yards of linen2 ounces of gold1/2 a ton of ironx com A, &c

1.The altered character of the form of value

All commodities now express their value (1) in an elementary form, because in asingle commodity; (2) with unity, because in one and the same commodity.(Marx, K. 1867: 70)

Here the roles of Token and Value are reversed, and linen becomes the Tokenthrough which the values are realized.

From here it is a small step to the final form of value, the money-form:

The Money-form

20 yards of linen1 coat10 lbs of tea40 lbs of coffee = 2 ounces of gold1 qr of corn1/2 ton of ironx commodity A

In passing from Form A to Form B, and from the latter to Form C, the changesare fundamental. On the other hand, there is no difference between forms C andD except that, in the latter gold has assumed the equivalent form in the place oflinen. Gold is in Form D, what linen was in Form C – the universal equivalent. Theprogress consists in this alone, that the character of direct and universalexchangeability – in other words, that the universal equivalent form – has now, bysocial custom, become finally identified with the substance, gold. (Marx, K. 1867:75)

Gold has become the universal Token. While Smith and Ricardo both arguedthat money itself had value, Marx treats any value money might have as irrele-vant. Money now signifies, or represents, value: it is the Token that representsvalue.


Summary and conclusions

Having laboured through Hobbes, Smith, Ricardo and Marx, what have wefound out? What is illustrated is the slow building over a period of 200 years ofa hierarchical knowledge structure. Hobbes plants the seed, Smith begins thetheory, Ricardo and Marx refine it, rework it and add to it. In doing so they usethe language of the field that they are building. What they have done appearsto satisfy Bernstein’s criteria for a hierarchical knowledge structure:

This form of knowledge attempts to create very general propositions and theories,which integrate knowledge at lower levels, and in this way shows underlying uni-formities across an expanding of apparently different phenomena. Hierarchicalknowledge structures appear, by their users, to be motivated towards greater andgreater integrating propositions, operating at more and more abstract levels.Thus, it could be said that hierarchical knowledge structures are produced by an‘integrating’ code. (Bernstein 1999: 162)

Smith outlines the theory, arriving at very general propositions: propositionsthat are intended to apply over all cases and all time. Ricardo works throughSmith’s theory and finds what he thinks are a couple of propositions that don’tquite work. He doesn’t throw the theory out or use a different language to talkabout it. He uses the same kind of language and argument that Smith used,and uses it to refine the theory: he tries to make its general propositions workbetter.

Marx then takes the level of generality and abstraction further than eitherSmith or Ricardo. Marx used his economic analysis to build an integrated socialtheory. However, with Marx the discipline of political economy came to a fairlyabrupt halt. After Marx, arguments about and analysis of value just about dis-appear from capitalist economics: the focus is on price. If you look in just aboutany standard Western economics textbook you will find mention, often alongwith pictures, of canonical figures such as Smith, Ricardo, and even Keynes, butno mention, and certainly no picture, of Marx.

What happened after Marx? What I speculate happened with Marx was thathis economic analysis was so flawless and his political/social analysis so threat-ening to those wielding power at the time that the best way for them to deal withhim was simultaneously to ignore him and turn him into a demon. Ignore hiseconomics, and demonize his politics.

The discipline called political economy ended with Marx, but the socialsciences didn’t. Political economy spawned a large family. Political economyitself didn’t hit a dead end. It underwent a name change to Marxism and con-tinued as an integrated theory of the economic and the social. Marxism wasn’tonly a theory. Attempts at the practical application of Marxism changed theworld in ways very few social theories have ever done.

Leaving Marxism aside, the social science disciplines of economics, politicalscience and sociology all emerged in the West from the splintering of politicaleconomy. Orthodox capitalist economics left Marx out of its pantheon and wentback to Smith as the founder of the discipline, taking on the economy part of


202

political economy. Political science picked up on the political part, and sociol-ogy drew on both parts and added some.

Sociology, through Durkheim, has a direct link back to Adam Smith.Durkheim’s canonical work The Division of Labour in Society (1893) picks up on,theorizes and makes technical the term the division of labour, which is discussedat length in Smith but not incorporated into his technical/theoretical frame-work. Durkheim’s mechanical solidarity also has echoes of notions of a state ofnature and thus can be linked back to Hobbes, Smith, Ricardo and Marx as wellas to philosophers like Locke.

The social sciences had a lineage that was an emerging hierarchical know-ledge structure. Following the rupture of political economy, did the varioussocial science disciplines simply become smaller hierarchical knowledge struc-tures or did they, as Bernstein argues, become horizontal knowledge structures?My answer is that both options were possible and that, depending on circum-stances, both paths were taken to varying degrees.

Despite these different paths, in their underlying grammatical and discoursesemantic patterns the various disciplines of the social sciences do represent thescience of the social. Wignell (1997) shows that the social science disciplines ofeconomics, sociology and political science use the same set of linguisticresources to identify and define technical terms, arrange those terms into taxo-nomic frameworks, use those terms to explain phenomena, and from therebuild further technicality. These patterns are both similar to and different fromthe physical sciences and can be traced back to being derived from a hybrid ofthe language of the physical sciences and the humanities. As it evolved, socialscience became a discourse which is both technical and abstract. However, thebalance between the abstract and the technical differs across the social sciences.That is, the balance between the influence of the language of the physicalsciences and the language of the humanities alters.

Of the three social sciences analysed by Wignell (1997), the language oforthodox Western economics was the most technical and the most like thelanguage of the physical sciences. The language of sociology was also highlytechnical but relied less on quantification than economics. Political science,although still technical, was much less so than either economics or sociology.

Since the language of the social sciences evolved as a hybrid of the languageof the physical sciences and the language of the humanities, there is always akind of dynamic tension between the science and the social in the discourse.What happened after the fragmentation of the discipline of political economyis complex and, I speculate again, stems from the dynamic tension between thescience and the social and from the balance of power between the science andthe social among the practitioners of each of the social sciences.

Following from Wignell’s (1997) discussion, orthodox Western economics isclearly an example of a hierarchical knowledge structure in the social sciences.It has continued to evolve from the seeds planted by Smith and Ricardo. Thediscipline of economics has an elaborate technical framework which appears tosatisfy Bernstein’s criteria for a hierarchical knowledge structure.

The case of a discipline such as sociology is less clear. Wignell (1997)examined sociological orthodoxy as laid out in undergraduate textbooks. These


textbooks showed the discipline to be an example of a hierarchical knowledgestructure according to Bernstein’s criteria. However, a textbook presentingorthodoxy to undergraduates cannot afford to spend a lot of time on dissentingvoices. I believe that the key to whether a social science discipline evolves as ahierarchical knowledge structure, a horizontal knowledge structure, or someblend of the two, depends on what happens with dissenting voices.

I hypothesize that the stronger the boundaries around a discipline and thestronger the concord within that discipline, then the easier it will be for that dis-cipline to prevent outside voices from intruding and for it to expel or re-educateinternal dissenting voices. These circumstances would foster the evolution of ahierarchical knowledge structure within that discipline since less time and effortwould be used up in dealing with internal bickering and attacks from outside.With the risks of revolution and invasion reduced, and with relative concordamong insiders, there is more energy available for hierarchical knowledge-building. This would be the case with a discipline like orthodox Westerneconomics.

Conversely, the weaker the boundaries around a discipline (the more itoverlaps with other disciplines), the more it listens to other voices, and themore dissenting voices there are within that discipline, then the more likely itis that it will evolve as either a horizontal knowledge structure or as a blend ofhierarchical and horizontal. I would place sociology as a blend of hierarchicaland horizontal. In its orthodoxy it is a hierarchical knowledge structure, but itis a discipline with both internal dissent and open to many influences. Forexample, as Bernstein (1999: 162) points out, there are Marxist sociologists,post-structuralist sociologists, and a variety of other sociologists. The morecontending voices there are, then the more disagreement there is likely to beeven on fundamentals. In such cases it is likely that one strategy to combatorthodoxy will be each faction inventing a new language. This would result inthe knowledge structure of the discipline over time becoming less hierarchicaland more horizontal.

References

Bernstein, B. (1999), ‘Vertical and Horizontal Discourse: an essay’, in British Journal ofSociology of Education, Vol. 20, No. 2. London: Taylor and Francis.

Durkheim, E. (1893), The Division of Labour in Society (De la division du travail social) (1964reprinting). New York: The Free Press.

Halliday, M.A.K. (1988), ‘On the language of Physical Science’, in Halliday, M.A.K. andMartin, J.R. (eds) (1993), Writing Science: literacy and discursive power. London: Falmer.

Halliday, M.A.K. (1987), ‘Language and the Order of Nature’, in Halliday, M.A.K. andMartin, J.R. (eds) (1993), Writing Science: language and discursive power. London:Falmer.

Hobbes, T. (1651), Leviathan (1986 reprinting). Harmondsworth: Penguin.Locke, J. (1690), ‘On Property’, in Macpherson, C.B. (ed.) (1980), John Locke: Second

Treatise of Government. Indianapolis: Hackett.Martin, J. R. (1997), ‘Analysing Genre: functional parameters’, to appear in Christie, F.

and Martin, J.R. (eds), Genres and Institutions: social processes in the workplace and school,pp. 3–39. London: Cassell.


204

Marx, K. (1867), Capital, Volume One (1974 reprinting). Moscow: Progress Publishers.Ricardo, D. (1817), Principles of Political Economy and Taxation (1973 reprinting). London:

Dent.Smith, A. (1776), The Wealth of Nations (1964 reprinting). London: Dent.Wignell, P. (1998), ‘Technicality and Abstraction in Social Science’, in Martin, J.R. and

Veel, R. (eds), Reading Science. London: Routledge.Wignell, P. (1997), ‘Making the Abstract Technical: on the evolution of the discourse of

social science’. PhD thesis, Department of Linguistics, University of Sydney.Wignell, P. (1992), ‘Abstraction and technicality in Social Science’. Paper presented to

the International Systemic Functional Conference, July 1992, Macquarie University,Sydney.

Wignell, P. (1990), ‘Flowers in the Sentence: a look at the discourse of the sociology ofmedicine’. MA (Applied Linguistics) Special Topic. Department of English andLinguistics, Macquarie University, Sydney.


Introduction

Bernstein (2000: 155) calls for a shift of focus in educational research beyondthe pedagogic transmission and acquisition process (e.g. Bernstein 1990;Christie 1999, 2002; Christie and Martin 1997) to ‘the forms of the discourse’being transmitted; ‘that is the internal principles of their construction and theirsocial base’. The nature of subject-specific discourses has been investigated fromthe systemic functional (SF) perspective for some time, however (e.g. Coffin1997; Halliday and Martin 1993; Martin and Veel 1998; Unsworth 2000; Veel1999), and increasingly, this approach is adopting a multimodal discourseanalysis (SF-MDA) perspective which extends the scope of study beyondlanguage use to consider the contributions of other semiotic resources (e.g.Guo 2004; Iedema 2003; Kress 2003; Kress et al. 2001; Lemke 1998, 2003; O’Hal-loran 1999, 2003a, 2005a; Thibault 2001; Unsworth 2001). As part of thisemerging tradition, this chapter explores how an SF-MDA approach to mathe-matics and science furthers our understanding of Bernstein’s (2000) formu-lations of knowledge, knowledge structure and grammaticality. In particular,Bernstein’s view of the grammaticality of mathematics and science (see Muller,this volume) is enhanced through the concept of multimodal grammaticality,where linguistic, visual and symbolic grammatical systems are seen to functionintegratively to construct mathematical and scientific knowledge. Lastly, peda-gogical implications of the SF-MDA multimodal grammatical approach arebriefly outlined.

Bernstein’s (2000) formulations provide a basis for exploring and comparingmathematical and scientific knowledge. Mathematics is seen to provide the toolsfor the control and manipulation of the material world through the scientificdescription of the relations of time, space and matter for the purposes ofprediction and the establishment of causality. Habermas (1974: 8) explains,‘[t]here is a systematic relationship between the logical structure of a science[i.e. mathematics] and the pragmatic structure of the possible applications ofthe information generated within its framework [i.e. science]’. The ensuingdiscussion reveals how mathematics and science form different knowledge structureswhich complement each other. Through this collaboration, the scientific project issuccessful in the material realm through technological innovation within ever-increasing terms of reference (e.g. physics, chemistry, biology, nuclear physicsand genetics), and increasingly the approach shapes our understanding of whatit means to be human (e.g. through the social sciences, psychiatry, medicineand so forth). An SF-MDA multimodal grammatical approach to understanding

10 Mathematical and scientific forms of knowledge:a systemic functional multimodal grammaticalapproach

Kay O’Halloran

206

mathematics and science as forms of knowledge enhances Bernstein’s (2000)theorizations of knowledge and knowledge structures.

Bernstein’s forms of knowledge and knowledge structures

Bernstein (2000) proposes horizontal discourse and vertical discourse as a firststep in conceptualizing forms of knowledge. Horizontal discourse exists withinthe realm of the everyday knowledge and ‘it is likely to be oral, local, contextdependent and specific, tacit, multi-layered and contradictory across but notwithin contexts’ (Bernstein 2000: 157). Bernstein (2000: 157) explains, ‘thecrucial feature is that it is segmentally organized’ across different situationalcontexts. On the other hand, vertical discourse takes ‘the form of a coherent,explicit and systematically principled structure, hierarchically organized as inthe sciences, or it takes the form of a series of specialized languages . . . in thesocial sciences and humanities’ (Bernstein 2000: 157). Bernstein explains thatvertical discourse is distributed through processes of recontextualization, whichmay be contrasted with the discrete segmented forms of horizontal discourse.Following Bernstein (2000), mathematics and science are viewed as verticaldiscourses which function through the principle of recontextualization.

Bernstein (2000) describes vertical discourse as consisting of hierarchical andhorizontal knowledge structures. In the first instance, hierarchical knowledgestructures are conceptualized as integrating discourses of increasing abstrac-tion, such as those found in physics and biology. In the second instance, horiz-ontal knowledge structures ‘consist of a series of specialized languages withspecialized modes of interrogation and criteria for the construction and circu-lation of texts’ (Bernstein 2000: 161). Horizontal knowledge structures aredistinct, and the result is an accumulation of non-translatable languages. ‘[T]hespeakers of each language become as specialized and as excluding as thelanguage’ (Bernstein 2000: 162). Bernstein considers mathematics to containhorizontal knowledge structures ‘as it consists of a set of discrete languages forparticular problems’ (Bernstein 2000: 163). These formulations raise someissues, which are explored below.

The classification of hierarchical and horizontal knowledge structures inscience and mathematics respectively may oversimplify matters, where thereality is closer to a hybrid mix of the two. Muller (this volume) explains thathierarchical knowledge develops through integration ‘towards a unitary con-vergent shape’ which Bernstein (2000: 161) conceptualizes as a triangle (seeWignell, this volume), while progress in horizontal knowledge structures occursthrough the introduction of new languages by new speakers to investigate newsets of problems. However, as Muller (this volume) points out, while mathemat-ics consists of horizontal knowledge structures with discrete fields of study (e.g.fields of study within pure mathematics, applied mathematics and statistics),forms of verticality associated with hierarchical knowledge structures existwithin these different fields. From this perspective, mathematics would looksomething like a series of triangles which, following Wignell (this volume), havethe potential to be in conflict with each other.


The same may be said of science, where specialization (e.g. nuclear physics)involves the hierarchical knowledge structure of physics and the horizontalknowledge structure specific to the field, in this case, quantum mechanics. Theresult would be triangles within triangles which together form an integratedhierarchy of knowledge, as shown in the diagram. The same may be said ofmathematics, which integrates smaller hierarchical knowledge structures withinthe larger horizontal knowledge structures of mathematics.

One major issue which remains unexplained is the reasons why mathematicsexpands through the proliferation of new languages which become separatefields of study (i.e. horizontal knowledge structures), while science progressesthrough the integration of different fields of study within a unified structure, i.e.a hierarchical knowledge structure (see Muller, this volume, on science andWignell, this volume, on the social sciences). This question is explored inrelation to Bernstein’s formulation of grammaticality. Before moving into thisdiscussion, however, some preliminary remarks regarding the ideological effectsof simple classifications of knowledge and knowledge structures are made (seeMartin, this volume, on taxonomies).

Dichotomies such as vertical/horizontal knowledge and hierarchical/hori-zontal knowledge structures have the potential to simplify the discussion, withone consequence being the maintenance of particular ideologies about thenature of knowledge (Muller 2004). The tendency for simplification throughdichotomization is one effect of using linguistic categories to describe what inessence are complex phenomena. For example, ‘to know’ and ‘to structure’become ‘knowledge structures’ which are differentiated into two types, hier-archical and horizontal. The semantic shift arising from grammaticalmetaphors such as ‘knowledge structures’ (e.g. Halliday 2004a; Halliday andMartin 1993; Halliday and Matthiessen 1999; Martin, this volume; Simon-Van-denbergen, Taverniers and Ravelli 2003) means that the discussion centresaround the description of metaphorical entities, which gain a sense of perma-nence with attached values. While Bernstein’s (2000) theorizations are usefulfrom a global perspective, for example, the concepts of vertical and horizon-tal discourse are usefully employed for a variety of purposes (e.g. this volume),including the investigation of school discourse (Bourne 2003; Muller 2000), itappears some caution is required with respect to the ideological side-effects ofsuch classifications.

The ideological effects of dichotomization become evident with ‘resemioti-cization’, which involves a shift in semiotic construal (Iedema 2001, 2003); inthis case, the visualization of horizontal and vertical discourses. Two conceivablevisualizations are the well-organized city landscape with flat planes for the

207MATHEMATICAL & SCIENTIFIC FORMS OF KNOWLEDGE • O’HALLORAN

Mathematics Science (e.g. physics)

208

segmentally organized horizontal discourse and tower blocks for vertical dis-course, as seen in Figure 10.1a, a three-dimensional Bollmann map of New York,and Figure 10.1b, a virtual city landscape. In these two city scenes, everydayhorizontal discourses take place in the streets while vertical discourses takeplace in the institutional tower blocks. Alternatively, the spatial metaphors hor-izontal and vertical are visualized as lines in Figure 10.1c. The visual mappingsin Figures 10.1a–c (and the triangles pictured above) each promote a well-ordered, stable view of knowledge with different values accorded to horizontal


Figure 10.1a Bollmann map of New York (Nr. 40 – New York 1962)

Vertical discourse(recontextual)

Horizontal discourse(contextual)

Figure 10.1b Stable city landscape

and vertical discourses if height is considered culturally representative of power,status and seclusion.

While keeping in mind the possible ideological effects of dichotomization,Bernstein’s (2000) formulation of grammaticality provides a platform forexploring the reasons why mathematics expands through the proliferation ofnew languages (i.e. horizontal knowledge structures) and science progressesthrough the integration of new knowledge structures (i.e. hierarchical know-ledge structures). From this point, multimodal grammaticality provides an alter-native view of the grammaticality of mathematics and science to provide anaccount of how mathematics and science differ from other forms of knowledge.

Bernstein’s grammaticality and knowledge structures

Bernstein (2000) formulates grammaticality as a form of knowledge variationwhich applies to the internal and external aspects of language description.Bernstein (2000: 163) suggests there are ‘strong/weak’ grammars within thedifferent languages of horizontal knowledge structures, where the distinction isbased on the relative ability of the grammar to model empirical relations in theworld. That is, a strong grammar has the capacity to describe and modelexternal relations in the world (e.g. economics, linguistics and parts of psychol-ogy), while the power of a weak grammar to model such relations is much lower(e.g. sociology, social anthropology and cultural studies). The strong grammarsof horizontal knowledge structures ‘often gain their power by rigorous restric-tions on the empirical phenomena which they address’ (Bernstein 2000: 163).On the other hand, the strong grammar of hierarchical knowledge structures isa feature which remains constant (e.g. physics).

Bernstein (2000) admits, however, that his formulations of the relationsbetween grammaticality and knowledge structures remain unclear. For instance,mathematics creates a problem because it is seen to possess the strongestgrammar of all horizontal knowledge structures, yet it does not have the empir-ical fit with the external world which is the definition of this form of grammar.


Figure 10.1c Horizontal and vertical lines

Vertical discourse(recontextual)

Horizontal discourse(contextual)

210

This contradiction warrants further investigation, for it helps us to explore thereasons why knowledge develops in different forms in mathematics and science.That is, mathematical knowledge expands through proliferation (i.e. horizontalknowledge structures) while science progresses through integration of newknowledge structures (i.e. hierarchical knowledge structures).

Bernstein’s view of strong/weak grammars as a means for investigating and dif-ferentiating between mathematical and scientific knowledge is undertaken in twosteps. First, mathematics and science are discussed with respect to their relation-ship to each other, and their relative ability to model empirical relations in theexternal world. Second, comparisons between mathematics and science provide akey to understanding the relationship between knowledge structures and the waysin which the two disciplines expand and progress through the principles of pro-liferation and integration respectively. From here, the nature of mathematical andscience knowledge is developed from the perspective of SF approach to multi-modal grammaticality in the next section: ‘SF Approach to Multimodal Gram-maticality’. This discussion is aimed at exploring the relationship betweenmathematics and science, and how they differ from other forms of knowledge.

The relationship between the disciplines of mathematics and science isintimate. Bell (1987), for example, sees mathematics as the ‘queen and servantof science’. While different in fundamental ways, mathematics is none the lesscalled ‘the queen of the sciences’ and is ranked first amongst the sciences:

It is customary to call mathematics a science and, after the positivist philosopherI.A.M.F.X. Comte (1798–1857), to place it first in the classification of the sciences.So as long as we remember the radical difference between mathematics and thephysical or biologic sciences, no harm is done in calling mathematics a science.Something of the distinction of the mathematical method and the strictly scien-tific, however, must be seen before we attempt to uncover the mysteries of math-ematical prophecy in scientific discovery. The matter is extremely simple, butnone the less profound. (Bell 1987: 259)

The ‘radical difference’ between mathematics and science lies in the relation-ship with the external world. ‘A science has “real” content [with respect to theworld], or claims to have’ (Bell 1987: 259), and it proceeds through experi-mentation and correlation with empirical data. On the other hand, ‘[t]he essenceof mathematics is deductive reasoning from explicitly stated assumptions called postulates’(Bell 1987, p. 261), which may have nothing to do with the external world.

Electromagnetism, for example, is the organized body of information that hasbeen acquired concerning electricity and magnetism as they ‘actually’ appear inhuman experience. Likewise astronomy, unless scrutinized too critically, is asystematized accumulation of ‘facts’ about heavenly bodies, these celestial objectsbeing assumed to have an ‘existence’ outside our ‘sense perceptions’. Withmathematics, it is entirely different . . . So far as the external world is concerned,mathematics is as empty as a game imagined in a dream and forgotten on waking.Nothing whatever is in mathematics except the rules of the game, and these rulesare prescribed at will by the player. (Bell 1987: 259)


Mathematics is seen to be the ‘servant’ or ‘handmaiden’ of science because itis called upon to provide the means for scientific descriptions of empirical rela-tions in the external world. Furthermore, the scientific experimental methoddepends on scientific laws which are mathematically expressed. Emeritus Pro-fessor Philip Davis (Department of Applied Mathematics, Brown University)explains how mathematics is applied to areas beyond the immediate concernsof mathematics itself:

Mathematics is the science, art and a language of quantity, space, and pattern. Itsmaterials are organized into logically deductive and very often computationalstructures. Its ideas are abstracted, generalized, and applied to concerns otherthan mathematics itself.

When mathematics mixes with ‘outside concerns’, the mixture is called appliedmathematics, and for reasons that are by no means clear and may even remain aperpetual mystery, mathematics has been found to be of utility and an indispen-sable aid to the physical sciences. It has often been called the ‘Handmaiden of theSciences’. The expositions of theoretical physics are completely mathematical incharacter. (Davis, personal correspondence, 17 March 2005)

There are several points to be made here. First, mathematics and science sharethe same grammatical basis, a point which is developed in the next section: ‘SFApproach to Multimodal Grammaticality’. Second, mathematics has weakempirical relations with the world only within certain fields of mathematics (e.g.‘pure mathematics’). Fairly obviously, applied mathematics is concerned withmodelling real, empirical data. If we put aside applied mathematics, and acceptpure mathematics as the benchmark of mathematical activity, mathematicsappears to be at odds with Bernstein’s view of strong/weak grammars because ithas the strongest grammar of horizontal knowledge structures, yet it is not con-cerned with empirical relations within the world. On the other hand, sciencehas the strongest grammar of hierarchical knowledge structures and thestrongest empirical relations with regard to the external world. Why is this thecase? Why the discrepancy between the strong grammaticalities of mathematicsand science, and their respective relations to the external world? And how doesthis discrepancy relate to the expansion of knowledge in mathematics and theprogression of knowledge in science?

The answer to the question concerning the strong grammaticality of math-ematics and science becomes self evident; mathematics and science both havestrong grammars because they both have a grammar based on mathematicswhich has the potential to describe empirical relations in the world (i.e. the caseof applied mathematics).

Furthermore, the methodologies of mathematics and science both involvereasoning based on mathematical postulates and scientific laws which are math-ematically expressed. For reasons which philosophers continue to debate (seeWigner 1960), mathematics is effective in describing empirical relations in theworld. In conclusion, the grammar looks inwards (internally) in the case ofmathematics, and the grammar looks outwards (externally) in the case ofscience (and applied mathematics). The strong grammaticality of mathematics


212

and science is at one with each other: the difference is a matter of application.One could reasonably ask, why is this the case? How does the close partnershipbetween mathematics and science work, and what are the mutual benefits? Andwhy does mathematics expand through proliferation of new horizontal knowl-edge structures, and science through the integration of new knowledge struc-tures to form a unified hierarchical knowledge structure?

The question of expansion and progression through proliferation and inte-gration is not quite so clear-cut. Muller (this volume) sees progress in science asthe product of the integrative nature of hierarchical knowledge structure ofscience which combines with a strong grammar to give science a strongerexplanatory power than other forms of knowledge. One speaks of progress, andone thinks of science.

In other words, grammaticality determines the capacity of a theory or language toprogress through worldly collaboration; verticality determines the capacity of atheory or language to progress integratively through explanatory sophistication.Together, we may say that these two criteria determine the capacity of a particularknowledge structure to progress. [my emphasis]

Muller’s (this volume) link between verticality and grammaticality is a signif-icant step in explaining the progress of science in terms of an increasingexplanatory power. The question which remains is the discrepancy betweenexpansion through proliferation in mathematics and progress through integra-tion in science. Given the strong relationship between mathematics and science,why do the two disciplines differ in this respect? The answer to this question hasseveral dimensions which are developed below.

First, one possible reason for the integrative nature of scientific knowledge isthat by seeking to explain the external world, science possesses a strongly unifiedagenda. Science seeks to explain external phenomena, and thus scientificendeavours are orientated towards achieving that common goal, in one way oranother. Second, science depends upon support from outside for scientificknowledge to progress. That is, the mapping of experience (and hence knowl-edge) is tied to political, economic and military interests (see Wignell, thisvolume) where associated bodies fund and support the development of science.Scientific progression, in particular, is linked to the accumulation of power andcapital through technological advancement. One could almost say that scienceserves that goal. From this perspective, the progression of scientific knowledgeis determined by factors outside the intellectual arena. Scientific knowledgeprogresses according to the interests of stakeholders, and most typically newknowledge is integrated within existing structures because if isolated outsideexisting fields of research, the new knowledge will disappear unless institu-tional, financial and intellectual support is obtained to develop that knowledge.If such support is obtained, the new knowledge becomes integrated within theexisting structures of the science which seeks to fulfill the same goal. The otheralternative is a major rupture (e.g. see Koestler, 1959; Kuhn, 1970), which shiftsthe axis of verticality of the scientific knowledge structure from one position toanother. But this rupture doesn’t bring down the whole hierarchical knowledge


structure of a science, because – and this is the third major point – science pos-sesses a strong mathematical infrastructure which can withstand alternativedescriptions of phenomena. That is, a range of mathematical tools can be inte-grated for the common pursuit of describing external phenomena. This bringsus to a fourth (inter-related) point. If a knowledge structure is concerned withspecific areas of experience – i.e. if it has strong relations with the externalworld – for progress to take place, the new knowledge structure has to offerbenefits above and beyond other competing approaches which explain thatsame phenomenon. Knowledge progression becomes a political/economic/military/intellectual exercise where alternatives co-exist in the same hierarchi-cal knowledge structure if they provide rewards to those maintaining thestructure. The perceived value of a mathematical and scientific theory relates tothe objectives which it may achieve. Muller’s (this volume) view of the ‘worldlycollaboration’ of science takes on a new meaning in this light.

Mathematics, on the other hand, does not compete with regard to explainingthe phenomena in the real world. Unless called into action to support a causein the external world, mathematics is generally focused inwards (with the excep-tion of applied mathematics, as explained above). New knowledge can, up to acertain point, proliferate unproblematically. Following Wignell (this volume), ifconflict does occur within a horizontal mathematical knowledge structure, thehierarchy is maintained through internal struggle, or alternatively, another hor-izontal discourse is established. Wignell (this volume) suggests that the outcomeof struggle depends on the nature of the boundaries which are established;strong boundaries provide protection for practitioners within horizontal know-ledge structures. Given that mathematics has strong boundaries, dissentingvoices in mathematics are either quietened or leave to form new horizontalknowledge structures. This is not to say that mathematics is free to expandwith unbounded proliferation; there are limits on what type of mathematics isdeveloped at different periods of time. Some constraints relate to the closerelationship of mathematics to science, which selectively calls upon math-ematics to solve particular problems. The ability of mathematics to expandthrough proliferation is, however, one key factor which explains the successfulprogression of science. This point is developed below.

The significant point arising from this discussion is that science progressesthrough accessing the strong grammar of mathematics which is free to proliferate to providealternative approaches and descriptive tools. These approaches are integrated withthe hierarchical knowledge structure of science to serve particular agendasrelating to the description of material reality. The relationship of mathematicsto science is a two-way affair, what Bell (1987: 2–3) calls the ‘right-hand,left-hand aspect of mathematics’:

[T]he pure serves the applied, the applied pays for the service with abundance ofnew problems that may occupy the pure for generations. The debt may then bereversed, when art for art’s sake [in mathematics] pays off in the solution of diffi-cult problems in science and technology.


214

As O’Halloran (2005) and Wignell (this volume) explain, science is a dis-course which rapidly expanded because so many stood to gain from the powerit unleashed: the bourgeoisie, aristocracy, military, merchants, capitalists,intellectuals and scientists themselves. Being the queen and handmaiden forthe sciences, mathematics underscores scientific progression and the develop-ment of new technology. Bell (1987), for example, surmises that war outrankseconomics with regard to being influential in the development of mathematics,a trend which must have increased in the past few decades. More mathematicswas developed in some departments ‘in less than a decade of [the second world]war than might have been found in half a century of peace’ (Bell 1987: 3). Themathematics was used to develop military technology and armaments.

The close relationship between mathematical and scientific forms of know-ledge provides a platform to understand how horizontal and hierarchicalknowledge structures function productively together as a related phenomenonwhich fuels scientific progression. That is, the verticality and strong grammarsof the sciences which progressively integrate knowledge depend upon a stronggrammar which proliferates in the horizontal knowledge structures of math-ematics. In what follows, the basis for Bernstein’s formulation of the stronggrammars for mathematics and science (which are one) is explored through theconcept of multimodal grammaticality, where ‘the languages’ of mathematicsand science are re-conceptualized as the product of the integration of threesemiotic resources: language, mathematical symbolism and visual imagery. Theaim of the ensuing discussion is to explore how mathematical and scientificknowledge provides the means for rewriting the external world.

SF approach to multimodal grammaticality

Multimodal grammaticality is proposed as an approach which furthers our under-standing of the grammaticality of mathematical and scientific knowledge beyondthat proposed by Bernstein (2000). The approach departs from Bernstein’sformulation of the strong and weak grammar distinction between mathematicsand science with the aim of investigating the underlying similarities between thetwo forms of knowledge which permit the ‘right-hand, left-hand aspect of math-ematics’ described by Bell (1987: 2–3). For this purpose, mathematics andscience are conceptualized as forms of knowledge constructed through the inte-grated use of grammatical choices from language, visual images and symbolism.In what follows, multimodal grammaticality is theorized from the SystemicFunctional (SF) social semiotic perspective of language (Halliday 1978, 2004a;Halliday and Matthiessen 1999; Martin 1992; Martin and Rose 2003), which hasbeen extended to visual images and other semiotic resources (e.g. Kress and vanLeeuwen 1996; Martinec 2000, 2001; O’Toole, 1994; van Leeuwen 1999, 2005).More recently, SF multimodal discourse analysis (SF-MDA) research efforts aredirected towards investigating the ways in which semiotic choices interact tocreate meaning (e.g. Baldry 2000; Baldry and Thibault 2006; Iedema 2003;Lemke 1998, 2002; Martin 2002; Martinec 2004; O’Halloran 2004, 2005; Royceand Bowcher, in press; Thibault 2000; Ventola 2004).


An SF multimodal grammatical approach to mathematics and science isconcerned with the individual and integrated functionality of language, math-ematical symbolism and visual images in the form of diagrams, and abstract andstatistical graphs (see Lemke 1998, 2003; O’Halloran 2005). In the first case, thethree semiotic resources are seen to be uniquely functional in the constructionof mathematical and scientific knowledge. That is, each resource makes an indi-vidual intra-semiotic contribution to meaning through the meaning potential ofthat resource. Second, the three resources function inter-semiotically to makemeaning. That is, mathematical and scientific discourse functions through thegrammatical integration of linguistic, symbolic and visual elements. Further-more, shifts from one semiotic resource to another (e.g. linguistic to visual tosymbolic) in larger discourse moves permit metaphorical expansions ofmeaning to take place, while simultaneously enabling the meaning potential ofthe new semiotic resource to be accessed. Using pre-established results, laws andtheorems, the construction of knowledge in mathematics and science takesplace through:

• the meaning potential of the three semiotic resources;• the integration of grammatical choices from the three semiotic resources;• the expansions of meaning which take place as the discourse shifts from

one resource to another.

One aim of multimodal grammaticality is to theorize the three semioticresources as an integrated phenomenon to explore the nature of mathematicsand scientific knowledge. One result is formulation of the ‘strong grammaticalinterconnectivity’ of the three semiotic resources within a restricted semanticfield (see section (2) ‘Strong grammatical interconnectivity: the visual imageand the symbolism’).

The SF-MDA approach to multimodal grammaticality is developed withrespect to the concept of visibility (see Miller 1984; Virilio 1994) where ‘to makevisible’ is to capture semiotically; in this case, using visual imagery, mathemati-cal symbolism and language. Visibility relates to the concept of semiotic reality(Martin 1992) where the discourse shifts from ordering the material realm toordering the semiotic realm. The scientific project is the result of the expandedvisibility made possible through semiosis: that is, semiosis involving visualimagery and symbolism, the two central semiotic resources upon which modernmathematics and science are built. Language is seen to play a significant sup-portive role in this enterprise. However, the functions of language are only dis-cussed in general terms in this paper (see Halliday 2004b; Halliday and Martin1993; Martin and Veel 1998) because the focus is directed towards the visualimagery and the symbolism, the distinguishing trademarks of mathematical andscientific knowledge. In what follows, the expanded visibility derived from theuse of mathematical symbolism and visual imagery is explored.


216

The mathematical and scientific project

The mathematical and scientific project is unashamedly aimed at the control andmanipulation of the material realm. In the beginnings of modern science,Descartes (1596–1650) sought to establish a ‘methodology for a universal sciencewhich is as concerned with increasing human power as well as understanding’(Barry, 1996: 35). Barry (1996: 35) explains that ‘Descartes’ programme is tech-nological, because it seeks to penetrate the secrets of the material realm in orderto manipulate them’. For these ends, Descartes dismisses that which is perceivedthrough senses and accepts only that which is conceivable through the mind (seeDescartes 1952). This decision is based on Descartes’ view of the unreliability ofsense data and on his success in using mathematical symbolism to constructgeometrical curves. Descartes has the mathematical symbolism as a tool for rea-soning and the geometrical images as a tool for seeing in what becomes a limitedand well-defined semantic field (see Descartes, 1683, 1954). Descartes moves toconstructing a new realm of semiotic reality, and in doing so, the everydaylanguage of the sensory world is dismissed and the mathematical symbolism andgeometrical imagery is developed. The result is a mind/body duality with a priv-ileging of two forms of semiosis. That is, the mathematicized world constructedusing geometrical images and mathematical symbolism is admitted as true know-ledge, while sensory data constructed through everyday language is dismissed assecondary knowledge (see Descartes 1985 and Muller, this volume).

Newton (1642–1727) develops Descartes’ symbolic and visual semiotic toolsto reformulate the material sensory realm in what becomes an empirical-basedscience which uses scientific instrumentation in experimental laboratories.Newton constructs his mathematical principles of the natural world based on amodel of linearity, where complex wholes are broken into constituent parts inorder to understand the system. Although the model of linearity has beenreplaced by non-linear dynamical systems theory, Newton nonetheless makesvisible the non-visible through the symbolic and visual formulation of entitiessuch as force, gravity, velocity and acceleration (Barry 1996), a scientific tradi-tion of visualization which continues today within computerized environments.In what follows, visibility is discussed in relation to the functions and grammarof mathematical and scientific visual imagery. From this point, the strong gram-matical interconnectivity between the visual imagery and the symbolism isexplored, before the grammar of mathematical symbolism is discussed.

(1) Mathematical and scientific visual imageryThe impact of visual imagery is so significant that scientific, political, religiousand artistic pictorial images, icons and models are sites of construction, destruc-tion and defacement (Latour and Weibel 2002). However, the major theoreticalapproach to visual imagery is often restricted to the cognitive domain for inves-tigating cognitive psychological processing of visual images, the physiology ofperception, and the dynamics of creative thinking. This emphasis on thecognitive may be explained by the available funding for research in the ‘scienceand technology of instrumented aiding of optically-guided spatial behaviour’(Koenderink 1993: viii) for military purposes, space exploration, medicaldiagnostics and other fields where the human operator is vital but often


performs as ‘the weakest in the chain’ (Koenderink 1993: ix). Koenderink(1993: ix) explains that ‘[a] fundamental understanding of human abilities andidiosyncrasies is necessary, but only fragments exist’. A second major approachto visual imagery is the graphic design perspective for developing effectivedisplays of information for advertising, signs, maps and so forth (e.g. Lester2000; Wildbur 1989; Wildbur and Burke 1998). The intended use of the visualimage thus plays an important role in determining the theoretical approach tovisual semiosis. The social semiotic approach to the semantic realm which math-ematics and science make visible remains a relatively unexplored field of study.This approach to visual imagery in mathematics and science is developed below.

God plays a central role in the beginnings of modern science (Funkenstein1986). The need for God, however, disappears with the ongoing success ofscience in formulating time, space, matter relations for purposes of prediction andestablishing causality. With an increased focus on geometrical shapes, lines andcurves, human figures in material settings gradually disappear from mathemat-ical and scientific drawings (O’Halloran 2003b, 2005a). Human bodies arereplaced by body parts, such as hands to depict material actions and eyes for actsof perception. Scientific procedures, for example, involve disembodied handsas displayed in Figure 10.2. Military objectives such as hitting targets are nolonger represented. Eventually the context of the problem disappearsaltogether and only the geometrical lines and curves remain. The ongoingsuccess of science meant that God became irrelevant and man became theinvisible agent in the programme for controlling the material world. Todaymathematical results rarely directly display the objectives for which they weredeveloped, and the reader sees non-violent applications of the theory after themathematical results are established in a de-contextualized environment. Anexamination of the beginnings of modern science is productive, however, forunderstanding the semantics and grammar of modern mathematical and scien-tific visual images. In what follows, Newton’s early work is closely examined.


Figure 10.2 Scientific experiment with frogs (Plate 1, Galvani 1792)reproduced in Stafford (1996: 91)

218

Figure 10.3 is the page opposite the title page of Newton’s (1736) The Methodof Fluxions and Infinite Series. The problem of men shooting flying birds is visual-ized, complete with hunting dogs to collect the victims. The visual image isinteresting for several reasons. First, it becomes possible to see how mathemat-ical and scientific images make visible the invisible. For example, the paths ofthe two birds are semiotically constructed over time using parallel lines AB andCD. Instances of time in the flight path of two birds are displayed spatially aspoints E, F and G and H, I and K. The conflation of matter (in this case, thebirds) with respect to particular instances of time in the depicted scene aremapped as points G and K. Three lines of fire are captured visually as dottedlines. The potential collision of the gunshot with the birds is depicted as pointsE, F and G and H, I and K, with the intended hit pictured at G and K. Thus thevisual image makes visible the flight path, the lines of fire, and the potentialinteraction between material matter (the gunshot and the birds) at differentinstance of time. That is, the drawing is concerned with the relations betweenspace, time and matter for the purposes of prediction.


Figure 10.3 Space, time, matter and prediction (Newton 1736: Frontispiece)

Second, Figure 10.3 demonstrates how spatiality is linked with temporality inways which are unique to mathematics and science. The experience of visualspace is explained by Saint-Martin (1990: xi–xii): ‘. . . spatiality can be definedas the apprehension of a simultaneous coexistence of multiple elements in anautonomous form of organization, which is considerably different from that ofthe temporal order of these elements’. From Newton’s (1736) drawing,however, it can be seen that mathematical and scientific diagrams capturespatiality with respect to a continuous temporal order, which simultaneouslyincorporates several specific instances of time. We find the conflation of spacewith time in a way that combines matter with predictive ability, a feat which doesnot seem to occur in other genres of visualization. This point warrants furtherdiscussion.

A photograph captures a single frozen instance of time; an instance wheresomeone may be crying, laughing or running. The sculptor Auguste Rodinexplains that sculpture and other art forms, however, have the potential to givethe impression of movement over time, rather than capturing the single instanceof the photograph. Rodin claims that in doing so, ‘[i]t is art that tells the truthand photography that lies. For in reality time does not stand still, and if the artistmanages to give the impression that a gesture is being executed over severalseconds, their work is certainly much less conventional than the scientific imagein which time is abruptly suspended . . . ’ [excerpt from Paul Gsell interview withRodin cited in Virilio (1994: 2)]. Rodin appears to be simultaneously right andwrong. Time is abruptly suspended in the mathematical and scientific image,but it may be suspended for any number of instances within a continuous timeframe.That is, the mathematical and scientific image contains repeated photographicsnapshots, where the potential exists for any moment to be seized indefinitely,as seen in Newton’s drawing in Figure 10.3. The peculiar capturing of unlimitedphotographic-type instances within a continuous framework in mathematicaland scientific visual imagery is explored in relation to Figure 10.3.

There is a sense of movement and action in Newton’s drawing, with the menshooting the birds and the dogs sniffing and watching. However, the math-ematical lines and points capture a time-sequence which includes and extends theaction sequence of the men and the dogs. That is, the time-frame realized by theflight paths of the birds includes a time before and a time after the huntingscene which is drawn. The hunting scene serves as a visual backdrop to context-ualize the spatial relations of matter over time for the purpose of prediction; i.e.to hit the bird at times G and K, or more generally, to hit a moving target atsome pre-determined time. However, the mathematical semiotic construction ofthe problem, the geometric lines and points, captures change over time in amanner which extends beyond other genres of visual representation such asdrawings, photographs and sculptures. The lines and curves in mathematicalgraphs display an uninterrupted continuity in terms of spatio-temporal rela-tions, where the temporal-frame of mathematical visual images extends to theinfinite. The mathematical visual curve typically depicts a relationship betweenentities; it makes the relationship visible, it makes the relationship concrete and itmakes the relationship continuous. From this point, the relationship is describedusing mathematical symbolism.


220

If we consider Newton’s (1736: 67) mathematical writings in The Method ofFluxions and Infinite Series displayed in Figure 10.4, relations of time, space andmatter appear as geometrical figures in the form of triangles, curves, lines andpoints, and numbered paragraphs consisting of linguistic and symbolic com-ponents. The de-contextualized nature of the visual images means that theexperiential and logical meanings are foregrounded. There are no visual dis-tractions in mathematical and scientific visual imagery. From this point, havingachieved visibility in the form of geometrical images, the strong grammaticalrelations between these images and their symbolic description (e.g. thenumbered paragraphs in Figure 10.4) provide one path for understanding thefunctionality of the mathematical symbolism.


Figure 10.4 Revisiting the moment mathematically (Newton 1736: 67)

(2) Strong grammatical interconnectivity: the visual image and the symbolismDescartes’ method of using mathematical symbolism to describe curves meansthat the space, time and matter relations displayed in the visual image arerelocated within a new semantic map; that is, the visual patterns of co-variationare exactly described symbolically, a feat which is impossible using language(Lemke 1998, 2003). The meaning potential of language simply does notpermit curves to be described in such an exact manner. Mathematical symbol-ism evolved from language, however, to develop new grammatical systems ofmeaning so that symbolic description and manipulation of continuous patternsof relations became possible. Before the grammatical strategies developed inmathematical symbolism to achieve these functions are investigated (see section(3) ‘The grammaticality of the mathematical symbolism’), the nature of thegrammatical relations between the visual image and the symbolism is discussed.

A point in the mathematical visual image represents an instance in the stateof the system which captures the spatio-temporal relations between entities (e.g.the birds and the bullet in Figure 10.3). The instance, and every instance, isvisualized in relation to other states of the system (e.g. the line segments for theflight path of the birds and the line of fire), just as the hand for dissecting thefrogs in Figure 10.2 is viewed in relation to the activity sequence of the experi-ment which is portrayed. The ‘temporal’ frame of the mathematical and scien-tific image is not necessarily related to clock-time, rather the framing relates tounfolding patterns of relations. For example, the tonnage of nuclear bombs andthe levels of radioactivity may be visualized as an unfolding system along a scalewhich is not a temporal-clock version of time. Rather, the system is a pattern ofco-varying relations where specific instances may be located and suspended.

Significantly, the semantic framework of the visual image has a strong rela-tionship with the semantics of the symbolism, so the relations of the entitiesvisually displayed can be described exactly using mathematical symbolism, withthe potential for describing the state of the system at specific instances. Thismeans that the visual image is re-contextualized within the semiotic landscape ofthe symbolism, and vice versa. The re-contextualization process is bi-directionallysemiotic; in other words, one form of semiotic construction (the visual image) isre-contextualized with respect to a second form of semiosis (themathematical symbolism), and vice versa. The linguistic text is also involved inthe re-contextualization process (e.g. Figure 10.4) making mathematics andscience multi-directionally semiotic.

There is room for such ‘bi-directional’ and ‘multi-directional’ re-contextual-ization processes within other semiotic landscapes; for example, the photo-graph in the newspaper is re-contextualized with respect to the linguistic text,and vice versa. However, the directedness of the re-contextualization process inmathematics leads to what could be called a strong grammatical interconnectivity.That is, there is little room for ambiguity in the re-contextualization process,and so the ‘semantic effervescence’ (Cheong 2004) arising from the interactionbetween the visual image, the mathematical symbolism and the linguistic text isminimal. In fact, mathematics developed as a written semiotic discourse wherespecific types of grammatical interconnectivity were established in order to solveproblems through the mathematical symbolism. The strategies for achieving the


222

strong grammatical interconnectivity between the visual image and the symbol-ism are discussed below.

Space becomes a metaphor for the temporal in the form of the unfoldingrelations in the mathematical and scientific visual image, and the temporalbecomes located within space, a visual space which is describable using math-ematical symbolism. For example, Newton’s (1736: 10) grid in Figure 10.5demonstrates how an instant * can be spatially suspended in relation to anunfolding sequence of relations, in this case represented by the dotted line.The capturing, suspension and mapping of temporal-spatial relations in two-and three-dimensional symbolic grids may be seen in Figures 10.6a Newton(1968: 31), 10.6b Prigent et al. (2003: 112) and 10.6c Golovin et al. (2003: 199).According to the symbolic organization or mapping of space, (i) a point (i.e.an instant in the relationship between entities), (ii) the line, the curve andthree-dimensional shapes (i.e. the actual relationship between entities oversome temporal frame) and (iii) space (i.e. the potentiality of the relationship)have symbolical formulations. That is, the strong grammatical interconnectiv-ity between the visual image and the symbolism occurs because the visualentities (the point, the curve, shape or object, and space) have symbolicdescriptions. These symbolic descriptions are manipulated to describe otherrelationships and to predict the behaviour of systems. As we see from Figure10.6c, new forms of visualization are appearing through the use of computergraphics (see section ‘Visibility in the computer age’).

Mathematical and scientific visual images are concerned with the spatial-temporal display of relations which are suspended within a symbolic grid. Thevisual entities have symbolic identities (e.g. x, y and z) and the symbolic config-uration of the relationship (the line, the curve and three-dimensional shapes)


Figure 10.5 Suspending the instant (Newton, 1736: 10)

takes place through the formulation of the patterns of variation using math-ematical operative processes (O’Halloran 1999, 2003a, 2005) (see section (3)‘The grammaticality of the mathematical symbolism’). The ongoing success ofthe symbolic descriptions means that visualization, however, is no longer a nec-essary requirement for mathematical and scientific theorizations. For example,geometry has moved beyond three-dimensional Cartesian space to higher non-visible dimensions, and calculations in nuclear physics predict the behaviour ofelectrons which cannot be visualized. That is, the verticality of mathematicaland scientific discourse has moved beyond the boundaries of perceptual space-time-matter reality through the semiotic potential of the symbolism. Being


Figure 10.6a Suspending the visual image: Newton (1968: 31)

224

caught in a visual grid and/or symbolic forms of reasoning, however, means thatan effort is required to escape the confines of the semantic field which is subse-quently established. The increasing potential to manipulate visual imagerythrough computer generated imagery (CGI) is providing semiotic tools to movebeyond traditional approaches to mathematics and science (see section ‘Visibil-ity in the computer age’). The grammaticality of the symbolism is investigatedbelow.


Figure 10.6b Suspending the visual image: Prigent et al. (2003: 112)

(3) The grammaticality of the mathematical symbolismVirilio (1994: 5) explains how war contributed to condensing and making effi-cient written characters: ‘[t]he tendency to make reading time as intensive asspeaking time . . . stemmed from the tactical necessities of military conquest andmore particularly of the battlefield . . .’. However, despite its written efficiency,language is limited for capturing precise degrees of variation in relationships(Lemke 2003), for example between time, space, and matter. While visualimages capture variation in the form of unfolding relations, traditionally visual


Figure 10.6c Suspending the visual image: Golovin et al. (2003: 199)

226

displays are manipulated and re-reassembled for predictive purposes with somedifficulty, although this situation has changed considerably with computer tech-nology (see section ‘Visibility in the computer age’). The semiotic resourcethrough which relationships are encoded and manipulated is the mathematicalsymbolism because, as those following Descartes discovered, the symbolismincorporates exact descriptions of visualized relations through the strong gram-matical interconnectivity described above. Particular forms of grammaticality ofthe mathematical symbolism (which differ from language) were developed sothat it could be used as a tool for reasoning about the visualized relations.However, despite the significance of mathematical symbolism, it remains largelyan untouched area of research in the humanities, as discussed below.

Elkins (1999: 134) explains that symbolic mathematical formulations arelargely under-theorized domains of writing in the humanities: ‘[t]raditionally,mathematical expressions, formulae, diagrams, and equations are excludedfrom publications in the humanities. They are virtually absent from texts thatexplore deconstruction, and it is even possible to write about philosophic logicwithout using formalisms.’ Following Rotman (1995), it is difficult to justify theexclusion of mathematical notation for the study and deconstruction of writtentexts. One reason for the omission is that mathematical symbolic expressionsare typically not accessible to a wide audience. The symbolism requires theo-rization, however, and the current situation highlights the need for transdisci-plinary studies to bridge mathematics, the sciences, the humanities andeducation. A step is made in this direction in this paper.

Mathematical symbolism is a ‘symbolic’ sign system according to Peirce’s(1991: 251) classification of sign systems, where ‘every sign is determined by itsobject’. The Icon possesses characteristics of the object it refers to, the Index hasconnections to the object, but the meaning of the Symbol must be acquiredthrough ‘habit’, i.e. the culturally assigned meanings and values must be learnt.Language and mathematical symbolism are such symbolic sign systems. Math-ematical and scientific visual images typically bear some relationship to percep-tual reality, although this changed with quantum mechanics and, more recently,visualizations made possible through computer graphics. The majority of tradi-tional visual images tend to be iconic (e.g. a three-dimensional cube to representa box) and indexical (e.g. the graph where increase/decrease appear asrising/falling curves). However, despite these relations to perceptual reality, themeanings of mathematical visual imagery need to be taught and learnt, and thedifficulties for students learning mathematics increase with the symbolic status oflanguage and mathematical symbolism. The difficulties associated with scientificlanguage and mathematical symbolism are explored below.

In the first case, scientific language departs from the congruence of everydaylanguage (Halliday 2004b; Halliday and Martin 1993; Halliday and Matthiessen1999; Martin and Veel 1998). For example, Halliday (1993: 69–85) lists thedifficulties of scientific and mathematical language as including interlockingdefinitions, technical taxonomies, special expressions, lexical density, syntacticalambiguity, grammatical metaphor and semantic discontinuity. The metaphoricalnature of scientific language is conceptualized through grammatical metaphorand other linguistic constructions which aid the flow of reasoning (e.g. Halliday


and Martin 1993; Martin and Veel 1998; Simon-Vandenbergen et al., 2003). Inthe second case, it appears that the major problem in teaching and learningmathematics and science relates to mathematical symbolism, the semioticthrough which theory is constructed and problems are solved. Mathematicalsymbolism is most clearly a Peircian symbolic sign which must be acquiredthrough learning, the problem being that most people fail to learn the ‘habit’.The question which plagues many mathematics educationalists is why this is thecase. In what follows I give some possible reasons.

Mathematical symbolism is the major tool for reasoning because it developedas a form of semiosis capable of (i) encoding the continuity of relations dis-played by visual images, (ii) suspending and capturing those relations at anypoint, and (iii) rearranging those relations to solve problems. In other words,the symbolism captures exactly the relations between space, time and matter toestablish states of being for predictive purposes. But how is mathematicalsymbolism organized so that these functions may be fulfilled? This question isexplored below in terms of semantics, grammar and grammatical strategies forencoding meaning in mathematical symbolism.

Mathematical and scientific symbolic descriptions construct a differentversion of reality compared to linguistic constructions. Firstly, the semantic fieldis radically reduced to time, space, and matter relations (i.e. experientialmeaning) for purposes of prediction and establishing causality (i.e. logicalmeaning). With respect to experiential meaning, Relational Identifyingprocesses establish states of being (e.g. =, ≤ and ∝) which are described usingOperative processes (e.g. ×, +, ÷ and ±) between generalized mathematicalparticipants (e.g. x, y and z) (O’Halloran 1999, 2005a). The mathematicalparticipants, in the form of numbers and variables, have direct referents withinthe context of the problem, or otherwise the referent may remain unspecified.The nature of Operative mathematical processes radically departs from theprocess types found in language.

Operative processes are symbolic processes of dividing, adding, subtractingand multiplying and so forth which capture the dynamic patterns of variation asa complex of interactions within and between process/participant configur-ations. Operative processes developed from materially based processes (sharing,increasing and decreasing material goods, for example) to describe degrees ofincrease, decrease and relativity with virtually unlimited boundaries of delicacy.However, Operative processes are semantically different from linguisticprocesses. Operative process/participant configurations, for example, do notnecessarily contain one major participant (the Medium) with the potential foragency (the Agent), which is the hallmark of linguistic constructions (seeHalliday 2004a). The potential exists within the mathematical symbolism toconfigure an unlimited number of Operative processes and participants; forexample, x1 × x2 × x3 . . . xn where x i and n take any value, and × can be replacedwith other forms of Operative processes. Given the narrow semantic field of themathematical symbolic descriptions, the choices for circumstance are limited.The nature of symbolic processes, participants and circumstance requiresfurther research in order to investigate the ways in which mathematics andscience construct the world (see O’Halloran 2005, Chapter 5).


228

Operative process/participant symbolic configurations are maintained sothey can be manipulated and rearranged to solve problems. This may becompared to the encoding of meaning through grammatical metaphor inlanguage where dynamic processes are lost, for example, in nominalized wordgroups. The means for maintaining the relationships between participants isembedding where process/participant structures are rankshifted to functionwithin other process/participant structures. Mathematical symbolism has thepotential to embed process/participant configurations to an unlimited degree.To make the embedding and consequent potential for rearrangement of rela-tions a viable option, an economy of expression is blended with conciseness in thesymbolism. This took place through the development of new grammaticalsystems. That is, modern mathematical symbolic notation developed as a writtensystem with the aim of solving particular types of problems, and thereforespecific grammatical systems for encoding meaning were developed.

The new grammatical systems include, for example, generalized symbolic par-ticipants (x, y and z), special symbols for processes (e.g. Σ, √ and ±, ×), spatialnotion (e.g. x

y indicate division, powers to indicate multiplication x2 andsubscript/superscripts x 1

2), use of brackets and other forms of spatial notation(e.g. in matrix notation). In addition, the Rule of Order for the unfolding of theprocesses (e.g. brackets, multiplication/division and addition/subtraction)means that the Operative processes do not unfold sequentially in a left to rightformat (see O’Halloran, 2005a, Chapter 5, for further discussion). The newsystems for encoding meaning are apparent if software applications for math-ematical notation are examined.

The grammatical systems in mathematical symbolic notation mean that thedescription of the relations can be rearranged for the solution to problems. Forexample, the equation in Figure 10.4 (Newton 1736: 67) may be rearranged toobtain an expression for y.

x3 = by2

y2 =b

y = b

Embedded process/participant configurations are indicated below usingsquared brackets [[ . . . ]]:

[[x × x × x]] = [[b × y × y]]

[[y × y]] = [[ ]]

y = [[ [[ ]] ]]


x3

x3

√

√

[[x × x × x]]

b

[[x × x × x]]

b

The symbolism deals with a limited experiential field primarily consisting ofRelational and Operative processes and variable participants. The aim ofencoding spatio-temporal relations of matter for predictive purposes means therange of circumstances which need to be symbolized is similarly narrow. Inter-personal meaning is steady and repetitive; giving information, and a restrictedrange of commands with a consistent level of modality with a high truth value.The net gain is an expansion of experiential meaning within a given domainaccompanied by an increased proficiency for logical meaning made possiblethrough the potential of the symbolism to reorganize the relationships via agrammar dedicated to this purpose. The basis for the logical relations is pre-established mathematical axioms, laws, theorems and results.

Patterns of variation are captured in mathematics and science, but onlywithin the semantic domains of Operative processes, participants and circum-stance. This is the Descartes’ price, the capture of dynamic relations at theexpense of the rich semantic realm of the everyday. Natural language has alarger repertoire of meanings, but the symbolism extends the semanticsafforded by language in terms of capturing the patterns of relations within theconfines of systems which are de-contextualized from the complexity of materialreality.

The grammar and semantics of mathematical symbolic notation requirefurther investigation, including exploration of how this form of semiosiscombines with visual imagery and language to construct a scientific view of theworld. However, at this stage we can see that the price for the semantic expan-sions afforded by mathematics and science is limitations with respect to the fieldof the description and the complexity of the situational context. The payoff isthe discipline of mathematics which expands through proliferation, thusoffering new tools for scientific description, explanation and prediction ofempirical data. The strong grammar of mathematics looks inwards, but thesemiotic tools are employed in the service of science which looks outwards tothe external world. The progress of science, the building of a complex hier-archical knowledge structure with increasing powers of description and pre-diction, is the result of the ‘left hand right hand’ relationship of mathematics toscience. It is hoped that the SF-MDA approach contributes to an understandingof that relationship.

Finally, SF-MDA social semiotic perspective is promising with regard to under-standing the relations between mathematics, science and technology, and fordeveloping the practices through which mathematics and science can be moreeffectively taught in schools. These two issues are discussed in turn below.

Visibility in the computer age

The impact of technology cannot be downplayed with respect to the develop-ment of mathematics and science. The printing press in the mid-fifteenthcentury led to the development of mathematical symbolic notation because, forthe first time, algorithms and symbolic notation could be extensively studied,developed and standardized (Eisenstein 1979; O’Halloran 2005a; Swetz 1987).


230

While the printing press aided the development of the grammar ofmathematical symbolism and the spread of the Hindu Arabic numerical system,the printed visual images were initially difficult to reproduce. In the ongoingdevelopment of modern science, the emphasis was directed towards thesymbolic notation and the written word (Davies et al. 1990). The symbolic wasthus favoured over the visual image. In other fields the race between the ‘trans-textual’ and the ‘transvisual’ continued until the advent of the audio-visual mix(Virilio 1994). Davies et al. (1990), for example, discuss the ‘picture explosion’from the 1950s onwards arising from television, video, the digital computer, theradio telescope, remote sensing technology and the Internet. The ‘pictureexplosion’ is a ‘meaning potential explosion’ afforded by manipulable dynamicvisual imagery. Development in semiotic forms of representation, and thusknowledge, are linked to technology.

Virilio (1994: 4) explains that instruments have a long tradition in the questfor visibility: ‘Al-Hasan ibn al-Haitam aka Alhazen’s camera obscura in the tenthcentury, Roger Bacon’s instruments in the thirteenth, the increasing number ofvisual prostheses, lenses, astronomic telescopes and so on from the Renaissanceon’ resulted in the ‘imperative to re-represent oneself, the imaging of the imagina-tion which was such a great help in mathematics according to Descartes . . .’.Visual images and scientific instruments were accompanied by three-dimensional geometrical models which were popular in the 1800–1860s inEurope, and models made from ‘wood and plaster, wire and paper, and glassand brass’ appeared in the late 1800s in Germany (Galison 2002: 304). Today,scientific visualizations are produced in a variety of forms through computer-generated imagery.

The use of computers is changing the face of mathematics and science. Forexample, mathematical results are increasingly visualized using computerizednumerical routines. These visualizations are slowly replacing the traditionalsymbolic solutions to problems favoured by Descartes and Newton. ‘Conjuringwith fluid patterns was one way to cope with dynamic phenomena eluding thereductionist thinking that had dominated the mathematical sciences sinceDescartes and Newton’, claims Waldrop (1992: 13), author of Complexity: TheEmerging Science at the Edge of Order and Chaos, cited in Stafford (1996: 94).Similarly, mathematical results are increasingly visualized using computerizednumerical routines which have replaced analytical solutions. Virilio (1994: 76)expects that ‘synthetic imagery’ (as a mode of statistical representation) shouldsoon be contributing to the development of the digitalized visual image as aform of reasoning. A systems-based multimodal grammatical approach to math-ematics and science is sensitive to such changes in orientation, which areafforded through computer technology.

Virilio (1994: 69–70) speaks of the current drive from visibility to invisibility inthe military-industrial complex, where ‘. . . what is perceived, is already finished . . .’.‘Repeating weapons’ have been replaced with ‘repeating images’ in the form ofcomputer graphics for the automatic recognition of shapes; the ability toremotely scan, detect and track objects means that once an object is located, itmay be destroyed (Virilio 1994: 70). The investment in decoy research is aimedat hiding real targets through deceptive techniques.


It is a war of images and sounds, rather than objects and things, in which winningis simply a matter of not losing sight of the opposition. The will to see all, to knowall, at every moment, everywhere, the will to universalised illumination; a scien-tific permutation on the eye of God, which would forever rule out the surprise,the accident, the irruption of the unforeseen. (Virilio 1994: 70)

While there has been a ‘shift of interest from the thing to its image, and espe-cially from space to time, to the instant . . . [from the] real-figurative dichotomyto the more relative actual-virtual’ (Virilio 1994: 70), tension still remains overthe roles of the visual image versus the symbolism in mathematics and science(e.g. Davies et al. 1990; Davis 1974; Galison 2002; Miller 1984; O’Halloran2005a). Despite the recent innovations in computer graphics, solutions areusually presented using text-based analytical means rather than visual imagery.The traditional prejudice against the use of visual images for establishing proofs(Davis, 1974) is gradually being eroded through the meaning potential madeavailable though computer-generated imagery and the ongoing expansion ofnew fields of mathematics, such as chaos theory and non-linear dynamicalsystems theory.

Multimodal grammaticality and the systems-based SF-MDA approach fore-ground mathematical and scientific knowledge as the product of interactingsystems involving technology and political, business and military interests.Complexity is acknowledged, for example, in the shift from conceptualizinglanguage as an isolated phenomenon to understanding language as a form ofsemiosis which developed – and makes meaning – in combination with othersemiotic resources. The systems-based multimodal grammatical approachpromises new insights across a wide range of fields, including mathematics andscience pedagogy.

Pedagogical implications of an SF multimodal grammatical approach

Some students successfully learn the discourse of mathematics and science,while many are relegated to failure. Patterns of achievement can be mapped toschools which differ in terms of school type (e.g. private versus state), socialclass, funding, resources and physical location. Students in private schools andstate schools which draw from affluent socio-economic populations enjoygreater success than those students from low socio-economic backgrounds.Increasingly, the political trend is to support these differences through fundingand tax breaks, creating sectors within the education system which range fromthe private/academic to the state/semi-quasi-academic-vocational. Within thisframework, cognitive ability is an acceptable guise for making invisible the polit-ical practices which enable wealth and status to be one major factor in successat school.

Muller (2004) and others (e.g. Bourne 2003) believe that successfulteaching/learning rests with the teacher. Muller (2004) states ‘teacher compe-tence is by far the most important factor in learner attainment; the second isthat in-service teacher training has had almost no effect’. The differences in


232

teaching/learning practices in mathematics classrooms differentiated on thebasis of socio-economic status have been documented (O’Halloran 2004,forthcoming). As the divide grows between different types of schools, so do theexperience, qualifications and salary of the teaching staff. These trends point tothe urgent need to develop theoretical and practical approaches for developingeffective teaching strategies, particularly for those teachers working withdisadvantaged students.

One answer lies with the systems-based SF-MDA approach, where recontextu-alizing practices in the classroom would include a discussion of the nature ofmathematical and scientific knowledge as seen intra- and inter-semioticallythrough the grammar of visual images, mathematical symbolism and language.This would include an explanation of the meaning potential of mathematicsand science, leading to an appreciation of how mathematics and science extendour understanding of the world, but only within limited domains. Descartes’price needs to be explained. Teachers can be provided with such knowledge sothat students understand what mathematics and science can do, and the meansthrough which this is achieved. In addition, the systems-based approach takesinto account how politics, business, the military and technology impact on thedevelopment of mathematics and science. Mathematics and science educationbecomes a discourse of critique, for an understanding of the mathematical andscientific view of the world means an appreciation of realms to which it can bereasonably applied. The scientific reductiveness of the social sciences, forexample, masks the complexity of the everyday human realm and preserves theinequalities generated through current political, economic, military andbusiness-based practices. Finally, the SF-MDA approach reveals the metaphori-cal nature of pedagogical discourse, where language serves as the metalanguagefor discussing mathematical and scientific knowledge (O’Halloran 2000, 2004,forthcoming).

Conclusion

Mathematics and science impact on our life every day. ‘The three great author-ities over life and death’, ‘the law, the army and medicine’, with ‘the power todetain, destroy and objectify’ (Virilio 1994: 43) embrace technology and ascientific view of reality. Enlightenment did not deliver the promised results(Horkheimer 1972; Horkheimer and Adorno 1972). Our material reality is theproduct of science, and we ourselves are increasingly the scientific subjects ofphysical, psychological, social and cultural innovation (Foucault 1991). Narra-tives of space, time and causality impact on lives in western consumer-orientatedsocieties in the form of day-to-day schedules and routines (see Tufte 1990). Ourlives are literally the product of mathematics and science, discourses we hardlyunderstand given the limited access to the semiotic tools of their construction:mathematical visual images, symbolism and scientific language. One place toaddress this imbalance is the education system, in order to make new genera-tions aware of alternatives to the path we ourselves have chosen with regard tothe development and uses of mathematics, science and technology.


Acknowledgements

Figure 10.1a is reproduced with permission from Bollmann-Bildkarten-Verlag GmbH &Co. KG, Lilienthalplatz 1, 38108 Braunschweig, Germany.

Figures 10.3–10.5 are reproduced with permission of the Bodleian Library, University ofOxford (Reference: Newton (1736), Call Number: SAVILE JJ 11).

Figure 10.6a is reproduced with permission of Dawson Books, Rushden, UK.Figures 10.6b–c reproduced with permission from Elsevier Ltd, Oxford, UK.

References

Baldry, A.P. (ed.) (2000), Multimodality and Multimediality in the Distance Learning Age. Cam-pobasso, Italy: Palladino Editore.

Baldry, A.P. and Thibault, P.J. (2006), Multimodal Transcription and Text Analysis. London:Equinox.

Barry, J. (1996), Measures of Science: Theological and Technical Impulses in Early ModernThought. Illinois: Northwestern University Press.

Bell, E.T. (1987), Mathematics: Queen and Servant of Science (first published 1951).Redmond, WA: Tempus Books of Microsoft Press.

Bernstein, B. (1990), Class, Codes and Control, Volume IV: The Structuring of Pedagogic Dis-course. London: Routledge.

Bernstein, B. (2000), Pedagogy, Symbolic Control and Identity: Theory, Research and Critique.Lanham: Rowman & Littlefield.

Bourne, J. (2003), ‘Vertical discourse: the role of the teacher in the transmission andacquisition of decontextualized language’, in European Educational Research Journal2(4), pp. 496–521.

Cheong, Y.Y. (2004), ‘The construal of ideational meaning in print advertisements’, inO’Halloran, K.L. (ed.), Multimodal Discourse Analysis. London and New York:Continuum, pp. 163–95.

Christie, F. (ed.) (1999), Pedagogy and the Shaping of Consciousness: Linguistic and SocialProcesses. London and New York: Continuum.

Christie, F. (2002), Classroom Discourse Analysis: A Functional Perspective. London and NewYork: Continuum.

Christie, F. and Martin, J.R. (eds) (1997), Genre and Institutions: Social Processes in the Work-place and School. London and New York: Continuum.

Coffin, C. (1997), ‘Constructing and giving value to the past: an investigation into secondschool history’, in Christie, F. and Martin, J.R. (eds), Genre and Institutions – SocialProcesses in the Workplace and School. London: Continuum, pp. 196–230.

Davies, D., Bathurst, D. and Bathurst, R. (1990), The Telling Images: The Changing Balancebetween Pictures and Words in a Technological Age. Oxford: Clarendon Press.

Davis, P.J. (1974), ‘Visual geometry, computer graphics and theorems of perceived type’,in Proceedings of Symposia in Applied Mathematics 20, pp. 113–27.

Descartes, R. (1683), Geometria, anno 1673 Gallice Edita / Postea Cum Notis Fl. de Beaune inLatinam Linguam Versa / et Commentariis Illustrata Opera Francisci a Schooten: Volume 1.Amsterdam.

Descartes, R. (1952), Descartes’ Philosophical Writings (ed. and tr. Norman Kemp Smith).London: Macmillan.

Descartes, R. (1954), The Geometry of Rene Descartes. New York: Dover.Descartes, R. (1985), The Philosophical Writings of Descartes: Volume 1. Cambridge: Cam-

bridge University Press.Eisenstein, E.L. (1979), The Printing Press as an Agent of Change. Cambridge: Cambridge

University Press.


234

Elkins, J. (1999), The Domain of Images. Ithaca, New York: Cornell University Press.Foucault, M. (1991), Discipline and Punish. London: Penguin.Funkenstein, A. (1986), Theology and the Scientific Imagination from the Middle Ages to the

Seventeenth Century. New Jersey: Princeton University Press.Galison, P. (2002), ‘Images scatter into data, data gathers into images’, in Latour, B. and

Weibel, P. (eds), Iconoclash: Beyond the Image Wars in Science, Religion, and Art. Cam-bridge, MA: The MIT Press, pp. 300–23.

Galvani, L. (1792), ‘Electrophysiology: experiments with frogs’, in De Viribus Electricitatis.Plate 1.

Golovin, A.A., Matkowsky, B.J. and Nepomnyashchy, A.A. (2003), ‘A complex Swift-Hohenberg equation coupled to the Goldstone Mode in the non-linear dynamics offlames’, in Physica D 179, pp. 183–210.

Guo, L. (2004), ‘Multimodality in a Biology textbook’, in O’Halloran, K. L. (ed.), Multi-modal Discourse Analysis. London: Continuum, pp. 196–219.

Habermas, J. (1974), Theory and Practice. Boston: Beacon Press.Halliday, M.A.K. (1978), Language as Social Semiotic: The Social Interpretation of Language and

Meaning. London: Edward Arnold.Halliday, M.A.K. (1993), ‘Some grammatical problems in scientific english’, in Halliday,

M.A.K. and Martin, J.R. (eds), Writing Science: Literacy and Discursive Power. London:Falmer Press, pp. 69–85.

Halliday, M.A.K. (2004a), Introduction to Functional Grammar. (third edn revised byC.M.I.M. Matthiessen; first edn 1985, second edn 1994). London: Arnold.

Halliday, M.A.K. (2004b), ‘The Language of Science’, in Collected Works of M.A.K. Halliday,edited by J. Webster. London and New York: Continuum.

Halliday, M.A.K. and Martin, J.R. (1993), Writing Science: Literacy and Discursive Power.London: Falmer.

Halliday, M.A.K. and Matthiessen, C.M.I.M. (1999), Construing Experience through Meaning:A Language Based Approach to Cognition. London: Cassell.

Horkheimer, M. (1972), Critical Theory. New York: The Seabury Press.Horkheimer, M. and Adorno, T.W. (1972), Dialectic of Enlightenment. New York: The

Seabury Press.Iedema, R. (2001), ‘Resemioticization’, in Semiotica 137, 1/4, pp. 23–39.Iedema, R. (2003), ‘Multimodality, resemioticization: extending the analysis of discourse

as a multisemiotic practice’, in Visual Communication 2(1), pp. 29–57.Koenderink, J. (1993), ‘Foreword’, in Ellis, S.R. (ed.), Pictorial Communication in Virtual

and Real Environments. London: Taylor & Francis, pp. viii–ix.Koestler, A. (1959), The Sleepwalkers: A History of Man’s Changing Vision of the Universe.

Harmondsworth: Penguin.Kress, G. (2003), Literacy in the New Media Age. London: Routledge.Kress, G., Jewitt, C., Ogborn, J. and Tsatsarelis, C. (2001), Multimodal Teaching and

Learning: The Rhetorics of the Science Classroom. London: Continuum.Kress, G. and van Leeuwen, T. (1996), Reading Images: The Grammar of Visual Design.

London: Routledge.Kuhn, T.S. (1970), The Structure of Scientific Revolutions. Chicago: University of Chicago

Press.Latour, B. and Weibel, P. (eds) (2002), Iconoclash: Beyond the Image Wars in Science, Religion,

and Art. Karlsruhe: ZKM.Lemke, J.L. (1998), ‘Multiplying meaning: visual and verbal semiotics in scientific text’,

in Martin, J.R. and Veel, R. (eds), Reading Science: Critical and Functional Perspectives onDiscourses of Science. London: Routledge, pp. 87–113.

Lemke, J.L. (2002), ‘Travels in Hypermodality’, in Visual Communication 1(3), pp.299–325.

Lemke, J.L. (2003), ‘Mathematics in the middle: measure, picture, gesture, sign, andword’, in Anderson, M., Sáenz-Ludlow, A., Zellweger, S. and Cifarelli, V.V. (eds),


Educational Perspectives on Mathematics as Semiosis: From Thinking to Interpreting toKnowing. Ottawa: Legas, pp. 215–34.

Lester, P.M. (2000), Visual Communication: Images with Messages. Belmont, CA:Wadsworth/Thompson Learning.

Martin, J.R. (1992), English Text: System and Structure. Amsterdam: John Benjamins.Martin, J.R. (2002), ‘Fair trade: negotiating meaning in multimodal texts’, in Coppock,

P. (ed.), The Semiotics of Writing: Transdisciplinary Perspectives on the Technology ofWriting. Turnhout, Belgium: Brepols, pp. 311–38.

Martin, J.R. and Rose, D. (2003), Working with Discourse: Meaning Beyond the Clause.London and New York: Continuum.

Martin, J.R. and Veel, R. (eds) (1998), Reading Science: Critical and Functional Perspectives onDiscourses of Science. London: Routledge.

Martinec, R. (2000), ‘Types of processes in action’, in Semiotica 130–3/4, pp. 243–68.Martinec, R. (2001), ‘Interpersonal resources in action’, in Semiotica 135–1/4, pp. 117–45.Martinec, R. (2004), ‘Gestures that co-concur with speech as a systematic resource: the real-

ization of experiential meanings in indexes’, in Social Semiotics 14 (2), pp. 193–213.Miller, A.I. (1984), Imagery in Scientific Thought: Creating Twentieth-Century Physics. Boston:

Birkhäuser.Muller, J. (2000), Reclaiming Knowledge: Social Theory, Curriculum and Educational Policy.

London: Routledge.Muller, J. (2004), ‘On the shoulders of giants: a digression on knowledge, curriculum,

and finally, the teacher’. Plenary paper at Reclaiming Knowledge Conference.Faculty of Education, University of Sydney, 13–15 December 2004.

Newton, I. (1736), The Method of Fluxions and Infinite Series; with its Application to theGeometry of Curve-Lines; With Comments by John Colson. London: printed by HenryWoodfall and sold by John Nourse.

Newton, I. (1968), Mathematical Principles of Natural Philosophy. London: Dawson.O’Halloran, K.L. (1999), ‘Towards a systemic functional analysis of multisemiotic math-

ematics texts’, in Semiotica 124–1/2, pp. 1–29.O’Halloran, K.L. (2000), ‘Classroom discourse in mathematics: a multisemiotic analysis’,

in Linguistics and Education 10(3), pp. 359–88.O’Halloran, K.L. (2003a), ‘Educational implications of mathematics as a multisemiotic

discourse’, in Anderson, M., Sáenz-Ludlow, A., Zellweger, S. and Cifarelli, V. V. (eds),Educational Perspectives on Mathematics as Semiosis: From Thinking to Interpreting toKnowing. Ottawa: Legas, pp. 185–214.

O’Halloran, K.L. (2003b), ‘Intersemiosis in mathematics and science: grammaticalmetaphor and semiotic metaphor’, in Simon-Vandenbergen, A.-M., Taverniers,M. and Ravelli, L. (eds), Grammatical Metaphor: Views from Systemic Functional Linguis-tics. Amsterdam: John Benjamins, pp. 337–65.

O’Halloran, K.L. (2004), ‘Discourses in secondary school mathematics classroomsaccording to social class and gender’, in Foley, J. A. (ed.), Language, Education andDiscourse: Functional Approaches. London and New York: Continuum, pp. 191–225.

O’Halloran, K.L. (2005), Mathematical Discourse: Language, Symbolism and Visual Images.London and New York: Continuum.

O’Halloran, K.L. (forthcoming), A Multimodal Approach to Classroom Discourse. London:Equinox.

O’Toole, M. (1994), The Language of Displayed Art. London: Leicester University Press.Peirce, C.S. (1991), ‘Prolegomena to an apology for pragmaticism’, in Hoopes, J. (ed.),

Peirce on Signs: Writings on Semiotic by Charles Sanders Peirce. Chapell Hill: University ofNorth Carolina Press, pp. 249–52.

Prigent, A., Gregoroire, G., Chate, H. and Dauchot, O. (2003), ‘Long-wavelength modu-lation of turbulent shear flows’, in Physica D 174, pp. 100–13.

Rotman, B. (1995), ‘Thinking dia-grams: mathematics, writing and virtual reality’, inSouth Atlantic Quarterly 94(2), pp. 389–415.


236

Royce, T. and Bowcher, W. (eds) (in press), New Directions in the Analysis of MultimodalDiscourse. New Jersey: Lawrence Erlbaum Associates.

Saint-Martin, F. (1990), Semiotics of Visual Language. Bloomington: Indiana UniversityPress.

Simon-Vandenbergen, A.-M., Taverniers, M. and Ravelli, L. (eds) (2003), GrammaticalMetaphor: Views from Systemic Functional Linguistics. Amsterdam: John Benjamins.

Stafford, B. M. (1989), ‘The new immaterialism: Kenneth Snelson imagines the atom’, inKenneth Snelson: The Nature of Structure (Exhibition Catalogue). New York: New YorkAcademy of Sciences, pp. 51–7.

Stafford, B.M. (1996), Good Looking: Essays on the Virtue of Images. Cambridge, MA: MITPress.

Swetz, F.J. (1987), Capitalism and Arithmetic. La Salle, IL: Open Court.Thibault, P. (2000), ‘The multimodal transcription of a television advertisement: theory

and practice’, in Baldry, A. P. (ed.), Multimodality and Multimediality in the DistanceLearning Age. Campobasso, Italy: Palladino Editore, pp. 311–85.

Thibault, P. (2001), ‘Multimodality and the school science textbook’, in Carol TaylorTorsello, G.B., and Nicoletta Penello (eds), Corpora Testuali per Ricerca, Traduzione eApprendimento Linguistico. Padua: Unipress, pp. 293–335.

Tufte, E.R. (1990), Envisioning Information. Cheshire: Graphics Press.Unsworth, L. (2000), ‘Investigating subject-specific literacies in school learning’, in

Unsworth, L. (ed.), Researching Language in Schools and Communities. London and NewYork: Cassell, pp. 245–74.

Unsworth, L. (2001), Teaching Multiliteracies across the Curriculum: Changing Contexts of Textand Image in Classroom Practice. Buckingham, UK: Open University Press.

van Leeuwen, T. (1999), Speech, Music, Sound. London: Macmillan.van Leeuwen, T. (2005), Introducing Social Semiotics. London: Routledge.Veel, R. (1999), ‘Language, knowledge and authority in school mathematics’, in Christie,

F. (ed.), Pedagogy and the Shaping of Consciousness: Linguistic and Social Processes.London: Cassell, pp. 185–216.

Ventola, E., Charles, C. and Kaltenbacher, M. (eds) (2004), Perspectives on Multimodality.Amsterdam: John Benjamins.

Virilio, P. (1994), The Vision Machine. Bloomington: Indiana University Press.Waldrop, M.M. (1992), Complexity: The Emerging Science at the Edge of Order and Chaos. New

York: Simon & Schuster.Wigner, E.P. (1960), ‘The unreasonable effectiveness of mathematics in the natural

sciences’, in Communications on Pure and Applied Mathematics 13, pp. 1–14 (reprintedin Mickens, R. E. (1990), Mathematics and Science. Singapore: World Scientific, pp.291–306).

Wildbur, P. (1989), Information Graphics: A Survey of Typographic, Diagrammatic and Carto-graphic Communication. New York: Van Nostrand Reinhold.

Wildbur, P. and Burke, M. (1998), Information Graphics: Innovative Solutions in ContemporaryDesign. London: Thames and Hudson.


Section 3

Research Prospects – Exploring Uncommon Sense


We have organized this closing stage of our conversation here as a kind of inter-view, based on an agenda of issues which we hope will inspire future research.The contributions are from the editors, Fran Christie and Jim Martin, alongwith Karl Maton and Joe Muller, who converse on a first name basis below.

1. Why talk?

Jim: As we signal early in this book, we’ve tried here to push along the negotia-tion of ideas initiated by Bernstein, Halliday and Hasan into a second genera-tion of research. What is it that enables this conversation to continue, acrossdisciplines which usually have very little to say to one another?

This is my own third intense engagement with Bernstein’s thinking. The firststemmed from my concern with educational failure, and its relation to thecoding orientations privileged or not by institutionalized learning; the secondhad to do with our struggle to dislodge the hegemonic position of progressiveeducation as far as literacy teaching was concerned; this time round we’refocusing on hierarchy and knowledge structure, and its implications for learnerpathways in school.

It seems to me that while my functional linguistics colleagues and I are goodat seeing how meaning is realized in texts, we’re very much weaker at under-standing how meaning is distributed in society. The ways Bernstein and his col-leagues talk about this unfolds for me as a kind of revelation, and an essentialunderpinning for any kind of intervention we try to make with social justice inmind. It is hard to read across our respective knowledge structures; but withoutthis reading we can’t engineer socially significant changes to educationalpractice – we tend to waste our time.

Fran: My path into the issues here was different from Jim’s. As a young schoolteacher I developed an interest in language in two senses: firstly, I found thattraditional school grammar was not relevant for most of the children I taught;secondly, I became interested in the ways teachers and children used languagein schools, and I observed that children performed in differential ways whichneeded explanation in some systematic way. It was through these interests thatI turned to Halliday’s SFL. It offered both a possible pedagogic grammar and atool for the analysis of classroom talk. The interest in classroom talk led me toBernstein’s work on pedagogy, and thence to notions of knowledge structure.

11 Taking stock: future directions in research inknowledge structure

Frances Christie, J.R. Martin, Karl Maton and Johan Muller

240

Bernstein’s unfinished work on a theory of knowledge structure provided apowerful response to the spread of progressivist and constructivist theories ofcurriculum, which had taken grip on much Western schooling by the end of thetwentieth century. What enables us to keep talking is that the two disciplinesprovide a theoretical frame for addressing the nature of knowledge structureand of curriculum design in such a way that each informs the other.

Joe: The work of Halliday and SFL has until recently been a part of my con-sciousness only in a general sort of way. It is only since my colleagues and I havebeen working on rendering the knowledge structure of curricula visible, usingBernstein’s theoretical tools, that I began to read the work of Fran Christie, JimMartin and their colleagues seriously. I found to my surprise that they had madeconsiderable progress in making visible the metalinguistic structure of disci-plines and curricula. This enterprise, I found, is not only directly parallel to ourwork in curriculum but shares many of the same epistemological and politicalconvictions. And this work makes clear not only that knowledge lives in texts,but how. The advance is considerable.

Karl: The question of what enables our conversation to continue is crucial.Contemporary social science is replete with calls for ‘multidisciplinary’, ‘inter-disciplinary’, even ‘post-disciplinary’ work, but all too often these are realized asnon-disciplinary monologues rather than fruitful cross-disciplinary dialogue.For me, three attributes of SFL and Bernstein sociology stand out here. First,both share what Bernstein described as an allegiance to a problem rather thanto an approach. Intellectuals in both traditions are willing to look beyond theconfines of their knowledge structures for conceptual tools that enable them toexplain better that part of the world they focus on. Second, they both attemptto generate strong external languages of description, concepts that get to gripswith problems in empirical research. This gives them the possibility of a sharedpurchase on the world, enabling dialogue. For a sociologist, SFL is extremelyimpressive for its technical armoury – its conceptual framework is elaborate,detailed and aims for clarity. Lastly, both theories downplay the significance ofwho is speaking and instead focus on the explanatory power of their approachesin addressing defined problems. The degree of detail SFL can analyse in texts iseye-opening to me, and working with linguists has made me more aware of thesignificance of language to my own interests. Whether we borrow from eachother’s conceptual toolkits or manage to integrate the two theories, or perhapsevolve from cross-fertilization towards integration, is something to be seen – it’san exciting prospect. Politically, what impresses me about both theories is thatthey start with questions rather than with solutions – what Isaiah Berlin called aculture of consequence, rather than one of commitment, recognizing it is notenough to be well intentioned, one also needs epistemologically powerfulknowledge.


2. Horizontal and vertical discourse

2.1 The borders of horizontal and vertical discourseJim: One thing I remain puzzled about, not addressed in this book, is why Bern-stein places crafts in vertical discourse. Since he characterizes these as havingweak grammars and tacit transmission (showing and modelling proceedingdoing), I expect he would include trades, hobbies, sport and recreation heretoo.

For various reasons, I’d prefer to treat these as horizontal discourses of aspecialized kind. They’re learned the same way as other horizontal discourses(via participation and ostensive definition), don’t traditionally involve institu-tionalized learning and don’t depend on grammatical metaphor to constructtheir expertise. What are the grounds for considering discourses of this kind‘vertical’? This is a very important educational issue since radical progressiveand constructivist educators seem to believe that all vertical discourses have tobe reconstructed as crafts in order for students, especially younger or lesssuccessful students, to really learn them.

Fran: Well, however we address the last issue that Jim raises here, I don’t think itwill be by revisiting what Bernstein said of the crafts in vertical discourse. Thefundamental issue that faces us for pedagogy is: how do we make it possible forpeople to move from horizontal to vertical discourse? As a first step we needbetter teacher education, which, among other things, must be more upfrontabout addressing the pedagogic discourse of schooling, and more politically sen-sitive in the best sense. Thus, for example, it will pay more attention to difference(though not the rather vacuous notions of ‘individual difference’ with whicheducational psychologists have long had a field day!). It will acknowledge thedifferent meaning codes that children act with when they come to school, and itwill not, like much twentieth-century educational theory, subscribe to a rathernaive notion of the idealized subject who is ‘self-regulating’, and who developsbenignly, with little assistance from formal instruction (Bernstein 2000: 43). Itwill also, for reasons explored by both Joe and Karl (this volume), not disguisethe essentially hierarchical nature of much school knowledge, making the dif-ferent knower codes and knowledge codes more explicitly available to students.Here is a nice instance of the ways the two disciplines assist each other: thesociology of Bernstein, Muller and Maton provides a theoretical frame to identifythe problem, while the SFL theory provides a language for addressing it.

Joe: Through a 2004 PhD study of a master-apprentice relation in cabinetmaking (a trade close to the carpentry-woodwork examples Bernstein oftenuses), Gamble illuminates the nature of a tacit knowledge base in craft. Herwork shows that while the pedagogic outcome is an external performance thatmust meet explicit and exact criteria, this performance rests crucially on aninternalized competence – or a capacity to visualize the relationship betweenparts and whole, in both space and time. Visualization stands in place of a non-articulable or ineffable ordering principle that acts as a ‘glue’ to make all thesegments hang together. It is this, what Polanyi calls an ineffable relationship

241FUTURE DIRECTIONS IN RESEARCH

242

between particulars jointly forming a whole, even though all the particulars areexplicitly specifiable, that makes modelling the only possible transmission-acquisition practice. Visualization thus compensates for the lack of a clearsyntax by allowing a formal principle of arrangement to be grasped inembodied form. This goes for the generation of novelty too: as Bernsteinexplained in an email, ‘There will be a strong visualizing of the new. That maybe why it is difficult to explain in words. You “see” the design rather thanverbalize it.’ This makes craft homologous to knowledge forms, which althoughfar more strongly developed in terms of vertical and codified chains of abstrac-tion, share a common feature in that they make the jump from ‘token’ to ‘type’and thus generate meaning at a remove from the everyday. Bernstein acknow-ledged in an email the difficulty of talking about craft as having an ‘internalgrammar’, and proposed the term ‘condensed recontextualization’ for the tacittechnical principles of craft operation.

Constructivist interpretations of apprenticeship subsume the nature of theknowledge transmitted and the form of its transmission into the notion of‘shared or situated practice’ that fits with post-modern versions of hetero-geneous local knowledges not transferable to context-independent knowledge.In this version the notion of apprenticeship becomes so pervasive that there isnothing that is not apprenticeship. The term loses its meaning.

Karl: Fran’s question is the key issue here: what enables the move into verticaldiscourse? Is it helped by recontextualizing existing disciplines to become morehomologous to horizontal discourse? As Fran suggests, it requires making thecodes underlying curriculum and knowledge structures more explicit – we needto work out ways of giving different kinds of pupils the keys to the code. Not allpupils arrive with the means to recognize or realize the code required to learnsuccessfully. We need to work at creating forms of curriculum and pedagogy thatprovide pupils with what we possess, as quickly as possible.

Calls to change education to be more like everyday knowledge fail to under-stand the difference between horizontal and vertical discourses; in effect, theycall for an end to education – they want to make the sacred profane. In the fieldof information technology education research, for example, a widely madeargument currently is that children are ‘digital natives’ who arrive at school withimmense amounts of knowledge. The argument that they already know how touse mobile phones, computers, ipods, etc. is then used to justify moves to peer-group learning and a reconstruction of the curriculum around everyday uses oftechnology. Such arguments can only be maintained so long as one does nothave a theory of education or knowledge. It relies on negating differencesbetween horizontal discourse and its forms of circulation, and vertical discourseand its specific forms of pedagogy. What research is beginning to show in thisexample is that such moves to ‘authentic learning contexts’ and peer-teachingtend to leave pupils where they began: with only context-dependent practicalmastery rather than context-independent symbolic mastery. This is why recog-nizing the differences between horizontal and vertical discourse is crucial forenabling pupils to succeed in education. Here SFL and sociology have a poten-tially crucial role to play in enabling everyone to learn.


2.2 Grammatical metaphor and vertical discourseJim: From the perspective of functional linguistics, grammatical metaphor is thekey resource used to construct the uncommon sense knowledge of vertical dis-courses and is characterized in functional linguistics as involving stratal tension,with grammar symbolizing alongside realizing semantics, and the meaning ofthe metaphor involving its grammatical reading, its semantic reading and theimport of the tension between the two. The technicality of the functionallinguistic account makes the concept a challenging one to explain to outsiders.Does anyone have any thoughts on strategies we might use to make this crucialunderstanding accessible to academics in other disciplines, and equally impor-tantly for teachers and students of vertical discourses in secondary school?

Fran: This is a hard issue. We already know, from curriculum experience inAustralia at least, that school teachers in discipline areas like the sciences aremore disposed than their counterparts in the humanities and social sciences toengage with technical discourse, and this reflects their training in the universitydisciplines. There is, in other words, a relationship between the discourse of theuniversity disciplines and their recontextualized forms in schools, although asJim notes, under the impact of constructivist models of curriculum, someaccounts of the secondary science curriculum have shed their commitment to avertical discourse. The dilemma for those who espouse constructivist models ofcurriculum in all areas of schooling is that they actually look for and rewardvertical discourse in their teaching and their assessment. This is probably theway into persuading academics and school teachers in the various areas ofknowledge to have another look at what they are doing: identify instances of texttypes that teachers reward in their students’ work, bring some of their linguisticfeatures to consciousness, and turn this understanding into teaching practice.

Karl: Excellent question, Jim. The truth is no guarantee of belief. Two inter-related issues spring to mind here: appealing to the people we wish to help, andmaking our technical language accessible. When presenting papers or writingin sociology, education or cultural studies, I sometimes hear reactions to therelatively simple conceptual framework I employ, suggesting the world is morecomplex than the concepts suggest or implying the analysis is somehow coldand needs more warm bodies of knowers. My own strategy is to appeal to therational, to the fact that in higher education we are still ostensibly concernedwith understanding and explaining the world. I simply ask whether the analysisseems to make sense, could the concepts be of use in explaining issues ofconcern, and if not let’s come up with something better. Similarly, as Fran says,one can appeal to the desire of teachers to help pupils learn. It’s an obviouspoint but, in terms of making a technical language more accessible, we need dif-ferent ways of talking to different audiences. In speaking across the languagesof the horizontal knowledge structure of social science as a whole, one has to bea little multilingual, or at least appeal to the codes they operate with. For anaudience of academics unused to the kind of language we use in the Bernsteintradition, I employ concepts such as ‘knowledge code’ or ‘knower code’ in a lessexplicitly defined way (such as ‘knower approach’ or simply ‘emphasizes the


244

knower’), without using classification and framing or epistemic relation andsocial relation. However, this may be easier to do than for systemic linguists,because our language is less technical, or at least less extensive. I’m not sayingwe should work against the technicalization of our discourse by dumbing downor not using technical language at all, but simply highlighting that academicsand teachers may be operating with different codes from our own, which mayrequire different modes of expression.

3. Horizontal and hierarchical knowledge structures

3.1 Verticality and grammaticalityJim: Joe (this volume) develops Bernstein’s notion of hierarchality with respectto what he calls grammaticality and verticality: ‘grammaticality determines thecapacity of a theory or a language to progress through worldly corroboration;verticality determines the capacity of a theory or language to progress integra-tively through explanatory sophistication. Together, we may say that these twocriteria determine the capacity of a particular knowledge structure to progress.’Following on from the discussion of grammatical metaphor above, I’m wonder-ing about the role of technicality in determining degrees of grammaticality andverticality. Technicality involves deploying grammatical metaphor to defineterms and place them in precise relationships to one another. In my chapter(this volume) I gave examples from Bernstein (e.g. ‘a horizontal discourse entailsa set of strategies which are local, segmentally organised, context specific anddependent, for maximising encounters with persons and habitats’). Technicalterms engendered by the process of definition subsume the knowledge createdby the grammatical metaphors, but are not themselves metaphorical. In effect,technicality lightens up the discourse, making room for further grammaticalmetaphors to move the development of hierarchical knowledge along – andthus contribute to verticality. At the same time technical terms contribute to thedevelopment of an explicit network of precise concepts which can be testedagainst data – and thus contribute to grammaticality (i.e. the possibility ofempirical disconfirmation). So, from a linguistic perspective, if grammaticalmetaphor is the watershed demarcating horizontal and vertical discourse, is dis-tilling grammatical metaphor as technicality the key to establishing degrees ofverticality and grammaticality?

Fran: Well yes, I think it is, though the whole issue needs a great deal of unpack-ing; this is one of the issues that Mary Macken-Horarik, Karl and I would like topursue in the future with respect to subject English – surely one of the most dif-ficult of all subjects in the curriculum in which to propose to teachers that thereis a ‘verticality’ for them to explore! The difficulty about subject English is thatthe language is both the instrument of teaching and learning and the object ofstudy. If English is to be ever understood as having ‘verticality’, this will be builtaround what we know of grammatical metaphor. Its emergence is a develop-ment of late childhood to adolescence (though many children with literacydifficulties don’t master it), and its emergence occurs when students enter the


secondary school, with its different curriculum (though the nature of thatcurriculum is often subverted these days by constructivist notions).

Successful control of literacy for the secondary school, in which grammaticalmetaphor plays a crucial role, facilitates entry to ‘uncommonsense’ experience,because it allows the writer to be distanced from the immediate experience,achieving a degree of detachment from the event. It is precisely this capacity todraw back from experience and build abstraction – be that achieved in writinga valued story, in reviewing a novel or film, or in writing an expository text onsome social issue – that subject English actually rewards, though for the mostpart, English teachers are not aware of this. In the first instance, Englishteachers need to develop an understanding of the technical terminologyinvolved in understanding how the rewarded language is constructed, and thenturn this into a metalanguage they can use in teaching and working with theirstudents. Thus, they can make rather more visible than presently applies what itis that people do with language in order to build the kinds of texts and theirknowledge that subject English rewards. An infinitely trickier issue, however, willinvolve taking the next step: how we take an understanding of the kinds ofcapacities in using written language laid bare in the first step just described, anddevelop a coherent and integrated account of the English curriculum, such thata degree of sequence and progression across the years of schooling is achieved.This is an issue that cries out for empirical research!

Joe: The issue of what would constitute technicality in horizontal knowledgestructures has not been addressed. I would suspect that in disciplines thatproliferate parallel languages, hence with weak subsumption, technicality is tooinfused with lexical metaphor to be stable, with the result that neophytes oftendon’t know whether they are using the perhaps imperfectly technicalized termscorrectly or not. This is why students often learn usage of such discourses not bylearning the principles but by mimicking the discursive style of their teachers,and in so doing acquire the disciplinary ‘gaze’, a rather weak form of con-sciousness specialization.

Karl: This is one of the questions that I think this book really presses us on asthe next stage for our thinking. To be speculative, if grammatical metaphor is akey to moving from horizontal to vertical discourse, then perhaps it is the formtaken by the process this comprises that shapes differences between hierarchi-cal and horizontal knowledge structures. As Joe is saying, in horizontal knowl-edge structures technicality is prolific but of a different order. Sociologists,for example, often ‘pack up’ (to be crude) a whole set of meanings,allusions, political attitudes, aptitudes and beliefs into particular terms, such as‘ideology’, ‘hegemony’ or ‘rhizomes’. There are extensive literatures on theirdefinitions, but in common use such terms often signal the author’s stance, witha high normative and political loading. One aspect of the difference may be thatunderlying grammatical metaphor in hierarchical knowledge structures is aknowledge code (see my paper, this volume), where its strongly bounded andcontrolled epistemic relation to the object and other knowledge helps regulatethe form it takes, enabling subsumption and a stronger grammar; while, in


246

horizontal knowledge structures, technicalization may be shaped by inter alia aknower code, where this epistemic relation is weaker and more emphasis isplaced on the social relation to dispositions, aptitudes, attitudes, etc. Thoughthis gives the resulting language a weaker external grammar, the process (andthe basis of legitimate insight) is still regulated, affecting the way it gets taughtand the kinds of consciousness specialized.

3.2 SFL and verticality Jim: Joe (this volume) characterizes verticality as having to do with ‘how theorydevelops. In hierarchical knowledge structures, it develops through integration,towards ever more integrative or general propositions, the trajectory of devel-opment of which lends hierarchical knowledge structures a unitary convergentshape.’ Bernstein uses the imagic metaphor of a triangle to symbolize this con-vergent apicality. Triangularity is a useful metaphor for considering the progressin canonical hierarchical knowledge structures such as physics and biology, andperhaps within some of the specialized languages of certain horizontal know-ledge structures (those with proliferation inhibiting verticality in Joe’s terms).But I wonder how appropriate the triangle metaphor is for a polysystemicknowledge structure such as systemic functional linguistics, and whether thenature of its object of inquiry, language (and semiosis in general) has engen-dered its multiperspectival path of evolution (so we end up with a prism, not apyramid).

Alongside hierarchy, in other words, I’d like to introduce here the notion ofcomplementarity – the idea that we can’t always find ‘ever more integrative andgeneral propositions’ but have to learn to live with (and enjoy) complemen-taries which taken together exhaust our understanding of the field but whichcannot be reduced to a single apical insight. A familiar example of such a com-plementarity would be light having to be interpreted as both particle and fieldin physics (consider also Joe’s complementarity of verticality and grammatical-ity as co-determinants of knowledge structure progression). Is the image of ‘yinand yang’ in other words a possible reading or extension of ‘ever more integra-tive or general propositions’, alongside the image of the triangle?

Joe: I would regard complementaries as belonging to the same apical node, soto speak. It’s the supplementaries that are the competitors, and that logicallycan’t co-exist on the same apical level. That a proposition can be broken downinto independent propositional clusters doesn’t mean that the clusters areconceptually incompatible. (Recall that it is of the essence of Derrida’sapproach to deconstruction that there must always be a supplement that can’tbe subsumed; for Bernstein, the more hierarchical the knowledge structure, themore intolerant of supplements it has to be.) From my point of view, comple-mentarity and hierarchy are, well, complementary. On the other hand,Bernstein was at least a closet Hegelian, developing his thoughts by antithesesfollowed by syntheses. Perhaps this gets us closer to yin and yang?

Karl: I agree with Joe – complementarity and triangularity are not mutuallyexclusive. A question: are the complements (the systems) brought together at a


higher level of abstraction? If not, can they be? Why stop and declare thecomplements as the height of the ambition to climb?

3.3. Science, social science and humanitiesJim: Joe (this volume) proposes degrees of verticality as a criterion for consign-ing disciplines to hierarchical or horizontal knowledge structures, and withinhorizontal knowledge structures to constrained proliferation or unconstrainedproliferation types. Degrees of grammaticality complement this with respect tothe possibilities for worldly corroboration. Does this give us adequate resourcesfor interpreting Bernstein’s hierarchical and horizontal opposition as a cline,and satisfactorily positioning the knowledge structures of social science along it?

I remain puzzled by my own field of linguistics, for example, which presentsitself as the science of language and as such strives for as much verticality andgrammaticality as possible. This would seem to predict that the proliferation ofnew theories of language would be highly constrained and that decades ofresearch would lead to integration. It might be argued however that prolifer-ation of incommensurable theories is in fact rampant, and that what we in factexperience is a language proliferating horizontal structure par excellence. Thismay simply be because linguists can’t agree on what language is, and withoutconsensus on data verticality charges off in different directions, and grammat-icality ‘tests’ these divergent models against radically different phenomena. Thiscertainly inhibits co-operation within the discipline, and can cause confusion inthe community as far as interventions and applications are concerned.

Fran: Most of Jim’s last observations apply equally to English, as that is conceivedin university departments of that name. All this suggests to me that we have away to go in straightening out the degrees of verticality that are found/arepossible in the hierarchical and horizontal knowledge structures in the socialsciences and humanities. About this matter, some of Halliday’s observations areworth noting. Thus, he is given to saying that the semiotic sciences in the twen-tieth century, including linguistics, stood in relation to the objects of their studyclose to where the natural sciences stood to their objects of study in the six-teenth century. The twenty-first century, he has said, is the century in which thesemiotic sciences will come of age, and he is of the view that linguistics will be apivotal discipline in the endeavour. The paradigms found in his linguistics willhave relevance for the other social science disciplines, though that does notpreclude adoption of other paradigms. He also has said that it is no accidentthat the natural sciences got their acts together rather earlier than the socialsciences: the phenomena of the natural world are certainly hard to study, butthose of the social world – ourselves in fact – are infinitely harder (Halliday1993). Hence he would agree with a point Joe makes (this volume) that we stillare clumsy about the ways we talk about the things we do talk about – we are stillshaping the intellectual tools to build our sense of the social world.

Joe: I wonder whether SFL and theoretical linguistics, say, are simply two alter-native languages of linguistics or whether, as is the case for sociology andanthropology, and human geography, these are disciplines which, though


248

taking the same broad class of objects, are more fairly treated as differentthough related disciplines? I would regard the range of approaches within socio-linguistics as different languages, for example; and the rest of us social scientistswould dearly love to see a synthesis within that grouping. But I would not havethought that it could be easily achieved across all the ‘hyphen-linguistics’ in thetextbook. But perhaps I mistake the integrative ambitions of SFL?

Karl: For me this highlights a couple of issues. First, a discipline is not merely itsknowledge structure. A strong grammar and high degree of integration andsubsumption are perhaps necessary but not sufficient conditions for avoidingproliferation of languages within the knowledge structure. One has also tocontend with the tradition of the intellectual field as a social field of practice,where actors are invested in different approaches. This brings us back to theearlier question of speaking across segments and the issue of persuasion.Second, perhaps what is required is a metatheory, a theory which can determinewhich of the languages of linguists are more ontologically sound than others.This is perhaps one reason why SFL is looking to approaches such as that ofBernstein – to provide the basis of a means for understanding what is going onin linguistics as a whole, and to help construct a ruler for determining whatapproaches are more or less empirically adequate and for what phenonema.Bernstein did not focus very much on issues of ontology and epistemology,though there are approaches compatible with his and, I believe, with SFL thatmay offer further tools for such a task. In the sociology of education, forexample, some of those working with Bernstein’s approach are drawn towardswhat is called social realism or ‘critical realism’, a metatheory that underlaboursother theories by providing a means of discussing and criteria for judgingbetween theories on the basis of which are more or less empirically adequate. Itis an ‘underlabourer’ because it describes what a theory would have to do, butnot what that theory would say. This has had perhaps its biggest impact in eco-nomics, where mainstream economics has been strongly critiqued for beingdivorced from the real world and lacking fruitful empirical application, and thenature of theories required to make sense of the economic world are beingoutlined, underpinning the rise of what is being called ‘post-autistic economics’.

3.4 SFL as metalanguageJim: What does it mean to model discourse through the metalinguistic andmetasemiotic prisms of SFL? As far as the professional discourse of knowl-edge production is concerned, this remains of course a linguistics exercise –and linguists need to keep in mind that this modelling involves recontextual-ization from one discipline to another, and that the amount of verticality adiscipline deploys relative to linguistics will impact strongly on the nature ofthe recontextualization.

For pedagogic discourse, the impact of such modelling is even more chal-lenging. For hierarchical knowledge structures like science, language educatorsseem to be suggesting that an additional language be introduced, in effecthorizontalizing the school discipline. For horizontal knowledge structureslike English or history, language educators have advocated introducing an addi-


tional language with far more verticality than the host discipline, in effect verti-calizing it. I remain convinced that intervening across disciplines along theselines is the only way to make the knowledge structures schools were invented totransmit democratically available, across the range of coding orientationsstudents bring with them. Is there any way of making these marriages of con-venience more comfortable ones for the host and intervening disciplines?

Joe: What Fran and Mary have shown (this volume) is how the school subjectEnglish, by moving in Australia from what Bernstein called a performancepedagogy to a competence pedagogy, has had its knowledge structure renderedinvisible. A central feature of competence modes is that they focus on theacquirer and her sensibilities, and it is the latter which are to be tutored andrefined towards an end which is invisible to the acquirer. At the same time, hier-archy – that is, progression and progress – is rendered invisible, and the acquir-ers are stranded in the present tense of their own productions. The paper alsoshows that the only recourse for concerned pedagogues (like Mary and Fran)who wish to render verticality visible is to reconstruct a hierarchy in the meta-language. For ideological reasons they are prevented from making verticalityvisible by retrieving a canon (making the knowledge structure visible), but, asthey show, they can at least make it visible metalinguistically. The ‘knowledge’option is blocked, but not the ‘metalinguistic’ option.

What this makes plain to me is that it is virtually impossible to speak about theknowledge structures of subjects in a competence mode because they are, inprinciple, invisible. To make them visible is to change the mode, which onecan’t do unilaterally. Therein lies the difficulty in talking about English as aknowledge structure: under prevailing conditions, one literally can’t. Does thatmean there is no verticality? No, it does not. Jim shows (this volume) thatcertain metalinguistic features are entailed by vertical discourse. Thus, byshowing verticality in the metalanguage Fran and Mary show, by implication,that the knowledge structure must be vertical; it’s an inference, but a prettywell-founded one because, as Jim says, ‘if no grammatical metaphor then no ver-ticality’, and vice versa. In this way, SFL allows one to circumvent the prevailingideology, and becomes a powerful tool for intervention.

This clarifies for me the power of the term ‘verticality’; it applies both to theknowledge structure and to the metalanguage, and allows one to make infer-ences from one to the other.

It may be that all knowledges are virtual, and come to light only by consciouseffort on the part of pedagogic recontextualizers. SFL researchers have donethis in two ways: Jim Martin, Veel and others have reconstructed the meta-linguistic structure of history and science, inter alia, from textbooks – that is,from already recontextualized texts, which they take for representative tokensof the knowledge structure. Fran and Mary have less to work with, and havereconstructed the metalinguistic structure of English almost from scratch. Theirnext step, I think, is to show that it holds up when measured by some or otherequivalent representative token of the knowledge structure. This can’t be donethrough the current curriculum, because it is an invisible/competencepedagogy, hence radically under-stipulated. They might then try with some


250

older textbooks that do reference representative text types. Or from tokensfrom another national system; or even more powerfully, through a selection oftokens from a number of national curricular systems. All of this begs the largerquestion: what is the necessary and sufficient global archive of knowledgesystems?

This brings to light the fact that knowledges do not exist naturally in afinished representational form. The object we call ‘a knowledge’ is virtual. Theobjects we have that we take for tokens of the knowledge are the product ofvarious kinds of representational reconstruction. The SFL community hasbegun to construct metalinguistic representations, as we have seen. In logic,maths, some of the sciences, one finds propositional reconstructions – the orderthat one can see retrospectively (i.e. when doing pedagogic recontextualiz-ation) in terms of the necessary sequence of logical subsumption. There areprobably other kinds of representational reconstruction too. None of these cansufficiently capture the entirety of the knowledge. Metalinguistic reconstruc-tion; propositional reconstruction; pedagogical sequence reconstruction; thereare probably other kinds as well.

This digression sheds light on why the question of knowledge arises in thelight of the need for its pedagogic recontextualization. We can agree with Bern-stein, and with Karl, that the knowledge structure is not curriculum structure,but it is only the question of curriculum structure that brings to light that theremight be a knowledge structure worth bringing to light. For scholars at thecutting edge, the knowledge structure is old news not worth reporting on – untilthey put on their pedagogical caps and try to figure out how to teach the retro-spectively accumulated store, as in: ‘I know I stand on the shoulders of giants,but what exactly am I standing on?’ To date, it is the SFL community that havegiven the most systematic answer.

Fran: Joe is right when he says Mary and I (this volume) have shown that in theunfolding of the English curriculum, there has been a decisive shift from per-formance to competence pedagogy. I am grateful to him for sharpening the dis-cussion here, for we can at least say that the road is cleared to think about wherewe can go next (see my earlier comments about subject English above). I amexcited by Joe’s observation that ‘all knowledges are virtual’, and that they‘come to light only by conscious effort of the pedagogic recontextualizers’. It isindeed in coming to terms with issues of knowledge and curriculum structurein schools that we face not only the fact of recontextualization, but also – whatis more difficult – what it is that is recontextualized anyway. If we confront theissue, ‘what is the knowledge we seek to recontextualize for the purposes ofschool learning?’, we are driven – as this book reveals – to revisit quite a lot ofhistory (Muller, Moore, O’Halloran, Wignell, Christie and Macken-Horarik).Thus, echoing some of Joe’s remarks to me by email, in revisiting history we ‘payregard to the retrospect and the prospect: where the knowledge has come fromand where it is going’. This, Joe further states, ‘illuminates the communallearning path traversed which in turn is the guarantor for the individuallearning path that neophytes must traverse’. Does this – I ask myself – mean thatthe various models of English that Mary and I identify as all to some extent


jostling for position in the English curriculum – mean we have to (re)accom-modate them all in whatever model we adopt anew? I think not, but followingwhat Karl and Joe (joint chapter, this volume) argue, we should be able to estab-lish which past theories of subject English (if any) can be shown to share somecommensurable elements, making possible the articulation of a ‘more vertical’model than presently applies.

Karl: I’m glad Fran and Mary’s chapter has come up, as it’s a real step forwardto take an aspect of Bernstein’s discussion of knowledge structures (the way inwhich they progress through integration and subsumption of past knowledge orthrough segmental addition) and homologously transfer that to the study ofcurriculum and learning. This forms the basis of an insightful exploration of thedifferent principles that have been underlying the recontextualization ofknowledge into educational knowledge (to become subject English), and theconsequences of these different principles for the form taken by transmis-sion/acquisition in the subject area and the educational experiences oflearners. As Joe rightly points out, the competence mode of subject English isassociated with rendering invisible the principles underlying the form taken bythe curriculum structure, if not the educational knowledge itself. Instead of avisible structure of educational knowledge we get a series of ideal knowers asthe basis of recontextualization. In short, the educational knowledge structureis made less visible and the educational knower structure becomes the key.This knower code renders integrative and progressive learning of knowledgeproblematic.

As Fran and Joe discuss above, the key question is: what is to be done? Twothings stand out for me. First the basis of the hierarchy in the curriculum as itcurrently stands must be rendered visible. There are, after all, means of judgingsuccess or failure in the subject and thus a basis of hierarchy. To not highlightthis is to leave the knower code tacit, making it an extremely difficult positionto get to grips with or change – it is all the more powerful for not being broughtout of the shadows and into the light. Second, the alternative basis of a know-ledge code can be articulated and make visible. This is what I think Fran andMary are doing with SFL. As Joe puts it, they’re not doing it directly – thesubject’s powerful knower code makes this difficult to do in terms of the know-ledge structure. It’s an interesting move to do this via the metalanguage.

A question raised for me by all this is how the nature of English as a know-ledge structure (not subject English but rather its field of knowledge pro-duction) shapes the form taken by the humanities and social sciences. If, asKay O’Halloran shows so well in this volume, mathematics helps enable theintegrative and subsumptive nature of science, does English help enable thesegmental, proliferative nature of humanities and social science? Is Englishthe key subject area for enabling more epistemologically powerful know-ledge in other subjects that use it as the principal medium of expression? Ifso, then SFL has an extremely powerful role to play as an underlabourer,determining what is and is not the best means of articulating knowledgeclaims.


252

3.5 ValuesJim: From an SFL point of view the discussion of knowledge structure in Bern-stein’s work seems very ideational. But we know from work on appraisal in SFLthat the way in which concepts are valued by members of a discipline are everybit as important as what these concepts are and how they are arranged – includ-ing consideration of how research is appreciated, how researchers are judgedand how researchers and their apprentices are positioned emotionally. Karl’swork on knower structures and Rob Moore’s on judgements (this volume) ispart of this picture – the kind of identity researchers and learners construct forthemselves is an essential feature of disciplinarity. How might our discussion ofknowledge structure be expanded to bring a concern with the disciplinarity offeeling into the picture?

Fran: I think the very notion of pedagogic subject positioning following Bern-stein has consequences for a view of knower codes and knowledge codes ascarrying values. It is in my view a relatively easy extension of both theories hereto view all areas of the curriculum as involving the adoption of certain valuepositions. The study I did a few years ago (2002) of classroom discourse did notuse appraisal theory, though with hindsight I see that it could have done.

Joe: Bernstein didn’t speak directly about disciplinary identities, but in the lastvolume he does discuss both pedagogic identities and local (cultural) identities.As one might expect, for him, identities are the product of symbolic formationand they relate to the ‘social base’ within which the identity is embedded. It isnot clear to me how this might dovetail with appraisal theory or ideation moregenerally, since for Bernstein identities of either variety are not self-constructed.

Karl: The notion of knower structures brings into the picture something thatwas tacit but already there in Bernstein’s account of knowledge structures. Ishould emphasize that it’s not the sole basis or equivalent of talking aboutidentities. Basil showed how identity is shaped by different forms of knowledgestructures and curriculum structures in various ways. So identity is already a partof the picture. But whenever talking about those structures of knowledge or cur-riculum that render knowledge less visible, then it was not clear what the basisof identity is. ‘Knower structures’ simply brings out what was in the shadows.There is always both a knowledge structure and a knower structure, where thelatter simply refers to the ways in which knowing subjects are arranged by thediscipline. For example, with physics one could say that alongside its ‘explicit,coherent, systematically principled and hierarchical organisation of knowledge’(Bernstein 1996: 172) – its hierarchical knowledge structure – it is also charac-terized by a series of strongly bounded knowers, each with their own specializedmodes of being and acting. That is to say, physicists represent a series ofsegmented knowers in terms of their (non-scientific) ‘gaze’ – it matters less whoyou are in terms of your social background, for example; they can be verydifferent people, so long as they use the procedures of physics. Conversely,disciplines with horizontal knowledge or curriculum structures, such as that ofEnglish explored by Fran and Mary, may be characterized by a hierarchical


knower structure – the basis is who you are, your aptitudes, dispositions,attitudes, etc. based on your biological, social or cultural background.

If we don’t get to grips with the knower structure, and especially the role itcan play in humanities and social science subjects, then we’ll find it difficult tounderstand why the existing state of affairs exerts such a powerful grip overmany people. If they’re not focused primarily on the knowledge structure orthey render knowledge less visible, then what is it that they’re focusing on? Thisis important not only for being able to understand what it is that different dis-ciplines do, but also for reaching out to other disciplines, being able to see whatis underlying their ways of working and so be able to speak across languages. Itmay help explain why it is difficult to get some academics and teachers toengage in discussion of ideas as ideas. The hierarchical knower structure ofEnglish, for example, has been the basis of criticisms of canons (see Moore, thisvolume) of the ‘patriarchal’, ‘bourgeois’, etc. basis of knowledge claims. It’s alsowhat makes many approaches so appealing – they seem to speak to the experi-ences of people, and they’re viewed as warmer, more caring and sharing, moreobviously human than cold, inhuman technicality. I have been recently workingwith Sue Hood on using appraisal theory alongside these concepts and it’sbecoming clear to me just how well SFL can get to grips with these issues.

3.6 SFL and hierarchical knowledgeJoe: I have at times, perhaps mistakenly, discerned something of an apologetictone in SFL discussions of ‘abstraction’ and ‘technicalization’, as if these termswere somehow politically incorrect. Yet everything in the theory tells us that dis-cursive power is made available only by this route. The first part of the questionis I suppose this: does this apologetic-ness stem from the neo-Marxist belief thatabstraction and power are intrinsically tools of domination? Or is it rather alliedto a more post-modern spirit of de-differentiation, as in the post-modern DavidBohm’s objection to an analytic that ‘divides things into separate entities’ thusdistorting ‘undivided wholeness’, and his wish to denominalize scientificlanguage (‘which gives the basic role to the verb rather than the noun’)? All ofthis I take to be quoted approvingly by Halliday (p. 108) in Halliday and Martin(1993). It is true that there are signs of a lingering progressivism in early Bern-stein too (my students fight every year as to whether he is actually for integratedcodes and against collection codes). But this is by and large gone by mid-periodBernstein, and by late Bernstein his colours are clear. Is this embarrassmentabout dealing with the necessary instruments of power something we shouldtalk about?

Fran: I don’t see that Joe’s interpretation is justified. This discussion is veryHallidayan in character, in that, as often, he is interested in the relationshipbetween language and the social construction of experience and the claimssometimes made that language can’t always cope with the pressures of change,and of the meanings people want to make. It all points to another issue: is thereanother reality that language somehow only partly captures? I don’t findanything particularly ‘approving’ about the ways Halliday quotes Bohm. In fact,he writes of Bohm and others that they may have overlooked one thing: namely


254

that ‘you do not need to keep engineering a language in order to change it; itwill change anyway – because that is the only way it can persist’ (Halliday, ibid.).

Jim: I take Halliday’s point as being that whereas language offers both morenominal (attic) and more verbal (doric) construals of experience, science hasby and large adopted the attic mode, and has become, now and again, self-conscious about the limitations of this bias. Halliday is suggesting, contra Bohm,that it’s not language which is at fault, but the language of science, and thatlanguage has the resources to serve the doric mode if scientists wish to takethem up. This of course begs the question of whether science in the doric modewould recognize itself as science. I suspect not, since it is not at all clear how ahierarchical knowledge structure can be developed verbally, without theresources of grammatical metaphor developed in the attic mode.

I hope I have never expressed a distaste for abstraction and technicality in myown writing, since I have always seen myself as mounting a defence of such inthe face of progressivism, constructivism, the plain English movement and soon. As Maton points out (this volume) hierarchical knowledge structures can bea great deal more democratic that horizontal ones, since what you know isprivileged over who you are. I have always found the non-technical discourses ofthe various languages in horizontal knowledge structures far more elitist andinsidiously dominant than discourses featuring verticality.

Joe: A follow-on to Jim: Halliday in the chapter just quoted goes on to speculatethat Bohm might be heralding ‘a new type of order’ where ‘communication’becomes of the essence – and perhaps linguistics can then become the newqueen of the sciences (‘Their coat-tailing days are over’). There is a similar kindof aspiration expressed at the end of Jim’s chapter. Do your ambitions forlinguistics rest on a similar foundation to Halliday’s, Jim? That is, do you thinkthat the future will be a de-differentiating one? Or do you think SFL will have tostiffen its spine somewhat (become a hierarchical knowledge structure) beforeit can challenge for pole position?

Fran: SFL theory is often heard as being very imperialistic in that it is said towant to overtake everything else in its claims to offer ultimate explanationsand/or ultimate accounts of experience. Well, that is a risk I suppose, but itmisses an important point. Halliday and Bernstein were bold and ambitious intheir scholarly visions, for both wanted to develop tools with which to explainthe nature of social life. Both might well be accused of being arrogant becauseof the boldness of their visions – they did want to do nothing less than explainsocial experience! I think the very adventurousness of what they both aspired todo is part of their appeal, but also part of the intellectual power they bothunleashed in the world. Theirs was a noble endeavour, and one that remainsrelevant in the twenty-first century.

Jim: The basic propaganda here as far as SFL is concerned is to argue thatlinguistics should not be modelling itself as the science of language, sincescience builds knowledge structures responsible for describing physical and


biological systems – and social semiotic systems are more complicated thanthese, having evolved out of them (their emergent complexity is of a differentorder). This means that the theories we build for social semoisis will have to bedifferent in kind from those we’ve built before, and we can expect them to havemore than enough power to interpret metalinguistically and metasemioticallywhat science has been saying about physical and biological systems all along. Ifthis metasemioic purchase makes linguistics the queen of sciences, so be it, longmay she reign. But let me stress that this purchase will not make her a scientist,because her object of inquiry is discourse, not material reality per se.

I don’t think linguistics has to stiffen her spine and become more of a hier-archical knowledge structure to achieve this, because becoming more likescience as we know it is not going to do the trick. As I said above, we need prismsnot pyramids, and so we need a theory that embraces complementarity along-side hierarchy. If complementarity as I presented it above can be construed as adimension of verticality, as Joe suggests, then, OK, yes, linguistics does need tostiffen its spine. Currently, the main technological problem standing in the wayis that we can’t process data fast enough and deeply enough to foster the gram-maticality we need to move on. In short, our computers can’t do discourseanalysis; they can barely do the kind of functional grammatical analysis we needin fact. So we can’t see very well what we are looking at. This should change inanother generation or so and then we’ll begin to see which languages oflinguistics have the right stuff.

Fran: I have not much to add here, other than to say that we are getting betterat seeing what we are looking at. The next generation will no doubt do betterstill, but we are on the way.

Karl: One thing I’d like to add is to echo Fran’s admiration for the boldness ofthe enterprise of SFL and Bernstein’s sociology. In the contemporary climate ofsocial science such boldness is easily denigrated as arrogance or imperialism.F.R. Leavis (someone often accused of all kinds of arrogance, not alwaysmistakenly) once said that the only way to escape misrepresentation is never tocommit oneself to any critical judgement that makes an impact – that is, neversay anything. What began as a much needed dose of modesty became a form ofintellectual cowardice. To always caveat everything as only being from our ownknower perspective, as if we should only share our own individual story, leavesus where we began in terms of understanding the world. This has become an alltoo common strategy in at least my own disciplines of Education and Sociology.The image that always springs to my mind is of a tower of knowledge. The aspir-ations of SFL, it seems to me, are akin to a grand undertaking such as buildinga cathedral, involving many thousands of workers, taking many decades, wherethe original architect and most of those who laboured will not live to see itscompletion. In contrast, many other approaches are akin to the relatively rapidcreation of suburban housing in a new development, each building beingconstructed quickly, requiring near-constant maintenance, with new membersof the workforce moving rapidly to begin building the next house. Instead of acathedral or tower, building upwards and aiming to last, this represents the


256

spread horizontally across the intellectual landscape of comparatively low-level,largely identical buildings. Hierarchical knowledge structures are intellectualcreation as the building of towers; horizontal knowledge structures are intellec-tual creation as suburban sprawl. As John Ruskin put it, when we build let usthink that we build for ever. One concern I would have though, concerningunderstanding knowledge and curriculum structures, is the possibility oflinguistic reductionism, that is seeing these objects solely as comprising theirlinguistic features. Knowledge is more than the sum of its linguistic parts.Exactly how, and what form the interface between our approaches will take, areexciting issues to explore.

3.7 The frontiers of knowledgeJim: One final question I’d like to pose has to do with discourses that don’t getmentioned in our discussions of knowledge structure. In our early 1990s workon uncommon sense, one such discourse we tried to tackle was the discourse ofadministration. This is full of grammatical metaphor, but in the service of pro-posals (directives, rules and regulations) rather than propositions (statementsabout the world). Are we suggesting that a procedural discourse of this kind isnot a knowledge structure?

Another discourse we considered was the discourse of technology, whichbridges between science and control of our material world. This turned out tobe more like horizontal discourse than I expected, although knowledge inscience industry clearly bridges across the horizontal and vertical discoursedivide. What do we do about bridging discourses of this kind?

Perhaps my general point here has to do with what we do about discoursesevolving in the service of other discourses. Kay O’Halloran (this volume) showsthe crucial role played by mathematics for the grammaticality of science. Thisinterplay in turn engenders the technology that manages material resources.Reasoning along similar lines, we might see the language of administration as atechnology that manages people, as informed by humanities and social science.

In short then, what qualifies a discourse as our object of inquiry? Why areeveryday conversation and traditional disciplinary knowledge in, and things likethe discourse of technology and bureaucracy out?

Fran: The answer to this is surely social: it is a matter of what gains recognitionand respect. I suspect that many of the discourses that evolve in the service ofother discourses are often the most invisible (and like women’s work, often leastrespected!). Their very invisibility is a source of their strength. In Bernstein’ssense of the word ‘pedagogy’, the discourses involved in the service of scienceor of administration are profoundly important as sites for pedagogic practice.SFL theorists might well re-open this whole matter in the future.

Joe: Reading through Jim’s comment above, it struck me that nowhere in thebook do we discuss a key distinction Bernstein makes between knowledgesingulars and knowledge regions. Our entire discussion has been assuming thatall knowledges take the form of singulars, when they evidently don’t. A regionis an ensemble of singulars combined sometimes with segments of everyday or


procedural knowledge. Architecture is one example of a region. The varioustechnical or technological discourses would be the same. Public administration,as a subject taught at university is also a region. What then is the discourse ofadministration? Bernstein’s answer would be to ask: is there integration at thelevel of meaning – in which case it is a vertical discourse – or is integrationsegmental? I would say that it is probably segmental, which makes it horizontaldiscourse. However, it borrows a number of features from its recontextualizedparent public administration, so that it exhibits features of a region. And so on.In other words, either we can deal with administration, and all types of discoursefor that matter, with Bernstein’s tools, or the theory is lacking. There can’t be aquestion of leaving out this or that class of discourse.

What the question highlights for me is that although we may say: if no gram-matical metaphor then no verticality, we can’t necessarily say: if grammaticalmetaphor then it must be a vertical discourse.

Karl: As Fran and Joe suggest, I don’t think the focus of our discussions has beento negate other discourses, such as administration and technology, as not beingworthy of study. But Jim is right to highlight that these must be brought into thescope of the discussion. As the concepts are applied to such discourses they willundoubtedly be pushed to their limits, necessitating theoretical development,and raising further questions for us all to consider. The dialogue between thetheoretical and the empirical will continue, as I hope will that between systemicsand sociology. This, we could say, is not the beginning of the end, but simply theend of a new beginning.

References

Bernstein, B. (2000), Pedagogy, Symbolic Control and Identity (rev. edn). Lanham andOxford: Rowman & Littlefield.

Gamble, J. (2004), ‘Tacit Knowledge in Craft Pedagogy: A Sociological Analysis’. Unpub-lished PhD thesis, University of Cape Town, Cape Town.

Halliday, M.A.K. (1993), ‘Language in a Changing World’. Occasional Paper 13, AppliedLinguistics Association of Australia.

Halliday, M.A.K. and Martin, J.R. (1993), Writing Science: Literacy and Discursive Power(Critical Perspectives on Literacy and Education). London: Falmer.



Page numbers in italics refer tofigures and tables.

absolutism 115, 119, 120abstractions

embedded clauses in nominalgroups 171, 176–7, 178

(proto-)grammatical metaphorand 145–8

and technicality 190–200 passim,202

activity sequences 34–6, 37, 41, 43–5Agent and Medium 54–5All-or-Nothing Fallacy 114, 122appraisal 178arts 112–13

high and low 114–15music 98–104sculpture and photography 219see also humanities; judgement(s)

Basic Skills, English teaching model159–60, 161, 162, 168–9

Bell, E.T. 210, 213, 214Berlin, I. 73–4Bernstein, B. 3–8passi, 14–29, 36, 38,

47–80 passim, 87–96 passim,104, 109, 110, 112, 126–7,131–9 passim, 157–67 passim,184–5, 201–14 passim, 239–44passim, 252

Bloom, H. 125, 134Bourdieu, P. 91, 111Brookes, B.C. 95

canons, critical approaches 110–12Carey, J. 112–21, 122, 123, 124–6Cartesian rationalism (Descartes)

73–4, 113–14, 216, 226, 229‘cause in the clause’ 45–7, 48childhood see language learning in

early childhoodChristie, F. 17, 56, 159, 160, 162,

239–57 passimand Martin, J. 34

class see social classClassics 91classification 17, 36, 38–40

see also taxonomy of valuecoalitions of the mind 121–4cochlear ear implant, example of

composition 40code(s) 6–7, 16–18

collection and integrated 96–7elaborate 6, 16, 22, 131–2élite 98–9, 100, 102–4integrating 8knowledge–knower 27, 96–8legitimation, of specialization

93–5, 96–8, 99, 100, 101orientation to meaning 16, 18, 20relativist 98restricted 6, 16, 22

codification of disciplines 69–70Collins, R. 65, 68, 109, 112, 121–6,

127common and uncommon sense

fields 37, 38, 66, 68–9composition 36, 37, 40, 42

Index

260

computer technology 224, 225–6,229–31

context 5decontextualization 122, 123language learning in early

childhood 139, 140–3recontextualizing rules 7, 19, 22,

28–9, 162–3social 35

Cultural Heritage, English teachingmodel 160–1, 162, 165, 166,167–70, 174, 177–8

Cultural Studies, English teachingmodel 164–5, 167, 168–9

curriculum 8England

English 156, 170–4music 98–104

NSW, Australia, English 156, 165,175–8

planning 72South Africa 79–82structures and knowledge

structures 28–9, 80–2

Davis, P. 211, 231et al. 230

decontextualization 122, 123Descartes, R. (Cartesian

rationalism) 73–4, 113–14,216, 226, 229

dialects, social 6dialogue

development 5–9interlocutory interpolation 121language learning in early

childhood 132–5‘situation-transcending’ 125

discursive saturation 69–70distributed social goods, knowledge

as 66distributive rules 7, 19, 162–3

division of labour, mental andmanual 19

Donne, John 74–5, 117Durkheim, E. 15–16, 19, 22, 28,

67–8, 75, 109, 112, 117–18,127, 202

economics see political economy;value (economic theory)

educational knowledge–knowerstructures 96–7

elaborate codes 6, 16, 22, 131–2elaboration 172, 176–7, 178élite code 98–9, 100, 102–4Elkins, J. 226elliptical clauses 171, 172embedding 56–7, 171, 178, 228empathy 167–74 passim, 175–8 passimEnglish

interpretation/symbolic readingof literary text 175–8

‘literary’ narrative 167–74, 178metalanguage of systemic

functional linguistics (SFL)179–80

and science, pupil/universityentrant perceptions 102

teaching models 158–67, 168–9see also literature

epistemic relation (ER) toknowledge structures 93–4,96–8, 101

ethical judgement 177–8ethical theory and aesthetics 120–1evaluative rules 7, 19, 162–3events and objects, joint attention in

early childhood 133external language of description 71

field(s) 34–8, 55–6, 111common and uncommon sense

37, 38, 66, 68–9

INDEX

production 88–96reproduction 96–104

framing 17Functional Language Studies,

English teaching models163–4, 166, 168–9, 174, 179

functional linguistics see systemicfunctional linguistics (SFL)

Gellner, E. 91, 95genealogical organization of living

things 38, 39generic reference, language

learning in childhood 142–3genre 9, 55–6, 57Golovin, A.A. et al. 222, 225grammar(s) 5, 26–7

matching 52–3modelling 136weak and strong 71, 78, 209–10,

211, 221–4grammatical metaphor 49–55, 58–9,

60–1, 173(proto-)grammatical metaphor

and abstraction 145–8and vertical discourse 53–61

passim, 243–4grammaticality

and knowledge structures 209–14of mathematical symbolism 225–9multimodal approach 214–15,

216–29and verticality 71–2, 244–6

habitus 120, 126Hacking, I. 68Haire, M. et al. 38, 39, 40Halliday, M.A.K. 3, 4, 5–6, 9, 17, 47,

51, 53, 54, 60, 61, 132, 133,145, 147, 163, 164, 172, 173,186, 187, 193, 205–9, 214, 226,239, 240, 247, 253–4

and Hasan, R. 9McIntosh, A. and Strevens, P. 9,

163and Martin, J. 226–7, 253and Matthiessen, C. 214, 226

Hasan, R. 3, 6–7, 17, 66Halliday, M.A.K. and 9

hierarchical knower structures 91,92–3

hierarchical knowledge structures 8,23–4, 70–1

political economy 186, 189,190–200

science 38–41, 90social sciences 189, 201–3sociology 202–3and systemic functional linguistics

(SFL) 253–6triangle metaphor 60, 90, 91, 92,

184, 206–7vs horizontal knowledge

structures 25–8, 109–10hierarchy, in discourse and society

65–70high and low art 114–15history

Australian Aboriginal people 45–7Australian population and

immigration 42–5, 55–6genre 57, 58, 59horizontal knowledge structure

42–9Philippines, Spanish colonization

47–9of science 73–8, 186, 187–9, 193of social sciences 186–203

Hobbes, Thomas 186, 187–8, 189,201

Hoggart, R. 111–12, 115horizontal discourse

definitions of 7–8, 22–3, 54, 66–7language learning in childhood

261INDEX

262

131, 132, 135–9science and humanities 95see also vertical and horizontal

discoursehorizontal judgements 112–17horizontal knower structures 92horizontal knowledge structures

8, 23–4, 25–6, 70–1English 167history 42–9humanities 90sociology 202–3vs hierarchical knowledge

structures 25–8, 109–10humanities

hierarchical knower structure 91

horizontal knowledge structure 90knower–knowledge structure 93–4and science

and social science 247–8‘two cultures’ debate 88–95,111–12, 119

idealization 67–8ideational metafunctions of

language 5identity see knower code; knower

structuresimaginative nexus 125–6information exchange, in early

childhood 136–7intellectuals 122–3internal language of description 71,

72, 78interpersonal metafunctions of

language 5

judgement(s)critical approaches 110–12ethical 177–8horizontal 112–17

literary 116–17, 118–21, 123,124–6

preferences and 114–17, 120–1,126–7

sociality of 112, 124–6vertical 117–24

Kant, E. 113, 118Kepler, Johannes 75, 76knower code 27, 96–8knower structures 91–3

see also knowledge–knowerstructures

knowledge 14–16, 29–30and dilemma of progress 70–2ethical theory and aesthetics

120–1making visible 21–2, 24–5in school 34

knowledge about language (KAL)57, 58–9, 163

knowledge frontiers 256–7knowledge progression 73–7,

212–14knowledge structures

and curriculum structures 28–9,80–2

differences between 25–8discourses and 22–5future research directions 239–57linguistics as 56–60and pedagogic structures 79–83science and humanities cultures

89–90knowledge theory development

14–16knowledge–knower structures

educational 96–7in production fields 88–96in reproduction fields 96–104science and humanities 93–6

knowledge code 27, 96–8

INDEX

language learning in earlychildhood

after three years old 140–53centrality of dialogue 132–5first two and a half years 135–9horizontal discourse 131, 132,

135–9lavatories, grammatical metaphor

49–51Leavis, F.R. 94–5, 160–1legitimation codes of specialization

93–5, 96–8, 99, 100, 101lexical metaphor 173, 176lexical motifs and technicality 190–3‘literary’ narrative in English

167–74, 178literature

judgement 116–17, 118–21, 123,124–6

teaching models 160–1, 165logical positivism 71, 77, 82McEwan, Ian 125Macken-Horarik, M. 164, 167, 174,

175, 244, 250–1magnification 60Mares, P. 43, 55–6Martin, J.R. 9, 34, 37, 38, 55, 66–7,

68, 70, 164, 214, 239–57 passimChristie, F. and 34Halliday, M.A.K. and 226–7, 253and Rose, D. 34, 164, 173, 174,

214and Veel, R. 226–7and White, P.R. 9, 55

Marx, Karl (Marxism)evolution of social sciences 186,

187, 189, 195–200, 201–2reductionism 111vs modernist history 47–8, 49

mathematical symbolism 215, 216,219–20, 221–4, 229–30

grammaticality of 225–9

mathematicsknowledge structure 80–2and science 210–14

computer technology 214–15,216–29, 231–2knowledge and knowledgestructures 206–9knowledge structures andgrammaticality 209–14SF-MDA (multimodalgrammatical) approach214–15, 216–29, 231–2

Maton, K. 8, 88, 91, 110, 239–57passim

Moore, R. and 4, 30, 74, 88, 93,110, 122

meaningknowledge as 66orientation to 16, 18, 20

Medium, Agent and 54–5Merton, R. 30, 69, 70, 75–6metafunctions of language/SFL 5,

35, 55metalanguage

development in early childhood142

systemic functional linguistics(SFL) as 248–51English teaching 179–80

metaphor 172–3, 178triangle 60, 90, 91, 92, 184, 206–7

modelling grammar, earlychildhood 136

Moore, R. 87, 118, 122and Maton, K. 4, 30, 74, 88, 93,

110, 122and Muller, J. 8, 77–8, 87

Morais, A. et al. 65Morris, M. 89Muller, J. 30, 55, 82, 158, 167, 207,

231, 239–57 passimMoore, R. and 8, 77–8, 87

263INDEX

264

multiliteracies, English teachingmodel 166, 167

multinocular vision 60Murray, L. 132–3music 98–104

New Critics, English teaching model161

‘New Cynics’ 77–8, 82New Literacy Studies (NLS), English

teaching model 166–7, 168–9Newton, Isaac 75–6, 186, 187,

216–22 passim, 223, 228nominalization/nominal groups

52–3, 171, 172, 176–7, 178, 192non-finite clauses 171–2

objects and events, joint attention inearly childhood 133

observational understandings of text143–5

O’Halloran, K.L. 214, 217, 222–3,227, 229–30, 231–2

operative processes/participantconfigurations 227–9

oral and written transmissions 68–9orientation to meaning 16, 18, 20

Painter, C. 67, 68, 133parochialism 123participant configurations/operative

processes 227–9pedagogic device 7, 18, 19–21, 22,

160pedagogic discourse 7, 21, 22, 56–7,

162–3pedagogic and knowledge structures

79–83pedagogies to knowledges 14, 22Personal Growth, English teaching

model 161–3, 164, 165, 167,168–9

philosophy 122–3photography, sculpture and 219political economy 186, 187, 189,

190–200Popper, Karl 24, 26–7, 77pre-linguistic communication 132–3preferences and judgements

114–17, 120–1, 126–7Prigent, A. et al. 222, 224production fields,

knowledge–knower structuresin 88–96

profane and sacred worlds 67–8

progression, knowledge 73–7,212–14

projection relations in pedagogicdiscourse 56–7

(proto-)grammatical metaphor andabstraction 145–8

‘proto-language’ phase of languagedevelopment 133–4

realism see sciencerecontextualizing rules 7, 19, 22,

28–9, 162–3reductionism 111register (variation) theory 5, 9relationalism 111relations between social groups 115relativism 111, 112, 114–15, 116,

121–2local 123–4subjective 118

relativist code 98reproduction fields,

knowledge–knower structuresin 96–104

researchadministration 72code theory 17–18future directions 239–57

INDEX

‘resemioticization’ 207–9restricted codes 6, 16, 22Ricardo, David 189, 193–5, 199, 200,

201Rodin, Auguste 219Rose, D. 58–9, 68, 83

Martin, J.R. and 34, 164, 173, 174,214

sacred and profane worlds 67–8science

biology 38–40codification 69and English, pupil/university

entrant perceptions 102geology 34–6, 37hierarchical knowledge structure

38–41history of 73–8, 186, 187–9,

193horizontal discourse 95and humanities

and social science 247–8‘two cultures’ debate 88–95,111–12, 119

knowledge progression 73–7,212–14

and mathematics see mathematics,and science

meterology 41, 42, 47post-modernist account 113

sculpture and photography 219‘segmental organization’ of

horizontal discourse 138–9self-contextualisation, language

learning in childhood 141–3semiotic(s)

‘resemioticization’ 207–9resources, mathematical and

science 215social 56–7, 60

Shapin, S. 74, 75

skills instruction, early childhood135–6

Smith, Adam 187, 188–9, 190–2,193, 199, 200, 201–2

Snow, C.P. 88–9, 90, 91–2, 93, 94, 95‘social base’ 15–16social class 3, 6–7, 16, 17–18

and science 91–2, 188, 231–2social context as register and genre

35social dialects 6social phenomenon of language 5–6social relation (SR) to knower

structures 93–4, 96–8, 101social sciences

historic evolution of 186–203relativism in 114–15science and humanities 247–8

social semiotics 56–7, 60sociality

of judgement 112, 124–6of knowledge 111, 118, 120, 122

sociologyof education, Bernsteinian

perspectives 3–8evolution of 201–3

spatial/temporal relationship217–29 passim

specialization 96–8legitimation codes of 93–5, 96–8,

99, 100, 101specialized lexis 41subjective relativism 118subjectivism/inter-subjectivism 115,

116–17, 121, 124–5Sydney School 34, 38, 56‘symbolic control’ 7, 15–16symbolism, mathematical see

mathematical symbolismsystematic knowledge building,

language learning inchildhood 148–53

265INDEX

266

systemic functional linguistics (SFL)3, 8–9

dialogue development 5–9English 179–80future research directions 239–57and hierarchical knowledge 253–6and knowledge construction

34–64as metalanguage 248–51

English teaching 179–80scales and complementarities 60SF-MDA (multimodal

grammatical) approach,mathematics and science214–15, 216–29, 231–2

and verticality 246–7systemic grammar 5

tacit acquisition of values 157tacit learning in childhood 136–8,

139tastes see preferences and

judgementstaxonomy of value 190–5technical lexis 41technicality

and abstraction 190–200 passim,202

history 42, 43, 45, 48–9linguistics 58see also science

textinterpretation/symbolic reading

175–8observational understandings

143–5and oral transmissions 68–9vs lived experience 143–4

textual metafunctions of language 5thematic non-finite clauses 171–2‘thingification’ 44, 45, 53time

activity sequences 34–6, 37, 41,43–5

spatial/temporal relationship217–29 passim

Token and Value 53–4, 199–200triangle metaphor 60, 90, 91, 92,

184, 206–7truth 118, 119–20, 121–2‘two cultures’ debate, science and

humanities 88–95, 111–12, 119

value (economic theory)commodities and 194–200and labour 191–2, 193–4, 196–8,

199taxonomy of 190–5

Value, Token and 53–4, 199–200values

aesthetic 111SFL 252–3tacit acquisition of 157

vertical discourse/verticalitydefinitions of 7–8, 22–3, 53–4,

70–2in English 157and grammatical metaphor 53–61

passim, 243–4and grammaticality 71–2, 244–6judgements 117–24kinds of 77–8and systemic functional linguistics

(SFL) 246–7vertical and horizontal discourse

borders of 241–2city landscape visualization 207–9definitions of 7–8, 22–3, 53–4,

66–7and knowledge structures 22–5maths and science 206see also horizontal discourse;

vertical discourse/verticalityVirilio, P. 219, 225, 230–1, 232

INDEX

visibilityconcept of 215of knowledge 21–2, 24–5see also mathematical symbolism;

visual imageryvisual imagery

city landscape 207–9computer technology 224, 225–6,

229–31

maths and science 216–20Vivo, Giambattista 73–4

weak and strong grammars 71, 78,209–10, 211, 221–4

weak and strong progression 73–4Wignell, P. 185, 202

267INDEX

Documents

[Frances Christie, J. R. Martin] Language, Knowled(Bookos.org)