Calise Lowi- Hyperpolitics

http://www.jstor.org

Hyperpolitics: Hypertext, Concepts and Theory-MakingAuthor(s): Mauro Calise and Theodore J. LowiSource: International Political Science Review / Revue internationale de science politique, Vol.21, No. 3, CyberPolitics in International Relations. CyberPolitique et relations internationales,(Jul., 2000), pp. 283-310Published by: Sage Publications, Ltd.Stable URL: http://www.jstor.org/stable/1601237Accessed: 30/05/2008 05:05

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at

http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless

you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you

may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at

http://www.jstor.org/action/showPublisher?publisherCode=sageltd.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed

page of such transmission.

JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We enable the

scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that

promotes the discovery and use of these resources. For more information about JSTOR, please contact [email protected].

http://www.jstor.org/stable/1601237?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp

http://www.jstor.org/action/showPublisher?publisherCode=sageltd

International Political Science Review (2000), Vol 21, No. 3, 283-310

Hyperpolitics: Hypertext, Concepts and Theory-making

MAURO CALISE AND THEODORE J. Lowi

ABSTRACT. Political theory begins with respect for the complexity of its conceptual universe. Hypertext, a language for organizing knowledge on a modular rather than a sequential basis, provides an opportunity to unscramble this verbal mixture and to reconstitute it with new understandings and a new source of creative political discourse. This article reports on an experiment in computer-assisted advancement of political theory. It presents the initial results of a hypertextual approach to the conceptual basis of the literature by: (1) selecting the concepts most frequently employed in a sample of political science dictionaries; (2) linking them to a number of textual and bibliographical sources for in-depth analysis; and (3) developing multi-dimensional linkages by placing each concept in a matrix of logically consistent relationships. Thanks to the Web, Hypertext opens the door to unlimited navigation through wider and wider domains. The Hyperpolitics platform is an attempt to make this compatible with the logical structure and informed guidance of higher learning.

Key words: Hyperpolitics model * Hypertext * Political concepts * Political science dictionaries

Introduction

From Hypertext to Hyperbooks

Although, thanks to the World Wide Web, hypertext has become the standard

language of modern communication, few successes have yet been recorded in the

adaptation of this language to higher learning. Hypertext opens the door to unlimited navigation through wider and wider domains, while higher learning calls for logical structure and informed guidance. But hypertext has not yet been

consistently developed into a means suitable for scientific inquiry and

0192-5121 (2000/07) 21:3, 283-310; 013086 ? 2000 International Political Science Association SAGE Publications (London, Thousand Oaks, CA and New Delhi)

International Political Science Review 21(3)

dissemination. Text eventually evolved into the modern bound book; hypertext is still in search of a format.

CD-ROMs have gone some way toward bridging the gap between excess of information and consistent hypertextual linkages. Yet this has usually been done on an idiosyncratic basis. CD-ROMs greatly differ according to the nature of their sources and objectives. Rather than stressing the logical aspects, they tend to emphasize entertainment or information through mere multimedia and retrieval capacities. Despite the wealth of materials, only a tiny fraction of the power of hypertext has been tapped; for purposes of education and learning, the road is wide open to develop a more consistent hyperbook format.

The Hyperpolitics project is an effort to develop a hyperbook prototype for political science-and social science at large. It consists of an analytical system to take full advantage of the wealth of scholarly information now available in the Web environment by constraining hypertextual links into a logical model compatible with scientific discourse.

Hyperpolitics: An Overview

As any experimental project, Hyperpolitics can be presented through its parts and its aims; or it can be used-and tested-as a working software prototype. In this article, we will necessarily confine ourselves to the scientific rationale of our project, which leaves enough room for evaluation and criticism. However, we welcome all requests to share the practical use of Hyperpolitics from all who wish to contribute to-and improve-our enterprise. We have learned from our own experience that biting into the Hyperpolitics cake remains the best way to actually move along a few bites further.

Hyperpolitics consists of three basic components: (1) selection of keywords: the computerized selection of roughly a hundred keywords, based upon the scanning of ten political science dictionaries from four languages and five countries; (2) selection of textual and bibliographical sources for keyword in-depth analysis: the selection of three textual reference sources to link to appropriate keywords- Definitions, Books, and Authors. Definitions consist of English abstracts of the definition of each keyword in the various dictionaries (see examples below). Books consist of four thousand reviews published in the American Political Science Review of titles containing the keywords. Authors are represented by biographies of prominent living political scientists, linking their major works to the keywords; (3) construction of a logical model for systemic keyword cross-tabulation: each keyword is placed at the center of a matrix of twelve related keywords, in a dialectic relationship (see below). This produces a closed logical system of analysis. The matrix also provides a tool for systemic navigation in our database and the Web at large.

As a hyperbook prototype, Hyperpolitics can be thus defined as a system for keywords relational navigation, allowing for in-depth analysis of keywords through comparison of related (con) texts and systemic cross-tabulation of related keywords.

However, with all due respect for new methodology and technology, our main effort has been to improve on our own capacity as political scientists. The fundamental unit of hypertext language is keywords, which in theoretical parlance can be referred to as concepts. The keywords of any field of science capture living thought; the keywords in politics capture most, if not all, of the range of experience with which political theory must deal. And when strung together with

284

CALISE/LOWI: Hyperpolitics

reason, these keywords can produce insight and thought that in many ways are superior to experience itself. We are not here preparing ourselves or our readers for a disquisition on language; we are trying to prepare a language for a disquisition on politics.

Theory-making in the Web Environment

Language serves thought. But different kinds of thought require different kinds of language, and different kinds of language influence the manner of thought. There has been an explosion of language in recent decades in political science, mostly as a consequence of its expansion on a world scale. A common reaction to this explosion, and the complexity following from it, has been single-mindedly to specialize. This involves specialization of sub-fields and methodologies, culmi- nating in the specialization of concepts.

However, as we have advanced in our ability to define specific concepts to gain more clarity in meaning and context, we have weakened the capacity of the field at large to engage in theory-making. Defining for clarity for one purpose (the limiting case being the operational definition) can block or seriously reduce the applicability of a concept-and even the larger thought arising out of it-to other analyses and to other contexts and theories. It is undeniable that the meaning of any concept depends not only on the defining words and sentences themselves but also on the choice of context. Designative words are one thing; concepts essential for a field of thought are quite another matter: they are never context-free. But recognition or construction of context should not be done seriatim, one concept at a time. The job necessarily begins one concept at a time, but it has to move quickly and self-consciously from discrete concepts to relationships between and among concepts. The essence of theory-making is the control of contexts by placing each one in relation to all the other concepts deemed necessary for the endeavor, and hypertext offers unprecedented help in this difficult task.

Hypertext is a language for organizing knowledge on a modular rather than a sequential basis. While in a conventional exposition concepts are associated through linear steps, across time and space, hypertext creates multi-level and multi-directional links between a word and its meaning: each term can be associated with (and described through) a number of related contexts. At the dawn of the computer era one of the leading computer scientists was asked his opinion of what was the most important thing we can do in social science that was virtually, if not actually, impossible to do before computers. Without hesitation he said, "factor analysis." We believe hypertext can now lead us a step further. For the power of hypertext is not merely one of calculations, pulling together an extraordinary amount of information on each individual concept; it brings an interactive and dynamic quality to theory-making.

However, a caveat is in order here. The recent spread of the World Wide Web on the Internet has greatly enhanced the possibility of surfing from one term-or site-to another through a hypertext-like interface. Yet, the philosophy of the Net is a bottom-up approach: making available for the user as many sources as possible, but with no closure. Unlimited-and unguided-access to a massive number of sources is equivalent to hardly any access at all; an excess of freedom can be a one-way ticket to chaos. This has proved to be the case for the most enthusiastic Web users. The search for systems and criteria for selecting and accessing relevant information for different users dominates the present stage of

285


Web explosion. Search engines and portals have become the gatekeepers to the Net's ever-expanding universe.

This becomes all the more the case if hypertext is to meet the challenges of theory-building. To this end, hypertext access needs normative choice and responsibility to become a course of action, a set of guiding criteria for conceptual navigation. As President Truman said of General Eisenhower, "He was a fine leader, as long as he had a boss": the bottom-up components of hypertext needs governing mechanisms to give it purpose. The governing mechanism we propose is the system of Hyperpolitics.

Concepts and Contexts

Stage 1: Choosing Keywords The quality of a science can be known not only by the number of its concepts but by its choice of concepts, by the names it gives those concepts, and by the appropriateness and stability of the properties to which each concept refers. Political science has gone part of the way, producing innumerable concepts, but falls back among the disciplines by its negligence regarding choice of referent and boundaries between and among related concepts. This becomes all the more evident as one turns to the dictionaries as the natural source for the basic political science keywords. We reviewed and analyzed most of the recently published and easily accessible political science dictionaries, and quickly found ourselves suffering from the lexicographer's curse: too many words, definitions either too long or too brief, and, above all, negligence and downright indifference toward the criteria for choosing words and how to define them, what words to include or exclude, what limits to set on words' boundaries-in short, guidelines on what concepts should contain and what concepts should exclude. Even the encyclopedic dictionaries, with definitions of concepts reaching 6 000-10 000 words, followed no strict or discernible plan and were more often padded with bibliographic references and historical allusions than by efforts to specify relevant properties to be included and irrelevant properties to be excluded.

In the end, we chose those dictionaries whose definitions were closest to the mid-point between very brief and very lengthy. However, since that still left us with an unwieldy number of dictionaries, we narrowed our choices according to our judgment of the status and distinction of the editors within the field of political science in their own countries as well as elsewhere. Although we regret that our limited resources prevented us from including all dictionaries, our program is an open invitation to add dictionaries, keywords, and all manner of cross-tabulation.

The more difficult task was choosing the actual words that would become the keywords in the Hyperpolitics system. Computer inputting and indexing produced an initial word count. Frequency distributions, as shown in Table 1, gave a spread that led to a preliminary list of 100 keywords by taking all the words that appeared as entries in at least six dictionaries. As a result, our list is a deliberate, quantified representation of the discipline in the five oldest and most advanced homes of political science.

To be sure, an arbitrary numeric threshold had to be chosen, as is the case with every dictionary. Our project's limit of roughly one hundred entries stemmed from two criteria. First, we wanted to confine our overall vocabulary to a number

286


TABLE 1. Frequency Distribution of Entries per Selected Dictionary.

Dictionaries All Inputted Average Entries Entries Entries length occurring

(words) in at least 6 dictionaries

Dizionario di Politica, Norberto Bobbio, Nicola Matteucci, Gianfranco Pasquino (eds), Torino, TEA, 1990

The Blackwell Encyclopedia of Political Science, Vernon Bogdanor (ed), Oxford, Blackwell, 1991

Dictionnaire Constitutionnel, Olivier Duhamel and Yves Meny (eds), Paris, Presses Universitaires de France, 1992

Dictionnaire de la Science Politique et des institutions politiques, Guy Hermet, Bertrand Badie, Pierre Birnbaum, Philippe Braud (eds), Paris, Colin, 1994

The Social Science Encyclopedia, Adam Kuper andJessica Kuper (eds), London, Routledge, 1996 (2nd edition)

The Encyclopedia of Democracy, Seymour Martin Lipset (ed), Washington, Congressional Quarterly, 1995

Pipers Worterbuch zur Politik. 1: Politikwissenschaft, Dieter Nohlen, Muinchen, Piper, 1985

The American Political Dictionary, J.C. Plano and Milton Greenberg, New York, Holt, Rinehart and Winston, 1985

A Dictionary of Modern Politics, David Robertson, London, Europa Publications Ltd, 1993

The HarperCollins Dictionary of American Government and Politics, J.M. Shafritz, New York, HarperPerennial, 1992

320 317

590 534

687 626

298 293

600

3660 75

540 84

1160 64

540 79

98 1220 53

417 208 2390 55

830 701 1700 71

1200 219

519 467

5000 309

300 63

430 87

150 82

Total 10790 3774 1700 99

Entries have been inputted according to the following criteria: skipping all biographies and other non-conceptual topics (such as countries); Piano and Shafritz entries have been considered only when in Bogdanor; Nohlen's entries have been excluded when specifically statistical (as indices and models).

that we felt we could master both on logical and intellectual grounds. To expand the number of the terms much further would not only involve an organizational effort that we were unable to afford, it would also push our universe of discourse

287


to a level of detail we were neither prepared nor interested to address. This provided one second criterion, that the terms we take into consideration should be placed sufficiently high on the ladder of abstraction. Moving beyond the 100 would have drawn us into overly descriptive terms. Indeed, this latter criterion has led us to exclude almost a third of the terms originally selected which, while present in most dictionaries, scored poorly on conceptual complexity. Merely designative words, such as "ombudsman" or "ballot," clearly stood outside the scope of our enterprise.

In the process of selecting words, we faced a problem of standardization. Working in four languages and five cultures (American, English, French, German, and Italian), we had to translate all words into one language-English. Then, we had to transliterate, providing a single word or wording for formulations that varied from country to country but had to be taken as identical-for example, participation, political participation, participation in politics.

Another problem we were forced to confront in the giving and taking of names is that some names are single and some occur in families. That is, some exist only in one form, for example, "bureaucracy," while others regularly appear both in their simple form and in compound forms that make up families of related but often different concepts-for example, "democracy," "industrial democracy," "social democracy," "participatory democracy," "mass democracy." Although we decided not to include each of these compounds as a discrete concept for definition, family words were taken into account in our weightings of concepts; a word was weighted because of the pure frequency of its occurrence (bureaucracy), or it could take on a greater weight due to the breadth of its usage singly and in family compounds (democracy). Again, our choices were those most heavily weighted in each class.

In the end, however, we were forced to compromise between our efforts to select terms on as much a quantitative basis as possible and our interest in including terms which, in our own judgment, belonged to the founding fathers of the discipline. We were, indeed, dismayed at finding that such keywords as "citizen" or "responsibility" could not qualify on a numerical basis, and that even the concept of "public," perhaps the cornerstone of modern politics, had practically disappeared in all political science dictionaries. We have thus added to the original list several terms which scored lower in occurrences but much higher in our own consideration and interest. Here again, users are free to subtract as well as add keywords.

We believe the final list of keywords to be an objective representation of the present political science vocabulary. We make no claim to having developed the one best list of political science keywords. We also recognize that some will dispute our use of the level of abstraction and saliency as criteria for amending the original list based upon frequency alone. But by putting forward our methods and offering such a list as the basis for concept development and theory-making, we think we have taken an important step toward clarification and consistency in political language, which should be the aim of every political science dictionary. In an age of extensive computerized collections of full-text databases, other, more effective, comprehensive, and systematic selections of keywords can be envisaged. We will be happy if our contribution will stir new and more enlightened endeavors, and we consider that progress can be measured by the speed with which our list is supplanted by a superior one (see Table 2).

288


TABLE 2. The Hyperpolitics Keywords (ranked in order of number of appearances in dictionaries).

Anarchism Authoritarianism Bureaucracy Capitalism Fascism Legitimacy Party Authority Constitution Corporatism Democracy Federalism Law Liberalism Liberty Marxism Pluralism Power Revolution Socialism State Centralism Charisma Coalition Conservatism Culture Dictatorship Ideology Majority Media Participation Populism Referendum Republic Rights Sovereignty Terrorism

10 10 10 10 10 10 10 9 9 9 9 9 9 9 9 9 9 9 9 9 9 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8

Totalitarianism Absolutism Alienation Cabinet Class Colonialism Communism Oligarchy Polling Representation Administration Aristocracy Clientelism Community Faction Feudalism Imperialism Justice Leadership Monarchy Opposition Patronage Police Racism Socialization Utilitarianism Utopia Violence Corruption Elite Executive Group Legislature Lobbying Nation Propaganda Radicalism

8 7 7 7 7 7 7 7 7 7 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 5 5 5 5 5 5 5 5 5

Agenda Conflict Crisis Election Government Institution Minorities Policy Regulation Religion Responsibility Trust War Welfare Agency Choice Citizen Civil society Collective Contract Court Market Opinion People Regime Consensus Decision Estate Individualism Interest Loyalty Order Public Rule Corporation Reform Individual

Stage Two: Choosing Contexts

Having completed the selection of keywords, we can move to the second stage, providing the proper context for each keyword. Here again we are confronted with the problem of overwhelming abundance resulting from subdisciplinary specialization, which must somehow be systematized. If the keywords are to be used as hypertextual links, sources must be selected that will provide a sound scientific guidance to hypertextual navigation. Our choices reflect the conventional wisdom of the discipline: We move from elementary definitions to in-

depth analysis through books and articles, to references to mainstream theories of

prominent authors. Since the Web provides many opportunities to develop in-

depth contextual hyperlinks, we are confident that our present collection will

quickly be obsolete. For the moment it allows us to test a system bridging a very wide number of contextual specifications with informed scholarly guidance.

4 4 4 4 4 4 4 4 4 4 4 4 4 4 3 3 3 3 3 3 3 3 3 3 3 2 2 2 2 2 2 2 2 2 1 1

289


The first step, elementary definitions, provides the conceptual backbone. The natural choice was to go back to the dictionaries we had already selected for sources of keywords and to make their definitions available as an introduction. However, and totally aside from the sticky problems of copyright, a mere inclusion of the original definitions would have limited access to those familiar with the original language of each dictionary; even in translation, variations in the links and stress cause more problems than they cure. Our aim, on the contrary, was to make a virtue of the differences of context and referents and to try to capture the often subtle nuances of difference in definitions due to cultural experiences or ideological, cultural, or methodological bias. Our solution consisted of translating, summarizing and characterizing each of the selected dictionary entries. In our own summaries we provided the number of words in each definition; whether the definition was approached by historical, bibliographic, or purely formal and linguistic references; and what was the author's normative slant. The following summaries have been edited for consistency (see Table 3).

The result is approximately 500 summaries that can be easily read and compared through the Hyperpolitics software platform, but could be just as easily accessed in a more traditional printed format. We thus make available what may be considered a dictionary of dictionaries, one that can be easily expanded to include other languages, other cultures, and additional definitions not only from other dictionaries but also from existing handbooks, such as Finifter's State of the Discipline, Polsby's Handbook of Political Science, or Grawitz and Leca's Traite de Science Politique. But a caveat is in order: care must be taken not to include so many keywords and definitions that consistency and cross-referencing of terms are compromised.

The next step in conceptual development is the choice of books appropriate for specialized analysis and research. Our quest for both quantity and quality produced a compromise consisting of a database with over 4 000 book reviews from ten years of the American Political Science Review, in full-text electronic format accessible for any kind of information retrieval, as well as reading of content. The link between each keyword and the book reviews is through the book title. Although more sophisticated relationships may be developed, under current circumstances we feel that the presence of the keyword in the title (or subtitle) remains the strongest available testimony of a book's relevance to that particular keyword.

When we started the Hyperpolitics project, more than six years ago, scanning over 4 000 book reviews seemed a formidable enterprise. In a short span of time, the scene has dramatically changed. The complete series of the American Political Science Review, along with several other political (and social) science journals, is now available on line through the JSTOR Web site. The International Political Science Abstracts, the most authoritative collection of abstracts in the field of politics, edited by Serge Hurtig, is now being published on CD-ROM as well as in the original paper edition. One can easily see a goldmine of opportunities for pursuing our methodology far beyond its present limitations. Expansion will, however, imply some problems of software development and copyright restraints, and, last but not least, exposure to the risks of gigantism. Overgrowth in the number of sources may easily turn out to be a curse on any attempt to obtain more consistency and clarification even as it expands contextual boundaries. Every attempt to push back the quantitative frontier risks transforming a useful reading list into an unmanageable library catalogue.

290


TABLE 3. Summaries ofDefinitions from Other Dictionaries (two examples from an average of six summaries per key word).

BGD Political parties. In about 1600 words, this entry discusses the origins of political parties, explains why parties are so signifi- cant politically, and offers a definition for the term.

Although their origins are essentially western, political parties exist in virtually every contemporary governmental system. The emergence of modern parties, constitu- tionalism, and representative government are closely linked. Party scholars generally agree about two ideas. First is the notion that the emergence of political parties marks a critical juncture in western political development because it represents the institutionalization of organized competition for power. Second, scholars also agree about two basic types of parties: those initially generated inside the legislature and those born as social movements or interest groups outside the legislature. The entry discusses each of these commonalities and offers examples as well.

Parties across the political spectrum share an important institutional trait-they provide the link between the formal government system and the different elements making up civil society. Thus, political parties exist along the boundary between the formal and informal segments of the sociopolitical system. Unlike interest groups, which seek only to exert influence, political parties endeavor to gain positions within the formal power structure.

The entry concludes by reemphasizing the defining characteristic of political parties: they are the linking mechanism between state institutions and civil society. Although most of the constitutional scholarship all but ignores this linkage role, it is vital to the proper functioning of complex modern societies. However, as the entry also points out, no definition of political parties would be complete without some reference to their role in winning elective office.

KPR Political parties. In approximately 1400 words, this entry notes the difficulty of defining political parties, distinguishes between political parties and interest groups, and presents three dimensions useful for classifying parties.

Those who study political parties have not been able to agree on the term's definition. Initially, in the nineteenth century, political parties were defined as "organizations that try to win public office in electoral competition with one or more similar organizations." This narrow definition, as Schlesinger points out, excludes certain organizations that are generally considered political parties. The effort to include such organizations, however, makes the definition too broad to be useful. His solution is to distinguish between two types of parties: governing parties in one-party states, and competitive parties in two- or multi-party states. In the latter, a further distinction between political parties and interest groups must be made. First, as Almond demonstrates, interest groups typically articulate interests, while political parties aggregate these articulated interests. Second, Blondel's work suggests that interest groups tend to have either a promotional or a protective character, while political parties simultaneously display both characteristics. Political parties can be classi- fied based on three dimensions: form of organization, party program, and party supporters. Duverger's work demonstrates a positive empirical relationship between these three dimensions. How a given political party operates in a democratic system depends on the party's own characteristics and its interactions with other parties. On this point, the literature distinguishes between two-party and multiparty systems. Sartori's and Blondel's theories about classifying party systems are introduced here. The entry concludes by briefly mentioning the empirical relationship between party systems, electoral stability systems, cabinet coalitions, and cabinets.

291


FIGURE 1. The Concept in Its Context.

After definitions and books, the third step in placing keywords in a proper context is to associate them with the major theories in the discipline through key authors. Indeed, several dictionaries append some sketchy biographies, mostly of such giants as Weber, Tocqueville, and Marx, as complements to some of the most relevant entries ("bureaucracy," "democracy," "class," and so on). We have tried to develop this feature in a more systematic way, by planning a directory with "conceptual biographies" of the most prominent living political scientists whose work is associated with each concept. Each author's main achievements-and books-are illustrated with a reference to some (four to six) of the selected keywords, thus creating one more link for placing each keyword in a wider intellectual and theoretical environment. To better integrate the various sources of hypertextual navigation, we plan to provide on-line book reviews of one or more author's main works (see Figure 1).

We have now reached a point where a wide array of specifications, sources that help contextualize the concept within its own subfield, is available for each entry. If we were to stop at this level, we would have confined our efforts to a narrow, although much enriched, territory; we would be missing the most dynamic part of

292


any theoretical discourse: that which explicitly links a given concept to others most

closely associated in the argument one intends to pursue. This is what we call controlled linkage: we place each concept at the core of a

matrix with the several concepts that are, to us, essential in the definition of the core concept. This is the normative dimension of our discourse, where we provide a set of relationships whose aim is to define boundaries. At this normative level, links are meant to expand meaning and at the same time to enclose it within a

clearly structured space. In many respects, we are following Sartori's rule that "in reconstructing a

concept, first collect a representative set of definitions, then extract their characteristics and eventually construct matrixes [sic] that organize such characteristics meaningfully" (Sartori, 1984: 41). Yet, thanks to hypertext language, we are able to apply this rule not to each separate concept seriatim and within a limited set of characteristics, but to a much wider spectrum of sources and

relationships.

The Thesaurus Puzzle

"Only connect ..."

The thesaurus can be considered the point of origin for conceptual analysis and

development, a first effort to break words, especially complex words, into their

parts, in order to determine how and in what respect each relates to other words. Peter Mark Roget recognized the problem when he began putting together in 1805, and eventually published in 1852, his Thesaurus of English Words and Phrases, the first guide to words and their connections to other words. In the Foreword to the 1979 edition, the editor reports that Roget "grasped what many writers, orators, and translators had long had to contend with: the necessity of

finding the exact appropriate word or phrase from a group of synonyms ..."

(Roget, 1979: v). The Editor's Preface of one hundred years earlier said it even better:

Many words... are found to be capable ... of being applied in many varied associations. Connecting links, thus formed, induce an approach between categories; and a danger arises that the outlines of our classification may become confused and eventually merged. Were we to disengage these inter- woven ramifications, and seek to confine every word to its main or original import, we ... [would] deprive our language of the richness due to an infinity of natural adaptations (quoted in Roget, 1979).

The thesaurus is an extraordinary accomplishment, but it is at once too broad and too narrow; too broad in that it covers the entire vocabulary of a language, or seeks to, and too narrow because it seeks to establish word associations on one level only, in this case the level of synonyms and antonyms. This gives it a universal applicability but no particular purpose. Yet, for all that, the thesaurus can be taken as the beginning of a "terminography"' that is essential to enriching, if not actually constructing, the theory of a defined field of knowledge, such as political science.

The next step then becomes one of moving up in the level of abstraction-and complexity. For only by linking concepts beyond their strict contextual boundaries can the language of politics serve its main purpose: theory.

293


Bringing the System Back In

One very large problem has always stood in the way of progress: to address a whole system, any whole system, requires removal further and further away from experience toward abstraction-from real actions and events toward their attributes, further still toward extraction of more and more microscopic attributes, and classification of those attributes until real experience can vanish altogether, leaving only the smile of the Cheshire cat. In the 1950s, when "the political system" was inaugurated, it was a pregnant idea because it encouraged political scientists to try to write political theory. None of the efforts has been successful, including the original, seminal effort by David Easton (1953), but it had a major influence, because it helped elevate political science discourse beyond and above the search for empirical political reality.

Several years before the full tide of search for "the political system," Harold Lasswell had anticipated systems thinking and the problem of thinking in systems with recognition of the need for a "configurative analysis." Configurative was Lasswell's word for context-the need to develop a context or framework within which particular phenomena could be studied without losing the sense of how they fit into the whole.

Lasswell thought he had the solution to the problem of meeting a happy medium between experience and abstraction, and he almost did, in Power & Society (Lasswell and Kaplan, 1950). Laswell's approach to the solution was to build context-to build toward "the political system"-with concepts as the actual unit of analysis, rather than trying to extrapolate from one reported finding to another with statements of increasing generality that might tend to rationalize that particular connection.

The culmination of Lasswell and Kaplan's encyclopedic work is a 64-cell table of concepts. These 64 concepts were derived in turn from eight concepts chosen for what we might today call their paradigmatic value, because each of the eight was treated first as a "base value" (for influence) cross-tabulated against the same concepts treated as "scope" values (the goal of influence), producing 64 property spaces into which the 64 concepts were literally logically derived from the interaction of the eight values paired and cross-tabulated. We have chosen to call this a "logically closed system" of concepts concerning power in society.

This is clearly a very large step beyond the thesaurus. It is not a thesaurus or a dictionary or an encyclopedia, although it possesses some of the features of each. That 64-cell table of concepts derived from the eight paradigmatic concepts has all the potential of a calculus, in which concepts have independent integrity by virtue of their definition but, beyond that, have a relational integrity that provides meaning, generates implications, and provides a driving force for new and unanticipated ideas. This is something that no dictionary, no encyclopedia, and no thesaurus can provide. Such a matrix is truly a system of thought, in that it has boundaries and all the parts within the defined space have a specified, and understood, logical place within those boundaries.

This is not, by itself, political theory; however, progress toward theory is hard to imagine without such a way to define concepts within a closed or closable system, while at the same time maintaining the ability to know roughly what is included and what is not, and what role each item plays within it. The matrix is empirical, to the extent that concepts embody experience; it is also abstract, without merely leaping to a higher level of generality without knowing what the intermediate

294


steps are. The steps are in fact quite knowable. Unfortunately, Lasswell and Kaplan devoted only 13 pages to a discussion of their invaluable table, and most of this was further dictionary-type treatment of many of the words that occupy the 64 cells. Yet, they set a precedent which can hardly be underestimated, and we gladly pay our tribute to their vision as we move toward our own proposal.

Closing Relationships into a Matrix

In focusing on the relational dimension of a concept, the first step is to provide a linking pattern: how many relationships are we willing to take into account, and how close do they need to be to the original core concept? The basic way to provide a format is to pose a simple network (as in Figure 2a), with one concept at the center, and arms extending out to other concepts most readily associated with the core concepts. Suppose we set a rule of four concepts deemed to be directly linked to the core.

concept

(CS' co\ ^_ /^ncep concept 2

concept

concept

FIGURE 2a. Links as an Unstructured Network.

Democracy, for example, could easily be linked to: individualism, majority, elections, and participation. This is, in fact, what several dictionaries may indicate as the cross-references. However, two problems arise from so loosely defined a pattern.

First, no indication is given as to the interrelationship existing among the concepts. Is there a direct link between, say, majority and elections, or individualism and participation? It is important that all relationships and interrelationships be explicit. The existence of such links within the chosen set of related concepts indicates that the network around the core concept is a tighter, more structured one. In terms of theory generation, this might indeed become the crucial factor. In fact, the logic of theoretical argument feeds on the wealth of self-reinforcing associations.

295

a 1-'k

S' . o

a _

ot I

.

I ' .

(^ 1-4-

N cls

1-

N.

N.

FIGURE 2b. Links as an Unstructured Network.


The second problem is that the cross-reference cannot be limited to direct or primary links. Indirect or secondary associations often play a major role in capturing the nuances of a given concept. By linking democracy to individualism, we are also relating democracy to the wider set of cross-references that belong to the individualistic discourse-rationalism, say, or market. How can we take that complexity into account? The way to find second- or third-level associations cannot be simply to surf from one cross-reference to another, a game that anyone with a computer can play out to the mathematical limit. With such an unstructured pattern, rather than gaining in complexity we would only be adding confusion. Moving from network to interlinked networks, by the third round we would reach 64 combinations (43), with the end not yet in sight. Once again, a one-way ticket to chaos (see Figure 2b).

Relationships among concepts to build a sound theoretical argument are not created by merely multiplying the number and levels of cross-referencing, or by just pushing a key or typing in a word for a word search. To generate theory that is both substantially rich and logically consistent requires some kind of a model to guide and specify the steps necessary for choices of relationships. This does not mean there has to be "one best way" to select and specify concepts that link up in some substantive way; in fact, we hope the model we propose will quickly be reformed or replaced with superior alternatives that are more substantive, more revealing, more logically powerful, more suggestive of further linkages, while keeping the logical system closed.

The model we present has axes rather than networks. The core concept is placed at the center of two intersecting axes (Figure 3). Axis A1,2 + Axis B1,2 are our

FIGURE 3. Toward the Model: Step One.

297


selections of primary concepts we deem necessary in the definition of the core concept. Already we can see our way out of the chaos so threatening in the utilization of the network format laid out above. First, by using dichotomous axes, we indicate a direct relationship within each pair of concepts. We follow a rule or practice associated with argument at least since Aristotle and more explicitly and formally since Hegel: we argue dialectically through antinomies-dichotomies of opposed or competitive concepts, or concepts that move along a continuum from least to most of the property in question.

Second, by intersecting the two axes, we now have a larger logic incorporating the format, because the interaction among the five concepts (core plus the four primary links) produces a calculus, a genuine calculus in the spaces around the core entry.

By use of this calculus, we can move one step further, from primary to secondary or indirect links-not on the basis of serendipity alone, but as a consequence of the interaction among the two primary axes. The four outside corners of the matrix can be occupied by what we call peripheral concepts (see Figure 4a). We now close off the spaces with secondary or peripheral concepts. In the selection of the peripheral concepts, the same criterion-the dialectic of antinomies and continua-was stressed. But, at this secondary stage, the concepts were actually derived from the cross-tabulation and intersection of the two primary axes and the core concept. Secondary links are thus logically driven and constrained by the previous choice of the direct links and by their potential for generating further interactions. Indeed, they perform as a consistency check on the original cross-tabulations.

FIGURE 4a. Toward the Model: Step Two.

298


Moving from the five original concepts to those which can be placed at the

periphery of our model is, in the first place, an exercise in the scale of complexity of a theoretical argument. In fact, finding two dichotomies that help define the axes of a given concept is an elementary task. A relatively large number of dichotomies can be selected, according to one's theoretical perspective.2 Yet, by cross-tabulating the two axes and thus moving to incorporate into the definition the secondary links resulting in the four peripheral corners of the matrix, each choice of the primary links undergoes a more challenging test. Not all dichotomies, once they are cross-tabulated, will produce secondary links that are both logically and substantially sound; the interaction may be barren, yielding only blank corners. This would be a sign that the argument cannot be pushed much further. In turn, success in producing secondary links reinforces the relationship between the primary axes and testifies to the logic and the creativity of one's original choice. Sound peripherals ensure sound primary links.

In the example we provide for party (Figure 4b), we chose, as defining axes, two mainstream dichotomies. The first one refers to Duverger's distinction between parties originating in the legislature and parties arising from cleavages in the civil society. The second dichotomy reflects the difference in the organizational pattern, contrasting the corporate party of the European continental tradition with the individualistic machinery that has prevailed in the United States. If we now move toward the second step of indirect links, the upper left peripheral we chose is participation, as parties of the corporate type-like the German Social Democratic Party in the pre-World War II era, or the Communist Party in post- World War II Italy-conceive of their presence in the legislative arena mainly as a way to maximize mass mobilization. On the other side, in the upper right cell, we suggested representation as the main relationship between individualistic parties and parliament. Moving to the lower half of the matrix, class is indicated as the conjunction between civil society and the corporate party, while market plays that role in relationship to an atomized, individualistic party. In our definition of party, the cross-tabulation of the individualistic vs. corporate and the civil society vs. legislature axes is thus conducive to four more keywords-participation, representation, class, and market: a complex but logical set of relationships for which we expect to be able to offer theoretically interesting arguments.

As with any theoretical argument, the example we are offering here for party can be criticized on substantive grounds. The choice of each concept in the matrix reflects our own view of the issue, a view which needs to be confronted with, and checked by, other statements of that very same issue. Needless to say, our model does not provide us, or anyone else, with a tool for automatically generating the right theory. A thinking machine for political science is a nightmare far exceeding the power-and aims-of our model. What our model does offer, however, is a system for controlling the logical consistency of conceptual links in the construction of a theoretical argument, at the same time giving the possibility of comparing, criticizing, and possibly modifying specific segments of such an argument. By clearly identifying each link within the overall matrix, step-by-step demonstration-or refutation-of one's argument becomes a simpler under- taking. We recognize that we are subject to cultural myopia and ideological bias; the difference here is that we provide any reader (or user) with all of the means possible to expose our weaknesses and to make up new matrices with different cross-tabulations. Science advances through weaknesses as well as strengths, so long as all of the steps along the way are explicit.

299


FIGURE 4b. Party 's Main Axes and Peripheral Links.

Along this line of argument-and constraints-the logic of our model can be pursued one step further. In fact, the indication of the peripherals truly closes off the scheme and permits the calculus to work at a more complex and abstract level. It provides genuine closure for each concept and all of its linkages to the eight other concepts. But it also provides known or knowable characterizations of the resulting property spaces, permitting us (or anyone) at this point to look for four additional derived linkages, one for each of the property spaces.

The third step-defining the property spaces-involves four sets of independent variables: the core entry plus one from Axis A and one from Axis B, plus one peripheral of each corner. The dependent variable is the spatial concept. Property spaces are thus composed of concepts that have a third-level connection to the original entry, through the peripheral and the axis. In an unstructured model, at this stage of the game, we would be left with an unguided choice among 64 options. In the model we propose, third level links become a constrained option, one which needs to be derived by a logical-and substantial-relationship to at least four more concepts which already belong to the matrix (see Figure 5a).

Every property space in each matrix can contain more than one concept, as the interpretation of the interactions of axes and peripherals in each space may vary from one author to another. However, at this level, variations can be expected to be of a small scale. The advantage of getting so far by complying with explicit logical rules is that each further step will be one of increasingly constrained choice. Once again, no automatic concept generation can be expected, even at this level. Quite to the contrary, property spaces will only be properly filled if our choice for the core axes proves to be a farsighted one. As we are left with fewer choices along the path we ourselves have set, it will become clearer whether we are

300


FIGURE 5a. The Model: Axes, Peripherals and Spaces.

heading in a direction allowing for a complex and consistent theoretical argument.

In the example we propose, election appears in the property space defined by the four concepts of the upper right quadrant. Party is often virtually defined by elections because so many parties in the Western-type democracies have used elections as their main means to power and influence. But as elections have become so much more a mass phenomenon, it would appear to be overly constrictive to confine election to one quadrant alone: Even though election is logically compatible in the upper right quadrant, what can be made of its absence elsewhere? Observers will inevitably differ on the particular spot among the coordinates where a given party system resides, but that is a constructive game in itself, because the differences of placement require clarification of meanings and re-examination of data (see Figure 5b).

Thus, as the us political parties move more and more toward polling, they also move that much further away from their historic legislative/representation position, even though legislative activities still take place. Movement toward polling also involves a more plebiscitary relation to individual citizens, and that moves any party toward the lower right quadrant-how far in that direction depends upon the weight a given observer puts upon plebiscitary and consensual relationships rather than traditional elections that lead to representation.

Additional understanding of this can be drawn with a move to the lower left quadrant of the matrix, where "oligarchy" best expresses the texture of political relationships. The idea of oligarchy is a driving force if not the driving force, for the corporate party dates back at least to the work of Robert Michels and is probably best represented by the Leninist and social democratic parties all

301


FIGURE 5b. Party: Axes, Peripherals and Spaces.

through the inter-war and post-war eras. But in our definition, oligarchy includes a wider logic of party behavior well beyond the original ideological boundaries of social democracy and communism. In contrast to polling, oligarchy stresses the

continuity that mainstream parties of a corporate type tend to present both in their organizational and in their ideological reality, especially once a class basis is established.

Yet, in the political realm, oligarchy requires consent just as much as any other

organizational type. Political scientists often refer to the "engineering of consent." But it is not a universal generalization; in fact, it applies best in this lower left

quadrant. That is, lacking the broad base of a mass or a representative system, the

oligarchy party must create an appearance of consent, usually by forms of internal

procedure and organizational representation that are ultimately cooptative rather than genuinely representational. We call it engineered consent; we could also call it virtual representation. This leads us to the upper left quadrant space, coalition, where we can say that consent is organized. As parties take on a more corporate form, they remain oligarchic, and they may very well remain strongly attached to a class, ethnic group, religion, or other social classification in a society, but they exist in a governmental system that comprises several corporate parties, each oligarchic, but each having to contend peacefully with the others. Whenever differences in the electorate remain frozen or otherwise attached to a substantial aggregate in the civil society and are represented through strong party organizations of a corporate type, the most likely outcome is a governing process arising out of coalitional accords with a minimum of direct electoral competition.

302


Matrix, Model and System In our presentation, we have often referred to matrix and model as inter- changeable terms for our logical compass in the conceptual universe. In fact, "matrix" is the descriptive word for a pattern of interaction among (conceptual) variables, whereas "model" refers to the predictive and normative aspects implied in the chosen pattern. As we have already said, our model can be changed and improved; different and better models can be chosen to replace it. However, using the Hyperpolitics platform implies-and encourages-respect for the basic rules of model-formation: production of explicit hypotheses, simplification and reduction of reality, use of the resources of logical argument, and testability on the basis of empirical evidence (see Hermet et al., 1994: 173).

In the beginning, all concepts are givens in the political science dictionaries; but their placement in each matrix is our own. We ourselves choose which concepts to include in a matrix and how to dichotomize these concepts. We choose them not only for their relevance as definers of the core concept in question but also in terms of our sense of which gives us the logical control we need-indeed, that any theorizing needs. The argument we have just given that accounts for our conceptual choices in the property spaces of the party's matrix can be further elaborated, with more examples and comparisons across time and space; or it can be cut short as a mere cue and invitation for discussion and criticism from different viewpoints. It can also be pursued further into the abstract, by pitting the normative implications of one quadrant against another, treating each one as a different ideal type of the political party. Yet, at this stage, the argument will be necessarily confined to the logical links and contextual references seen as proper to a particular matrix. The matrix acts as a powerful calculus as well as a container.

There comes a point, however, when the limits of the model are too constraining, especially to those who see the still larger, undefinable universe out there-the state, or the polity, or the political system. We have then to recognize that the beauty and the power and the danger of concepts must be confronted, just as if they were rational individuals: they come to have a life of their own. And, as in Menger's methodological individualism, the beauty of each concept having a life of its own is this: the potential they add to the richness of the political mind is practically infinite.

We cannot offer logical closure around all of politics within one integrated system. But we can work our way to the higher and less controlled level of abstraction by moving at least one step further. That is, we can move outside a concept's matrix by connecting it to all the other matrices where that concept has been included-whether as a member of an axis, as a peripheral concept, or as a special property. This is navigation at the systemic level, as is graphically demonstrated on Figure 6a.

Working through relational matrices contributes to our understanding of each concept as a prismatic universe, with its meaning depending on the links it incorporates. As we argued before, each concept when treated as the core entry can engender different axes, and therefore a different theory. Moreover, each concept presents different meanings when it appears not as the core but as an axis or as a peripheral or as a space attribute, according to the other matrices where it is being used. For example, "participation" as used in the case of the relationship between the corporate party and the legislature will be different from its meaning

303


FIGURE 6a. System Navigation.

in conjunction with the concepts of, say, social movement or citizenship. Our model offers the possibility of clearly identifying those differences, and accounting for the variation in meaning from one type of linkage to another.

What we see in Figure 6a is only a partial representation of the number of

systemic links that can be traced among the different matrixes for each concept. On purely statistical grounds, we may expect an average of twelve systemic links outside the original matrix where each concept is placed at the center. In fact,

304


Boo-k Rviw

The conept' rsL \ books' APSR

Ten years full teo The Concept

book

\ in te ^ a

\ Context \he a 5 500 English

from 10 \ dictonanes' \

\ definitions \ \ four \

\ languages \ man books ~\^~~~~ \ ~reviewed

The once0 t s authors

Directory of

One Hundred

Lvmng Polbtical

Sc.entsts

FIGURE 6b. System Navigation in Context.

305

et xt


several concepts fare much higher on the scale of systemic "popularity": democracy, community, state will probably appear as definers of the core concepts in a much larger number of matrices. However, this will only depend on the judgment of individual scholars, who remain the only ones responsible for filling the various links in each matrix according to their own theoretical understanding.

By weighing the overall presence of the various concepts according to different authors, we obtain a preliminary test of conceptual preferences-or bias-in the various systems of thought each scholar is pursuing. This quantitative measure is, however, only one of the several indicators one can extract from the Hyperpolitics platform. For instance, it may be useful to consider concepts in view of their predominant place in the different links within the matrix. Some concepts may come to be used mainly as axes, that is, at the primary level of relationship with the core concepts. Other concepts will be more often chosen as space attributes, that is, at the more abstract level of intersection between core concept, primary and secondary links. Thanks to the computer, we can have at our fingertips an overall classification and count of concepts as axes, peripherals, and spatial attributes to guide ourselves into better understanding of the use we are making of each of them within the system as a whole.

Aside from this numerical and classificatory information, the most challenging use of Hyperpolitics at the systemic level consists in the possibility of reconsidering the place a concept has in one matrix in light of its use in another one. Are those two-three, four, five...-uses complementary or are they conflicting? Is there theoretical compatibility between, say, the way party is used in its own matrix and the ones that have democracy or elections as core concepts? This systemic confrontation of concepts involves more than an exercise in logical construction. Since each concept embodies empirical experience, combining the experience embodied in a concept as it moves from its role as a core concept to an axis, peripheral or spatial role produces tension. But this is creative tension, because it requires new sets of statements about these hitherto unrecognized patterns of contrast. This kind of creative tension among competing experiences is the essence of theory-or theory-making. Theory-making is its own political process.

This process is assisted by-and integrated with-the contextual links that the Hyperpolitics platform provides. Reassessment and replacement of concepts at the systemic level is based both on the comparison among matrices and on the reconsiderations of sources we have originally used-or overlooked. Through hypertextual links from within each matrix, we go back to the dictionaries' definitions, book reviews, authors' intellectual biographies-or whatever other source we may have selected as part of our platform-in order to confirm or modify our original matrix. This is, in the end, a dynamic representation of hypertextual thinking. Bottom-up, top-down, and repeat: concept formation through contexts, through logical linkages, and back again, through successive approximations.

All this said, systemic navigation is limited by its own complexity and interactivity. Because of the number of diverse implications and relationships each concept may encounter in each individual matrix where it appears, hypernavigation is mainly a work of redefinition and revision. Many of the matrices will be improved by retrofitting concepts within them. Yet, taken as a general pattern applied to each and every one of our concepts, systemic navigation would inevitably turn into a Sysiphean ordeal, with each entry being constantly, and endlessly, modified (see Figure 6b).

306


To escape the infinite motion of systemic hypertextual navigation, conceptual order must be maintained by means of the self-limitation of hypertheory. Much as our model offers a way to exert better control on a wide theoretical spectrum, one must still choose a number of conceptual domains for the most fruitful application of one's own knowledge-and imagination. The trouble with most inquiry, especially inquiries into the really important issues, is that we often set out without being clear where we were coming from and are totally unclear as to where we are going. There is no set remedy for this, but we must be sensitive to it; we must keep trying to escape the problem set forth best in Through the Looking Glass. Our rendering of Alice's story is this: if you don't know where you're going, any road will get you there. The hypertextual model laid out here provides at least some basis for calculating the direction and goals an individual may wish to pursue. One thing is certain: the more uncertain the direction, the greater the need for a calculus and a set of contingency plans in the hands of the navigator.

Conclusions

Hyperpolitics is an experimental project. Six years in the making, the Hyperpolitics project has been developed as a working prototype on a Macintosh- Windows crossplatform. In its present format, Hyperpolitics is available on CD-ROM to all those interested in using and improving it. We plan soon to make Hyperpolitics downloadable on line, through a Web-site on the Net. And, more important, to make it Web-compatible through software re-engineering allowing for direct plug-in to at least some of the on-line political science archives now accessible on the Web.

The very nature of the Hyperpolitics platform calls for continuous adaptation and improvement in light of each individual's understanding of political science discourse. In illustrating a few conceptual matrices as examples of possible applications of a closed logical system of relationships with other concepts, using "party" as our case study, we have already underlined the fact that several options are available in the definition of each core entry. Definitions-and theories-will continue to differ according to changing focus, country, and epoch.

The model we propose is thus simply a way for more clearly controlling for- and comparing-the set of logical relationships each author is incorporating into his or her own theoretical argument. Yet, the consistency of a matrix should not be evaluated on the grounds of sound logic. As the matrix is the main engine of the whole hypertextual system, it also offers the possibility of double-checking each separate term according to its bibliographical sources, other dictionaries' definitions, and other theories by major living authors. After placing a concept on an axis-or on a peripheral or spatial dimension-according to one's own creative, logically consistent vision, the next hypertextual step calls for a careful, in-depth screening of that choice against the wide array of sources that the system is providing. Does our placement comply with the placements that arise from the contexts where that concept is most commonly used in the political science literature? By reading and confronting a dozen-or a hundred-book reviews, comparing cross-cultural definitions and revising mainstream theories by contemporary authors, one has the possibility of strengthening (or abandoning) his or her initial thesis.

In a way, this appears to be no different from the way scholars proceed ordinarily. But differences emerge rapidly. One difference lies in the scope and

307


quick availability of sources. Through on-line hypertextual links, speed of access makes comparability more practicable, and comparability contributes to theory- making. A second, even more important, difference concerns the fact that all information must pass through a matrix of interconnecting links, thus multiplying the operation of in-depth conceptual scrutiny and transforming it into a more dynamic and relational enterprise. This task alone calls for a challenging revision in attitude toward theory-making, by enabling the mind to work with a larger number of variables that are rich both in context and linkage.

As with any intellectual game of this sort, Hyperpolitics can be played at different levels of complexity. While scholarly use for research and theory-building can exploit all of the system's resources, classroom utilization can be made on a step-by-step basis, keeping the different logical segments-and database modules-separate or only partially integrated. Teaching students to compare the same concepts as defined in different cultures and languages is in itself an important achievement. So is the possibility of having available all the recent book reviews dealing with that same keyword. At the same time, introducing a concept by being able to present the intellectual biographies of authors mainly responsible for its theoretical development goes well beyond the quick citation to which most classroom teaching is confined.

In moving from the first to the second level-from context to relationships-of the Hyperpolitics platform, one can keep to a more gradualist approach. The matrix in its basic format, the two axes, is in itself a good exercise in logical construction. At this stage, it strongly resembles the usual two-by-two matrix that is a familiar sight in most undergraduate courses. Yet, within a hypertext system, it reveals its potential for associating logical, more abstract links to a rich domain of contexts and sources, and it introduces students to genuine reasoning in social science.

All this should help explain why direct experience is so crucial to the development of Hyperpolitics. When we first started on this project, we had only a rough idea of what its overall implications might be. At the outset, we were animated by a rather naive enthusiasm about the extraordinary amount of new information that on-line archives have made available for scholarly research. Along the route, we have learned the lesson that in order to use any information for improving theory-making, information must be mastered and channeled within a clearly set course of action. That course of action is a hypertextual project binding context to linkage in a careful logical design, the basic aim of any theory- making, which at the same time must be constantly alert to the blessings and curse of the contemporary conceptual explosion.

While sharing the preoccupation that, in social sciences, "all disciplines have lost their discipline" (Sartori, 1984: 50), we are confident that much can be done to restore a stronger control over the meaning and use of concepts without losing their growing empirical, historical, and cultural wealth. Hypertext, if properly understood and used, can be an important step in that direction. The Hyperpolitics system we have developed is but a first step in an effort which we hope will be much improved by others who share our concern and vision: that (with Lord Byron) "every man [in political science] can be his own Aristotle."

Notes

1. As defined by Fred W. Riggs, "Terminography is very useful in any well established field of knowledge where the relevant concepts are familiar to specialists who have evolved

308


appropriate terms to designate them in an unambiguous way" "Coming to Terms with Social Science," mimeo, IPSA xvII World Congress, Seoul 1997.

2. For instance, in defining party, one might want to stress the role of elections in the development of parties as a key actor of modern democracies. In this case, the choice of the two primary axes may be, say, that of juxtaposing elections to organization (A1,2), as two different ways to establish party's supremacy, and market to class (B,2), as two competing principles for the development of party ideology. If one's interest lies rather at the institutional level, two different primary axes may be selected; e.g., civil society vs. the legislature (A12), regarding the party's representative function, and individual vs. corporate actors (B1,2), accounting for a different pattern in the party's social and ideological background. Or, if one's interest in party is more behavioral, then the primary axes governing the definition might well be class vs. status (A1,2) and attitude vs. affiliation (B,2).

References

Easton, D. (1953). The Political System: An Inquiry into the State of Political Science. New York: Knopf.

Hermet, G., B. Badie, P. Birnbaum, and P. Braud (1994). Dictionnaire de la science politique et des institutions politiques. Paris: Armand Colin.

Lasswell, H. and A. Kaplan (1950). Power and Society-A Framework for Political Inquiry. New Haven: Yale University Press.

Roget, P. (ed.) (1979). Thesaurus of English Words and Phrases. New York: Avenel Books. Sartori, G. (1984). "Guidelines for Concept Analysis." In Social Science Concepts. A Systematic

Analysis (G. Sartori, ed.). London: Sage.

Biographical Notes

MAURO CALISE is Professor of Political Science at the University of Naples, "Federico II". He has been a Research Fellow at the Center for European Studies at Harvard University and a Visiting Fellow at Cornell University. He has conducted research on political parties and governmental elites in a comparative and

developmental perspective. His latest publication, II partito personale, addresses the transformation of the political party in the age of mass communication and strong executive leadership. He is presently working on a book comparing the role of

corporate authority in Europe and the United States for the development of the modern state. ADDRESS: Dipartimento di Sociologia, Universita di Napoli, Vico Monte di Pieta, I - 80138 Napoli, Italy. [e-mail: [email protected]]

THEODORE J. LowI is John L. Senior Professor of American Institutions at Cornell

University. He received his doctorate at Yale in 1961, served on the Cornell faculty from 1959 to 1965, and on the political science faculty at the University of Chicago from 1965 to 1972. A member of the American Academy of Arts and Sciences, he has received honorary degrees from Oakland University and from the State

University of New York at Stony Brook as well as the Doctorat honoris causa from the Fondation Nationale des Sciences Politiques of the University of Paris. ADDRESS: Department of Government, Cornell University, 105 McGraw Hall, Ithaca NY 14853, USA. [e-mail: [email protected]]

Acknowledgements. Developing the Hyperpolitics platform has required a considerable effort, in terms of both human and financial resources. We wish to acknowledge the contribution

309

310 International Political Science Review 21(3)

of several individuals who have shared our enthusiasm in this enterprise: Annarita Criscitiello and Rosanna De Rosa have been in charge of the development of our keywords database. Rosanna De Rosa has also contributed to the drafting of the summaries for the selected entries, along with Marco Cilento, Brenda Holzinger, and Christoph M. Kunkel. The English editing of the summaries is by Michael Harvey. Robert Spitzer is the author of the intellectual biographies of American scholars drafted up to this point: we owe him special gratitude for his outstanding intellectual contribution. Gianluigi Visco has acted as main software consultant in the overall development of the Hyperpolitics project. Last, but not least, we wish to thank the Italian National Research Council (CNR) and the John L. Senior Fund for their generous support.

Earlier versions of this paper have been presented at the IPSA xvII World Congress in Seoul and at the ISA World Congress in Montreal. We gratefully acknowledge the criticism and support of our discussants at those panels: Rainer Eisfeld, Ada Finifter, Alberto Martinelli, Stella Quah, Fred Riggs, andJohn Trent. We have also very much benefited from our discussion with a number of colleagues, among whom here we may mention David Lytel, Robert Parks and Robert Spitzer.

Documents

Calise Lowi- Hyperpolitics