Why Anthropologists Don t Like Children

8/2/2019 Why Anthropologists Don t Like Children

1/17

LAWRENCE A. HIRSCHFELD

Why Don't Anthropologists Like Children?ABSTRACT Few major wo rks in anthrop ology focus specifical ly on childre n, a curious state ofaffairs given that virtually al l contem-porary anthropology is based on the premise that culture is learned, not inheri ted. Al thou gh ch i ldren have a remarkable and undis-puted capacity for learning generally, and learning cultur e in particular, in signif icant m easure anthro polog y has sho wn l i tt le interest inthem and their l ives. This article examines the reasons for this lamen table lacunae and offers theore tical a nd em pirical reasons for re-pudiating it. Resistance to child-focused scholarship, it is argue d, is a byproduct of (1) an impoverished view ofcul tura l learn ing thatoverestimates the role adults play and underestimates the contribution that chi ldren make to cultural reproduction, and (2) a lack ofappreciation of the scope and force of chi ldren's cu lture, particularly in shaping adult cultu re. The margina lization of chi ldren and chi ld-hood, it is propos ed, has obscured our unders tanding of how cultural forms emerge and why they are sustained. Two case studies, ex-ploring North American children's beliefs about social contamination, i l lustrate these points. [Keywords: anthropology of ch i ldhood ,children's culture, acquisit ion ofcul tura l knowled ge, race]

THE TITLE QUESTION of course is only half serious andclearly incomplete. Half serious in that anthropologistsas individuals presumably enjoy the company of children asmuch as anyone else. Incomplete in that my in tention is notonly to draw attention to the marginalization of children,but also to persuade that there are good, indeed quite com-pelling, reasons tha t children deserve a broad-based scholarlyregard.1

Many readers might object that anthropologists havedone a good deal of research on children, as the substantialliterature concerned with the intersection of culture, chil-dren, and childhood attests. As one observer put it, there are"enough studies of children by anthropologists to form a tra-dition" (Benthall 1992:1). To cite a few examples familiar tomost anthropologists: the work of Margaret Mead (1930,1933); Beatrice and John Whiting (1975; Whiting 1963;Whiting 1941); Brian Sutton-Smith (1959); Mary Ellen Good-man (1970); Helen Schwartzman (1978); John Ogbu (1978);Charles Super and Sara Harkness (1980); Robert LeVine(LeVine et al.1994); and linguistic anthropologists BambiSchieffelin (1990; Schieffelin and Ochs 1986), Elinor Ochs(1988), and Marjorie Goodwin (1990).2 Critically for the dis-ussion at hand, this work has not coalesced into a sustainedof child-focused research. Nor, as a chorus of re-

1981; Stephens 1998; Toren 1993), has it succeedednging children in from the margins of anthropology.Admittedly, mainstream anthropology (tacitly) acknow-

is a reasonable pursuit. By and

large, however, it is accepted th at it is a pursuit that can be ig-nored. I believe that it cannot. My goal is to address, ques-tion, and suggest ways to redress the neglect. In this article'sfirst section, I review this curious marginalization, considerwhy it is so widespread, and suggest that there is ample rea-son to believe that child-focused research should occupy theattention of both specialists and those in the mainstream. Ina second section I offer a brief empirical case study illustrat-ing this last point. With it I attempt to show that attendingto children, their singular cultural forms, and their uniqueconceptual architecture paradoxically reveals significant in-sights about the nature of adult cultural experience. Manyadult cultural beliefs, I suggest, are sustained precisely be-cause of the way the child's mind is organized and the wax-children organize their own cultural environments. Manycultural forms are stable and widely distributed just becausechildren find them easy to think and easy to learn (Sperber1996). Pursuing this argument affords an informative yet un-appreciated perspective on the relationship between individ-ual psychological phenomena and their role in the constitu-tion of cultural forms.

In the briefest terms, mainstream anthropology has margi-nalized children because it has marginalized the two thingsthat children do especially well: children are strikingly adoptat acquiring adult culture and, less obviously, adept at creatingtheir own cultures. Although it is uncontToversial that children

AMERICAN ANTHROPOLOGIST 104(2):611-627. COPYRIGHT O 2002. A M E R I C A N A N T H R O P O I D * AI A S M M A H O N


2/17

2 American Anthro pologist Vo l. 104, No. 2 June 2002cquire the wherewithal to participate in the cultures theyinhabit, the processes by which this happens has drawn rela-ively limited attentionpresumably because most anthro-pologists consider these processes unremarkable and unin-formative to the field's principal concerns. Children alsocreate and inhabit cultures of their own making, culturesthat in significant measure are independent of and distinctfrom those of the adults with whom they live. In making

their own cultural traditions, children deploy singular con-ceptual skills that significantly constrain and mold not onlyheir own cultural productions but also those of adults. Byviewing children as vehicles into which culture is poured, an-thropology has put the cart before the horse. In this article'ssecond section I empirically flesh out this argument. By ne-cessity, however, this presentation is brief and illustrative;my goal is simply to demonstrate one way in which children,their cultural acquisitions, and their cultural productions cane studied and suggest ways that this knowledge can be used toextend our understanding of cultural environments generally.As I just pointed out, anthropology's indifference tochildren is not a function of an absence of child-focused re-search. What is disappointing is the modest effect this re-search has had on the mainstream. According the AmericanAnthropological Association's 2000 AAA Guide, there are 155active full-time sociocultural anthropologists appointed inthe National Research Council's ten top-ranked departments.Of these, only nine (including the author) declare child- andouth-related topics among their interests. I do not meanprimary interest. Among these nine scholars, child- orouth-related work is on average the third area of interestlisted. Moreover, four of the nine are primarily concerned

ith youth or adolescence, an age that by definition is at theorders of childhood.Publications follow interest. Between 1986 and 2001, ac-cording to the Eureka database of peer-reviewed journals, thehas published three articles on chil-ren (excluding book reviews and studies of nutrition). Aearch for articles from the American Anthropologist in whichchild," "children,' "child-care," or 'childhood" figured as aubject or keyword yielded 14 hits since 1904. If introductory

0 popular anthropology textbooks, Erika Friedl found a pat-epresented in our texts but also undertheorized and outrighteglected is both strange and disconcerting: "With few ex-t only are underrepresented in our textsalso undertheorized and outright neglected" (2002:19).

It might be countered that not every interesting phe-enon is of interest to everyone. There is no a priori rea-

mitm ent to the Idea that culture is learned,

display elaborate and culturally specific ways of making-meaning and modes of behavior that are manifestly well de-veloped. Although often not considered fully expert, by ado-lescence children are adept participants in a culturaltradition.This is not a latent fact or hidden dynamic that m ust beadduced through close and systematic analysis. Virtually allfolk traditions recognize, and many explicitly comment on,this state of development. Consider, for example, the age-dependent "costs" of transgression. Young children's errors,from a cultural perspective, are relatively low cost, provokinglimited or no condem nation (Lancy 1996). In contrast, errorsin cultural performance committed by postadolescents aregenerally seen as more serious and are more likely to arousedirect criticism, punishment, or other sanctions. Plausibly,the reason for the difference in cost is the conviction that byor shortly after adolescence a significant degree of culturalcompetency is expected. This observation is not a function ofany particular way of construing culture. Regardless of one'sorientationregardless of whether culture is identified withconceptual skills, or a specified range of sentiments, or thewherewithal to produce cultural performances, or the capac-ity to transact cultural relations, or the processes by whichunequal cultural capital is extractedit is during childhoodthat the overwhelming bulk of these achievements arereached. Because so much anthropology is devoted to identi-fying, understanding, and conveying what people do, itseems uncontroversial th at exploration of how they come todo it would be a central preoccupation of the field.This last point might be termed a theoretical imperative.Another compelling reason to study children derives from animperative to "embarrassment managements Imagine thefollowing scenario: An ethnographer works among a popula-tion whose social structure is sharply stratified. One group,whose cultural identity derives from a notion of maturityand attributed competence, wields significant power over asubaltern class, whose identity derives from a notion of im-maturity and attributed incompetence. There are culturallyspecified economic, emotional, and social relationships be-tween these two groups, both of whom are named by termsroughly translatable into familiar English terms. There is ex-tensive ongoing contact, and subjacent conflict, between thesubaltern and elite. Although the former significantly out-number the latter, there remains a closely managed relation -ship of power that systematically disenfranchises the subal-tern population. In this culture, like many others, the elitetalk endlessly about their subaltern clients and seem to de-rive an important dimension of their sense of self and com-petency by therelativesuccesses and failures of their clients.Moreover, the ethnographer spends a great deal of timeamong the subaltern group, informally observing their ac-tivities, their publicly punished transgressions, and theirpublicly praised successes. Remarkably, the ethnographernever. In all her writings, mentions th e subaltern population,making scant reference to the economic and emotional rela-tions that dominate the elites' interaction with them.


3/17

Hirschfeld Why Don't Anthropo logists Like Children? 6 13My poin t here is transparent. Such an account would beseen as fundamentally flawed; similar lapses in attention togender issues, once effectively challenged, brought on a sig-nificant reorientation of the field. Ethnography and theorythat ignored women were recognized as impoverished andmisleading.Our hypothetical ethnographer, however, m ight seek to

deflect criticism by observing that unlike gender and genderrelations, which are constructed and interestingly varyingacross culture and time, the particular elite/subaltern com-plex she omitted involves a relationship of power, authority,economy, and sentiment that is universaland many an-thropologists do not feel compelled to attend close-ly to in-variant domains of experience. It is not obvious, however,that gender is less a universal and more a construction thanchildhood. Both are universally encountered and both aresystems of inequity, disadvantage, an d sustenance. Bear inmind that an adequate treatment of gender in culture in-volves more than simply acknowledging gender relations. Anadequate treatment of childhood and children similarly in-volves more than acknowledging that adults and childrenstand in a particular relation. "Add children and stir" is nomore insightful tha n 'add women and stir." In both cases, agenuine change in gaze yields a reconfiguration of the field.There is a final motivation to move children in from themargins that turns on methodological and theoretical com-patibility. Children's lives and experience simply lend them -selves to anthropological inquiry. In its simplest guise, an-thropology is the study of the n ature and scope of differencesn the ways different populations act, think, and speak. Chil-

ren's behavior, thought, and speech differ systematicallyrom that of adults. The tools of observation and analysis thatnti and N uer, or for th at matter gays in San Francisco or

as such can be as usefully explored by anthropologists asgalese street merchants in Marseille, Vietnamese ric eAs the nature and practice of everyday power be-

lives also become increasingly central.

(e.g., Levy-Bruhl 1979) and that have continued to be drawn(e.g,, Hallpike 1979), plausibly evoke discomfort among con-temporary anthropologists. Even if they themselves do notenvisage doing it, the image of these embarrassing com-pari-sonslike the image of stage theories of social evolutionmay have turned some away from the study of children alto-gether. Second, children are closely associated with womenand their traditional spheres of influencethe home andhear th. As a result, some have suggested that children suf-fered the same systematic exclusion from the anthropologi-cal gaze as their m others (Caputo 1995; James an d Prout1990). Intrinsic dullness is another explanation (although itis seldom, of course, framed in such pejorative terms). Formost anthropologists, the commonly evoked image of chil-dren is that of adults-in-the-making. Liminality, somewhatironically given the considerable anthropological interest inother forms of age-related status, in this instance generallytranslates into the notion of children as culturally incompe-tent creatures, who are, at their most interesting, simply "ap-pendages to adult society'' (Bloch 1991; Caputo 1995; Jamesand Prout 1990; Schwartz 1981; Toren 1993).These are not entirely satisfying explanations. Like Sar-tre's (1948) anti-Semite, who, as a result of a disagreeable en-counter with a Jewish tailor, despised Jews but not tailors,anthropologists uncomfortable with their predecessor's awk-ward comparisons of children's and primitive thought didnot end up abandoning the study of native populations,only children. Nor are children th e only population wh o suf-fered invidious comparisons under the ancien regime. Nine-teenth- and early 20th-century anthropology had muchworse things to say about blacks and w omen , yet this did no tcause anthropologists t o exclude race or gender from subse-quent study. "Gender, race, and class" as categories havegreat currency in contemporary scholarship and the atten-tion paid to how they are constructed and sustained are notexceptions but the rule. The feminist turn in anthropology isnow several decades old. Women and their lives are closelydocumented in the bulk of contemporary anthropologicalwork. In contrast, mainstream interest in children remains(figuratively) in its infancy, to cite Sharon Stephens's (1998)bon mot. It does not seem plausible that children are off theradar because their m others urre.3

Nor does intrinsic dullness necessarily translate into lackof attention. Even the most adult-centric perspective doesnot in principle preclude interest in children. Arguably, itmakes children all the more interesting. Children, after all,are appendages to adult society often being the central (if fre-quently silent) figures in both the commonplace and pri\i-leged activities. Rites de passage, the very logic of age-grades,and notions of descent and alliance all turn on the fact andpresence of children and youth. Even if children were siniplvadults-ln-the-making, this should invite rather than repelmainstream interest. Children are meaningful creatures invirtually all societies. Practices peculiar to their care and sus-tenance are indigenously recognized and frequently topicsfor conversation. Indeed, virtually even- cultural system takes"the child" to be a natural t,u t, even if the particulars of w hat


4/17

61 4 American Anthro polog ist Vo l. 104, No. 2 June 2002constitutes a child vary (Aries 1962). Notions of "child,""adult,'' "parent," and "offspring" are all cultural confec-tion s. It is hard to imag ine m ore preternaturally 'natural"everyday concepts. People everywhere and at all times havesome beliefs about what children are and what should be"done" with the m . Issues surrounding cultural and social im-maturity are of vital interest to t hose wh ose beliefs and prac-tices anthropologists closely attend. In conceiving of chil-dren as mere appendages to adult society, anthropology hasconceived of them as lacking inherent interest. This conveysthe situation, it does not explain it. Being an appendage doesnot mea n bein g of m arginal interest: An arm is an app endagebut it is no t of marginal interest to those w ho study limb s.

We need to seek anthropological resistance elsewhere.Anthropology's faint image of children reflects a more gen-eral tendency in American social science to view childhoodas a way-station on the path to the "complete, recognizable,and . . . most significantly, desirable" state of adultho od(Jenks 199 6:9). On this view , children are engaged primarilyin becoming "with perhaps some minor variations . . . whattheir elders already are" (Toren 1993:461). Studying the me-chan ics of ho w th is happ ens is the purview of specialists. Theunderlying image is that children, especially the very young,are radically distinct from and unequal to the adults aroundthem. Importantly for this discussion, they are located intransition to cultural competence rather than as havinggenuine mastery of it. As a consequence, discussion of chil-dren is typically transformed into talk about adults and theways they organize the environment in which children de-velop so as to facilitate th e acqu isition of the cultural comp e-tence appropriate to the society in which they live. Toren as-tutely observes that for the field generally discovering howchildren b ecom e "what their elders already are .. . [has] littleor no bearing on an analysis of the relations between adults'"(1993:461). Knowing about the relationship between adultsand their children accordingly provides little insight into re-lations between adults, the principal phenomenon of anthro-pological interest.

The body of research exploring the lives of childrenmost familiar to anthrop ologists is th e literature on socializa-tion that reflects this refraction of childhood through anadult lens. By focusing on the adult end-state and adult influ-ence on "achieving" it, children's activities are cast as ancil-lary or subordinate. As a consequ ence, the con tributions thatchildren make to their own development are often obscuredif not effaced.4 The rich literature on children's play andgames illustrates this well (e.g., Goldman 1998; Lancy 1996;Sutton-Smith 1976). Clearly these studies focus on childrenand, particularly, on routines and cultural forms that oftenhave neither clear parallel in adult activities nor Involve thedirect participation of adults. Play provides a wealth of Infor-mation about the scope of children's cognitive, cultural, andocial skills. Notwithstanding these Insights, these studies

adult activities andy th e ways In which rou tlnized play, Includingy (199 6) describes games and play as

enduring artifacts, a permanent part of a society's reper-toire, reused with each generation . . . sh elt er ed ... learn-ing opportunities. [In play] children couldrisk"getting itwrong" without serious consequences. At the same time,from society's vantage point, games are clever de-vicesthey are fun to play . . . and, thereby seduce thechild into learning things society thinks are important.[1996:94]Sutton-Smith's (1976) history of toys, for example, is a fasci-nating study of the way industrially produced toys serve totrain children to expect and enjoy the sort of solitary exist-ence that produces a specific kind of bou rgeois citizen.

I am not criticizing this work; as I said considerable in-sight is afforded by it. Indeed, I relied on it in the earlier dis-cussion of th e low cost of children's transgressions. The pointhere is that this particular focus on play and games dovetailswith a perspective that conceives of children and their activi-ties as functional correlates to adult society and adult goals.In so doing, children's activities are explored to the extentthat they contribute to adult outcomes; and importantly"ownership" of the means of cultural reproduction rest withadults. Socialization theory the idea that adult dispositionsare achieved largely through adult interventions in children'slivesthus obstructs the appreciation of the contributionchildren make to th e acqu isition of cultural sensibilities.

Perhaps, most regrettably, socialization theory oftenoverestimates the influence adults actually wield. Severalstudies have sho wn that ad ults frequently do a rather poorerjob of sh aping their children's cognition s, personalities, andattitudes than is often presumed (Harris 1998; for examplesfrom anthropology see Toren 1993 for an excellent review;see also Hirschfeld 1989a; Mead 1932). Grasping how re-markably good children are at acquiring culture and ho w diffi-cult ''cultural transmission" actually is are difficult whenadults are overly generously credited with teaching it Chil-dren do generally become much like their elders in criticalways. However, they do so in virtue of more than being "'so-cialized" into adulthood .s

For cultural reproduction, overly generous appraisal ofsocialization is not only a function of what researchers lookfor but also where they look. On a widely accepted view, theappropriate environment in which to study cultural trans-mission/acquisition is the one inhabited and controlled byadults, a strategy that makes sense if adults are the principlesocializing agents. Accordingly, if it is assumed that adultscreate the cultural worlds into which children are inductedand that adults largely control the processes by which thishappens, then attention to the adult world seems fitting.However, if the goal Is to understand how children contrib-ute to making culture, a more appropriate focus wou ld be thearena in which children do most of their culture making:nam ely, In their lives with oth er children, what is sometimescalled "children's culture.

The Idea that children have their own cultures may seemfar-fetched to som e readers. Children m ay be uncomfortablein adult culture and even, as they often seem to be in NorthAmerican and Northern European societies, resistant to it, theframe of descriptive reference is still the dominant cultural


5/17

Hirschfeld Why Don 't Anthropo logists Like Children? 61 5t radi tion. And the re is the rub. It is now a comm onp lace thatcultural environments are manifold. At any given momentthe cultural environment which an individual inhabits isfragmented, fluid, noisy, and negotiablefrom the perspec-tives of both information and power. All cultural environ-ments are themselves comprised of mult iple, contest ing,competing subcultural env ironm ents. Recognizing that chi l-dren's culture is one of these should pose no a priori diffi-culty. In fact, it has not for many scholars (Corsaro 1997;Goodwin 1990; the contributors to James and Prout 1990;Maltz and Borker 1986; Opie and Opie 1960; see also Eckert1989; Willis 1981). Child ren no t on ly live in the culturalspheres of the adults with whom they share a life spacealargely trivial obse rvation but they create an d m ainta in cu l-tural environments of their own. The cultural environmentin which cultural rep rod uctio n takes place is according ly n otnecessarily, nor even principally, the cultural environment(s)relevant to adults:

A child's goal is not to become a successful adult, anymore than a prisoner's goal is to become a successfulguard. A child's goal is to be a successful child.... Chil-dren are not incompetent members of adults' society;they are com petent members of their own society, whichhas its own standards and its own c ulture. Like prisoner'sculture and the Deaf culture, a children's culture is looselybased on the m ajority adult culture within w hich it exists.But it adapts the majority adult culture to its own pur-poses and it includes elements that are lacking in theadult culture. [Harris 1998:198-199]These observations are more than the recognit ion that

not all childr en's activities are fully ma tch ed in adu lt society.It is unco ntroversial tha t child ren participate in special-pur-pose cultural activities from wh ich adults are largely exclud-ed. Nor is it disputed tha t chi ldren develop an d m aintain so-cial practices, networks of relationships, and systems ofmeaning that are distinct to their own social and physicalpaces. The notion of a culture of childhood, however, cap-ures more than special-purpose child-focused activities.hi ldren's cul tures encom pass su bstantial and elaborated en-

the adult environ me nts in which th ey are embedd ed.

. . adults [who] kno w no thin g of the m. . From generat ion[sic] world and quite as little

r ow n cultural worlds.

create and sustain on their own have obscured how centralan anthropology of chi ldhood to put a nam e on the projec tI am promo t ingwould be to an y unders tanding of cul tura lreproduction. Theories of cul tural reproductionor, for thatmatter, theories of the disruption of cul tural reproductionare adequ ate only to the exten t that they are based on a real-istic account of the lives and forces that shape the lives ofprecisely the individuals who bear that reproduction. I havesuggested, however, that the broad mainstream(s) of anthro-pology are more than uninformed about chi ldren. Anthro-pology has displayed an enduring aversion to children, a re-sistance to an anthropology of chi ldhood. It is not simplytha t chi ldren do more than man y acknowledge . That they d oit so exceptionally well is a point of tension for many anthro-pologists.

It is uncontroversial that children are great learners gen-erally and talented students of culture in particular. Theyrapidly and readily acquire the ab ilities to slip seamlessly in tothe cultural l i fe around them. This seamless el ision into aspecific cultural existence is a function of some "deep" abili-ties. Few would be satisfied with an account of cultural acqui-si tion based on simply m imicry. Children d o no t ape culture,they learn or acquire i t . They come to represent cultural infor-mation, manipulate these representat ions, and use them asthe basis for making sense of the w orld and organizing act ionin i t . Representat ions, manipulat ions, and computat ions areinterna l to individuals and thus psychologica l phenomena.Anthropologyincluding the bulk of psychological anthro-pologyhas long resisted acknowledging let alone system-atically exploring internal mental states. It hardly seems anovers ta tement to say tha t anthrop ology s ince Durkheim hasdemonstrated a sustained aversion to things psychological(see Strauss and Q uin n 1997, chap ter 2, for a com preh ensiv etreatment; see also Bloch 1998; Hirschfeld, 2000a; Sperber1996).

Cast ing children's experience and development in termsof psychological phenomena is not the equivalent to claim-ing that cul ture is ideat ion, that i t can be reduced to ideat ion,or even that there is some subset of our knowledge that isspecifically cu ltural. It is to suggest th at p articip ating in a cul-tural environment means part icipat ing in a part icular set ofcausal cognit ive relat ionships through whose agency a cor-pus of knowledge is distributed (Sperber 1996). It is a funda-mental tenet of anthropology that cul tural forms pass fromone generat ion to the next through the agency of teachingand m odeling, bo th direct and indirect . As observed abo ve,this argument is implici t ly causalcertain parental routinescreate the condit ions for certain childhood experiences thatshape individual and collective identities and it identifies thepredominance of cul tural pract ices with their own reproduc-t ion.

There is no dou bt th at cul tural learning is empirical ly in-evitable. Humans are, from birth, cul tural creatures shapedby the cultural environments they inhabit . Culture so thor-oughly sa tura tes the envi ro nm ent tha t no t acquiring i t seemsalmost unth inka ble. The usually tacit assum ption is tha t byexposing an individual to a range of cul tural knowledge, the


6/17

61 6 American Anthro pologist Vo l. 104, No. 2 June 2002individual acquires a more or less faithful version. Implicithere is the further assumption that most cultural knowledgeis expressed in patterns of behavior, speech, and artifact thatthe learner eventually com es to recognize. Reflecting on howeasy college teaching would be if this were the case shouldsignal that this view is at the very least an incomplete ac-count. Even if this theory of learning were enhanced withsome constraints (e.g., that knowledge is acquired only if thepresentation of th e kn owledge is clear, informative, and rele-vant to the learner and that the learner is well motivated),the theory remains insufficient as anyone who has tried toteach an adolescent common courtesies can attest. Theresimply is no psychologically plausible account of how knowl-edge, particularly th e sort of abstract know ledge encom passedin cultural schem a, m odels, key sym bols, or regimes of truth,could be acquired from exposure in the absence of consider-able med iation o n t he part of the learner (Hirschfeld 200 1).

This may strike many as a curious claim. After all, cul-tural information, cultural knowledge, is conspicuous to thepoin t of extravagance. It virtually saturates the environ men t.It is both promiscuous and redundant in that every act, everypublic representation, has a cultural character, a cultural di-mension that indelibly marks it. That children would learnto be cultural actors, that they could readily induce culturalknowledge solely in virtue of living in a culturally rich envi-ronment, seems more than plausible. All the more so ontheories that cast cultures as bounded, relatively stable, andhomogeneous environments, populated with actors whoconsisten tly share interests and kn owledge. If cultures cul-tural environmentsare spatially discontinuous, frag-mented, fluid, contested, and ever transforming worlds, asincreasingly has been argued (Brightman 1995; Dirks et al.1994; Gupta and Ferguson 1997), then no matter how con-spicuous or culturally saturated these environments might

e , learning about them is no trivial task. Learning "that X"X were true,

nd places a common moral value on X would presumably

enge and exception. Learning "that X" under these condi-

ignoring what is not.Anthropologists have never credited children with the

view them to a greater or lesser extent, as passive objects,as helpless spectators In a pressing environment which af-fects and produces their every behaviour. They see thechild as continually assimilating, learning and respondingto the adult, having little autonomy, contributing nothingto stKidl values or behaviour except the latent outpour-Ings of earlier acquired experiences. 1197J.87)

knowledge actually is, encourages anthropologists to over-look the crucial contribu tion that children make to th e crea-tion, sustenance, and distribution of cultural forms (Sperberand Hirschfeld 1999). It emb oldens the field to see th e acqui-sition of knowledge as a straightforward process, or more ac-curately, as a straightforwardly simp ly p rocess. As several an-thropologists have recently observed, it is not (Bloch 1998;Strauss and Q uinn 1997; Wertsch 1998).

Acquiring cultural knowledge is an asymmetrical achieve-ment, not because the expert is an expert and the novice anovice, but because the child brings to bear specialized cog-nitive skills and domain-specific programs that make devel-opm ent possible (Hirschfeld and G elman 1994). In a sense,the novice is the expert: an expert at learning. Without thesingular architecture of children's minds culture would beimpossible. This claim is more th an the throw-away acknow-ledgment that humans have individual minds and thereforesomething interestingbut largely irrelevant to anthropol-ogycould be said of the mind. I intend a much strongerclaim, namely, that culture cannot be understood except interms of the cognitive architecture of children and the spe-cialized learning mechanisms that the architecture affords(Hirschfeld 1996; 1997).

1 turn n ow to tw o specific cases with w hich I illustrate theprocesses that I have described: namely, a distinct children'scultural tradition, a similar adult tradition, and the cognitivesusceptibilities and competences that shape the relationshipbetween the two. Both traditions are cultural confections ofAmerican society (although both have counterparts in a widerange of cultures). The two formspreadolescent cootie loreand "adult" construal of raceat first blush surely seem in-comm ensurate. I will try to establish that they are n o t Inparticular, I seek to show how the operation of a specializedlearning program, deployed in the context of a specific cul-tural environment, creates the conditions by which childrencreate and sustain a "simple" game and by which adults or-ganize and sustain fundamental access to power, authority,and resources. Because I have written about racial thinking atlength elsewhere (Hirschfeld 1989a, 1996, 1997, 2000), mydiscussion of it here will be brief. The m aterial on coo ties hasnot b een reported, and I will present it here in more detail.Over the past several years, my collaboratorsSusanGelman, Rachel Heiman, Gail Heyman, Katie Hinds, BarbaraHofer, Oren Kosansky, Ivelisse Martinez, and Heidi Schwein-gruberand I have investigated a constellation of practicesand beliefs of North American children about "cooties"(Martinez et al. 1999). We observed children in 2nd and 4thgrades during free play, Interviewed children in small groupsand Individually at school, and eventually asked anothergroup of kindergarten, 2nd and 4th grade children to partici-pate In several experimental tasks. Children were drawnfrom two different areas. The first consisted of schools In andnear a Midwestern college town whose catchment is largelywhite and middle class. The second consisted of schools


7/17

Hirschfeld Why Don 't Anthropo logists Like Children? 6 1 7about 50 miles away, whose catchment was rural and smalltown and largely working class. About half the students inthese schools are white.In the most general terms, cooties are a social contami-nant that pass from one child to another, a form of interper-sonal pollution. According to one source, the term cootie is atransformation of a British colonial word for lice that waspopularized by returning World War I veterans (Samuelson1980). Consistent with this interpretation, The Oxford EnglishDictionary defines cootie as body louse and suggests that itmay be derived from the Malay word kutu meaning "a para-sitic biting insect.'" Contemporary North American chil-dren's usage sometimes follows this gloss. More frequently,however, children describe cooties as something that cannotbe seen but that is disagreeable (a fair number of childrenused one of three tropes in describing cooties: the invisibleparticulates associated with germs, farts, or "boogers").Adults, presumably "recalibrating" memories of their ownchildhoods, sometimes use cooties and lice interchangeably.One author entitled an article "Cootie Control" th at advisedparents on how best to treat head lice infections despite thefact there was no men tion of cooties in th e body of th e text(Nathanson 1997).

The overwhelming association with "cooties," however,for both adults and children in North America, is attributionsby children of an invisible contamination that passes fromone child (often a member of a stigmatized group) to anotherand the prophylactic routines used against cootie contami-nation. There are a number of these prophylactic routines,including "cootie catchers" fashioned out of paper, pretendinjections, and special ways of crossing fingers or hands: "ifyou cross your hands you won't get them. If you get on apole you won't get them "; "if you do n't like person and youtouch them, you can get cooties unless you cross your fin-gers"; or "if cooties are in your house, you lock up all of thedoors and turn off all the lights."There has been little scholarly work on cooties. The twonoteworthy studies are by Samuelson (1980) and an ex-tended discussion in Thome's (1993) Gender Play. There aremany more fictional treatments. Several adult-authored nov-ls, some targeting adolescents but some meant for adultudiences, have used cootie lore and its practices to convey a

sense of deep and inextricable social taint, dread, and exclu-sion (Hayter 1997; Holub 1998). On the whole, however,ties, enacted by ch ildren, regulated by children, and expe-

is exemplary.Cooties are invisible; cootie practices are not. The para-

a social category: "If1 say Rhhhhhl"; "If you touch some-

and the boys don 't want girl cooties . . . the boys run fromthe girls.'These activities may sound, as they sometimes actuallyare, like the hyperbolic, physicalized, and playful rehearsalsof routines of gender avoidance common among preadoles-cents. Cooties attributions, however, are often more seriousand the sometimes playful quality of the "game" not very ef-fectively masks much deeper emotion.An interaction that occurred during an earlier phase ofour collaborative ethnography of children's culture illustratesthis vividly. One of the lead researchers on the project,Ivelisse Martinez, witnessed the force of cootie fear firsthand. Martinez was chatting with a group of nine-year-oldchildren in a classroom (not about cooties) when a girl ap-proached th e group and sat in an available chair. Almost im-mediately and quite suddenly she became demonstrably up-set. Martinez asked her w hat was wrong. The girl breathlesslyreplied that she just realized that the last child to use thechair had cooties. The remark was not a function of the dis-cussion nor provoked by anything else Martinez observed.She was convinced that the girl's reaction was neither pre-tense nor fanciful. It was an expression of actual fear of pollu-tion.

In one respect the event was exceptional. Characteristi-cally, cooties involve attributions of pollution or the dangerof pollution that occur when both the cooties contaminatorand potential contaminatees are actually in the midst of in-teracting. Attributions tend not to be delayed. Despite this"real-time" quality of cootie attribution, there is also a char-acteristic uncertainty about cooties. A child not only neverknows when he or she will "get" them, but also never knowswhen he or she might be accused of "giving" them. Unlike,for instance, other cultural forms in which social contam ina-tion occurscaste pollution in South Asia seems an appro-priate point of comparisoncooties contamination is not aconstant threat for which regular, routinized protections areavailable. Nor are there specific circumstances in which coo-ties threat is always present or even specific to a particularperson or a particular groupas there is for menstruatingwomen in some societies. In short, there are neither specificcontexts nor a class of persons for whom or from whom therisk of cooties contamination is invariably present. The sameperson can be the source of cooties at one moment but be-nign the next. Indeed, some of the "excitement" of the coo-tie phenomenon seems to be derived from this lack of pre-dictability.

Given the association with disgust and pollution, notsurprisingly, cooties also serve as an offensive "weapon , aswhen one child tries to contaminate another in "cootie tag"(Samuelson 1980; Thome 1993). Opie and Opie (1969:75-76)classify a British variant ("the Dreaded lnrgi") under the ru-bric "the Touch having Noxious Effects." in their section on"Chasing Games." Although attribution ot cootie contami-nation Is often gamellke and playful, they can, as observedabove, be both strategic and serious. They are frequently usedin a familiar and cruel scheme for excluding unpopular, new.or otherwise stigmatized children. The novels 1 mentioned


8/17

61 8 American Anthropologist Vol. 104, No. 2 June 2002earlier convey this aspect dramatically. In one, entitled IvyGreen, Cootie Queen, and meant for early grade school-agedchildren, Holub (1998) tells the story of a youn g girl stigma-tized by a rumor circulated by the "popular" girls that shehas coot ies. More dramaticat least for adultsis a novel in-tended for an adult audience called Revenge of the Cootie Girls,in which Hayter (1997) vividly describes the lasting traumasocial exclusion in grade school by having been labeled a"cootie girl.' The force of this is underscored when she ex-plains to an uncomprehending Japanese friend that a cootiegirl is a social pariah.

The sense of genuine emotion evoked by cootiesandthe enduring memory of that emotionwas intriguinglyconveyed in a piece in the New York Times during the 2000presidential primary season. The article, published in theSunday Week in Review section, illustrated the growing at-mosphere of tension between the two leading Democraticcandidates with the following description: "The vice presi-dent has gone on referring to Mr. Bradley as his friendatleast un til their chat o n 'Meet the Press,' wh en Mr. Gore putou t his han d . . and Mr. Bradley stared at it with a look thatsaid, 'Ewww, coot ies' " (Henneberger 2000 ).

Simply mentioning cooties to undergraduates in mycourses triggers a flood of memories and invariably provokesa collective spasm of squirms and nervous laughter. Like im-migrants, whose memories of childhood often have a specialclarity because they evoke not only a particular moment intheir lives but also particular cultural and often sensory expe-riences (Hoffman 1994), North American adults have littledifficulty instantaneously retrieving memories of cootiespractices and re-experiencing th e affect inscribed in them .

We found, paradoxically, that despite cooties' instru-mental force and lasting salience, children give surprisinglyvague answers when we asked what cooties actually are, apattern that Sam uelson (1 980) also reports. Cootie lore is notconceptually orderly. In response to our questions, one childwou ld use language appropriate to describing familiar but in-visible particulates like germs ("cooties is like germs, it hasgerms on it."; "cooties can give you germs.' "They give youbad germs that can kill you.") or those associated with viola-tions of personal hygiene ("usually if one of my friends usesthe bathroom on themselves they have cooties''; "whensomeone digs up their nose or eats their boogers"; "cooties

re where somebody licks the bottom of the chair or eats pa-per"; "when you are by somebody and they fart maybe thatan give you cooties"). At other times, children use language

more typical to descriptions of ephemera like bad associa-ions or misbehavior ("if you don't like a person and you

them , you can get cooties"; "people who have cootiesle, and they beat up girls. They are bad"; "usually

somebod y does someth ing nasty"). Notnfrequently, children use mixed descriptions, particularly

ou cou ldn't see them"; "1 don't like them .

'Cause it's (ust a game that I don't play. I used to, but don'tanym ore. I really, really don't like it's gross. . . All differentkinds of p eople get it."

These vague and varying statements do not mean thatcoo tie lore is inco nsta nt, rather, as I said earlier, that it is un-tidy. One measure of this is that degree to which the same"untidiness" is distributed over time and space. Thome(1993) observes that the same pattern of practices and beliefsis found in communities in California (mixed Chicano/Lat-ino/Anglo), in Michigan, and on the East Coast. Similarly,our interviews, like those conducted by Samuelson (1980),reveal considerable cons istency b etween adult reminiscencesand contemporary children's routines. Perhaps most strik-ingly is a recurrent pattern in children's "games" across cul-tures. Samuelson (1980) reports versions of cooties in Eng-land, Spain, Madagascar, and New Zealand. Opie and Opie(1969:75-78) recount a British game called "The DreadedLurgi," whose name derived either from a then-contempo-rary radio show or, and intriguingly, from a supposed ail-ment "found" in East Anglia from which "the idle" suffer.They note , h owever, that basic form h as existed "generationafter genera tion." Lurgy tag (East Anglian spelling) in volves akind of tag in wh ich the chaser through skin-to-skin contracttransmits "something evil or sickening." like cootie attribu-tions, Lurgi tag is not always playful. One headmistress toldof discovering the game w hen a student cam e crying because"everybody said she had 'it'," a troubling complaint, theOpies report, which had been reported in "a number ofschools, large and small."

As with cooties, lurgy picks out and highlights negativesentiments toward stigmatized children: "In Norwich wefound the word 'lurgy' in everyday use amongst children'You're lurgy'but with the restricted inference that the per-son was 'stupid, goofy, looney, nuts, [or] a nit'." Similar rou-tines are found throughout Great Britain and "such gamesseem to be played around th e world. In Auckland, New Zea-land, when a boy is tagged by a girl, the others deride himshouting 'You've got girl fleas' " (Opie and Opie 1969:77).The Opies describe similar forms outside the United King-dom:

In Valencia the ordinary game of chase is Tu portes lapusa' (You carry the lea).At Massa in the Bay of Naples,the game is 'Peste'. And In Madaga scar... the child whodid the chasing was boka, a leper, and when he touchedsomeone his leprosy was conveyed to the one he touched,who in turn rid himself of the disease on someone else.[1969:77]A comparable children's cultural form exists in Japan.

Engacho are prophylactic routines tha t Japanese children useagainst a form of social contagion that shares a number offeatures with the forms discussed above. A typical situationIn which an engacho routine would be enacted would startwh en a child Is polluted by being soiled, say through contactwith dog feces (or who might otherwise be stigmatized, pos-sibly In virtue of being an ethnic minority). Successive con-tamination to other children, however, is not through contactwith the Initial polluting substance, but in virtue of it. Thus,


9/17

Hirschfeld Why Don 't Anthropo logists Like Children? 6 1 9as with coo ties and several of the o ther forms just discussed,social contamination results from contact with an invisible,essentially abstract, substance that transfers from one childto another. To prevent co ntam ination or remove i t, engach oroutines are enacted, which typically involve finger gesturesalso reminiscent of cootie lore. As in the United States, con-temporary authors have used engacho to evoke nostalgicmemories of childhood in adults. The material presentedhere was provided by Yu Niiya,6 who in addit ion to drawingon her own childhood memories, discovered an internetchat room devoted specifically to engacho. In it there was adiscussion, which she reported in an email to me, of a recentanimated film entitled:

Sen to Chihiro no Kami Kakushi (Sen and Chihiro SpiritedAway) by Hayao M iyazaki, [which] depicted a scene of En-gacho. Sen (the heroine, a 10-year old girl) goes to a won-derland and kills a little evil monster by stepping on it.Then she makes two crossedringswith her thum b and in-dex and says "Engacho!" The old man sees it and cuts thering with his palm, saying "Engacho kitta." And Sen fi-nally feels relieved, [personal communication]

The debate appa rently be gan d urin g a discussion of regionalvariat ion in engacho routines and the possibi l i ty that someviewers mig ht n ot b e familiar w ith it.

The point of these various and varied examples is notthat cootie lore, eng acho , th e D readed Lurgi, etc. are versionsof each other; nor are they expressions of some universal de-velopmental moment that children everywhere pass through.Later I will suggest that to understand these cultural forms isto understand specific aspects of children's conceptual archi-tecture. These cultural forms, nevertheless, are massively un-derdetermined by that architecture. They are literally themind-making contact with public representat ions.

For the present discussion, however, what is relevant isthat al l these "games," rou tines, and othe r cul tural forms arewhat Opie and Opie (1960) call children's unselfconsciousculture, reproduced w ithou t adult interv ention. For exam ple,to return to cooties in particular, while adults raised in theappropriate culture readily recognize and remember cootielore, children learn cootie lore only from other chi ldren and

nact it only w ith oth er chi ldren . Children give cooties o nlyto each other and get them only from each other.

The tethering of cooties to the con ceptual , physical , andelational space of childre n creates possibilities th at o ften re-ain obscured. Although cooties are extremely evocative for

t relations of power am on g children. Cooties mark and po-

social contaminat ion tha t regu-

Although cooties lack the systematici ty and apparent inde-pendence of context that are characteristics of systems of ra-cial and caste thinking, there are crucial similarities bothwith respect to beliefs and the use of those beliefs in the serv-ice of systems of power and authori ty. To repeat this impor-tant point : Cooties are about power and authori ty within chil-dren's culture. Cooties are used to establish and maintainunequal social relations between children. At the very leastthey are a means of signaling and ul t imately enforcing suchrelations. As Th om e notes "recoil ing from physical pro xim-ity with another person and their belongings because theyare perceived as [having cooties] is a powerful statement ofsocial distance and claimed superiority" (1993:75).

This power derives from at least two aspects of cootielore; the untidiness or lack of systematicity in cootie attribu-t ions and the putat ive nature of cooties themselves. Attribu-t ions of cootie contamination are typical ly unpredictable,one never knows when, in what context , or by and aboutwh om they wil l be mad e. "[Cooties] com e from differentpeople. Any kind of person can have i t, but I do n't kno wwhat makes you have i t . ' ' In one sense, then, a cooties at t ri -butio n at any given m om en t calls at tent io n t o exist ing socialdistance rather than creates it .7

In this respect, cooties functions much like race andcaste, and does so presumably in virtue of being at tributedwith a singular "natu ral" natu re. A central elem ent of racialand caste think ing is natural ized difference an d th e use m adeof it to police social difference. Part of the reason that cross-racial and cross-caste contact is thought to be pollut ing isthat such contact supposedly violates or disrupts a natural or-der. Children share these beliefs. Even preschool-age chil-dren apply essentialist reasoning to racial and caste differ-ences and they expect both racial and caste groups toreproduce in much the same way tha t other na tura l thingsdo, part icularly nonhuman l iving species (Hirschfeld 1996;Mahalingam 1999; Springer 1996). One possible explanationfor the frequent association of cooties with germs and lice isthat cooties too represents a naturalized (as distinct from anatural) image of the world. That is to say, cooties evokegerms and lice not because of empirical parallels between coo-t ies and small vermin but because of a conceptual parallel.Germs and l ice are biological or natural phenomena. Theconceptual paral lel suggests that juvenile modes of estab-lishing and signaling social distance are naturalized, just asare adult forms tend to invest important value-laden aspectsof sociality with a natura l basis.

Finally, a particular kind of social relation is policed andsignaled through cooties at t ributions. Cootie routines reflectbeliefs abou t group contamination, especially gender (Powlishta1995; Thome 1993). Gender, however, is not the only inter-group relat ion expressed (or publicly commented on)throu gh cooties. As already ob served, cootie at t ributio ns fre-quently pick out individuals st igmatized because of theircharacter or appearance (e.g. , chi ldren who misbehave orwho are overweight). Cooties are also l inked to st igmatizedgroups , as one boy mad e unst ' l t con sdo usly c lear : They|cooties| are grovs, people run I mm people wh en som eon e farts.


10/17

6 20 American Anthro pologist Vo l. 104, No. 2 June 2002or licks their toes, or dig up and eat their boogers. Africanpeople, Panama people. Girls have them more than boys.*'(The intersection of cooties, intergroup relations, and preju-dice was not missed by at least one (adult) playwright. InMarch 2001, the San Francisco Bay Area Discovery Museumpresented an interact ive play meant "to promote toleranceand diversi ty" ent i t led Cootie Shots: Theatrical Inoculationsagainst Bigotry.)

One explanation for the paral lels between cootie loreand race, caste, and gender on the one hand, and germs orlice on the other, is that cootie lore is a version of these otherbelief systems. Cooties could simply be an analogue by trans-fer of propert ies whose "base ' ' resides in another domain,one largely structured (supposedly) in and by adult culture.Cooties, on this view, would be a transfer, perhaps even a de-graded transfer, of relations and justifications for the waypower and authori ty are distributed and regulated in thedominant , adult cul tural t radi t ion. Cooties might , thus, be ajuvenile version of dominant cul ture, much in the way so-cialization theory anticipates that children's cultural formsare way-stat ions on the pathways to adulthood.To explore this possibilityand to ultimately rejectitwe need to look at how closely cootie lore and theseother systems of relations are structured. Before doing this,however, we need to rule out an even more straightforwardpossibility: cooties may simply be a way of signaling friend-ship (and lack thereof).

Friendship and Enmity: In order to consider this possi-bility, we asked a group of children to imagine the followingscenario (a dapted from a task used by Rozin et al. 1994):

You are going to another school to sing with your classand your teacher tells you to get a class shirt, you know,one that says the name of your school and your grade.You need to get it from a pile of shirts in a corner of yourclassroom. Your teacher wants you to get ready. You hurryup and grab a shirt, you put it on, and go join the groupfrom your class.

Half the children participated in a positive scenario:Think of a person, a kid you know, that would make youhappy and make you feel good if you found out tha t he orshe had worn the same shirt that you picked and put onbefore you wore it.

Half the children participated in a negative scenario:Think of a person, a kid you know, that would make youunhappy and make you feel bad and upset if you foundout that he or she had worn the same shirt that youpicked and put on before you wore it.

We then asked the child to rate how likely was each of a listof explanation s for their discomfort (at discovering that th eyhad worn a piece of clothing last used by a desirable or anundesirable ch i ld). We told th e child that th e l ist of explana-t ions was collected from the remarks that chi ldren had givenus when w e coniiucted the same experiment at another school(the list was In fact drawn from remarks children had made

uring unstructured play that we had observed during the ear-ier, ethnographic phase of the research). Each explanationeflected one of several modes of contamination: contamina-

t ion by associat ion ("made me think about how much Ididn' t l ike that person"), social contamination ("other kidsmight know who wore i t before"), contamination because ofcharacteristics of shirt ("it wouldn't fit right after the personwore it"), or contamination because of particulate transfer("it would have germs or tiny little things that would makeme l ike that p erson"). Children rated contam inatio n by asso-ciat ion, social contam ination , an d co ntam ination because ofcharacteristics of shirt as plausible explanations. There wereno reliable differences across ratings of the three explana-t ions. Contamination involving part iculate t ransfer/ t rans-mission, how ever, was rated less plausible, essentially no bet-te r than chance .

Following the first set of questions, we told each childthat chi ldren at another school sometimes brought up coo-ties in their explanations of their discomfort in wearing thesweat shirt of the undesirable chi ld. We told them that wewanted to take the opportunity of these one-on-one si tu-ations to tell us what they knew about cooties ("What canyou tell me about cooties?"). Their responses were coded intoseveral broad categories.In marked contrast to the explanations they gave fortheir discomfort at wearing the undesirable child's sweatshirt, children w ere significantly mo re like-ly to bring u p par-t iculate t ransfer/contamination when talking about cooties.The greatest portion, 31 percent, of their descriptions men-tioned the transfer of some substance (body excretions,germs, disease, etc.), 17 percent pretense (cooties is just agame), and 11 percent social (social categories or relation-ships). Thus, unlike their explanations of "simple* socialcontaminationthat emphasized associat ive thinking andsocial relationships cooties evoke a sense of p articulate pol-lut ion that conformed to a natural ized mo del .

These data help situate cootie lore relative to simple so-cial likes an d dislikes. Clearly coo ties can b e used as a way tocall at tent ion to desirable and undesirable chi ldren and to re-hearse notions of social distance and hierarchy. We knowfrom oth er stud ies that stigmatized g rou ps are likely to be as-sociated with cooties (Thome 1993). In addition to askingchildren to explain their comfort /discomfort with wearingthe sw eat shirt, we also asked childre n to tell us about th e de-sirable/undesirable child (what they looked like, how thechild's family was similar to and different from their family,whether their parents would be upset i f the child played atthe desirable/undesirable child's house, etc.). Children reli-ably associated desirability/undesirability to social status,gender, and race. Children enjoy contact with children whoare socially like them but are uncomfortable with even indi-rect contact w ith child ren w ho differ from t he m in race, so-cial class, an d g ender . There is, how ever, a n important differ-ence between simple deslrability/undesirability and cootieattributions, a difference that indicates that cooties is a sin-gular cultural form. C ootie conta min ation is l inked to thetransfer of particulate but invisible material, whereas simplelikes and dislikes are no t.

Germs: This then raises the possibility discussed earlierPerhaps cootie lore Is just a version of childr en's beliefs about


11/17

Hirschfeld Why Don 't Anthropologists Like Children? 6 2 1another naturalized particulate transfer, germs. There areobvious parallels between children's beliefs about cootiesand children's and adults' beliefs about material contami-nants. The wealth of children's routines with medical-likenames, such as "cootie inoculation,' "cootie vaccination,""cootie spray," and "cootie immunization" (Samuelson1980; Thorne 1993) suggest that children conceptually linkcooties and germs and infection and hence that cooties mayrehearse children's understanding of these phenomena, asseveral of the quo tes above suggest. We should, nonetheless,resist drawing too close a connection between cootie lore andthese adult cultural and medical forms. Cooties may evokenotions of germs and lice but there are significant differencesbetween cootie lore and children's naive beliefs about dis-ease. For one thing, as the quotes above also show, cootiesare thought to be like germs, or to increase susceptibility togerms. They are generally not claimed to be germs.

There is another difference between children's theoriesof cooties and germs; namely, their respective causal rela -tions to other events in the world. While questions remainabout how best to interpret children's beliefs about germs,there is a broad consensus that from preschool age on chil-dren understand that exposure to germs is causally linked tosubsequent symptoms and illness (Kalish 1996; Kister andPatterson 1980; Siegal 1988; Sigelman et al. 1993; Solomonand Cassimatis 1999; Springer and Ruckel 1992). The basicmental model is cause and effect: germs cause illnesses, likecolds; lice cause skin discomfort, like itching and rashes. Pre-school children also understand that invisible particles (e.g.,sugar dissolved in water) have causal properties (e.g., theyproduce a sweet taste) (Au et al. 1993).

Unlike germs, however, children do not believe thatcooties cause anyt h i ng . Getting, passing on, an d getting rid ofcooties is central to cootie lore; what happens materially afterthis is not, as the following remark shows: "Or someonemight do something really disgusting. It's when you dig upyour nose and you touch somebody, and you have to crossour fingers.' Even preschool children understand tha trossing fingers is not an effective way of avoiding germ-bomnfections. Rather a cootie routine is about bracketing anding social interac tion ("They are not nice. . If they

recall the version from Madagascar describedOpie and Opie in w hich a child w ith "leprosy" rids him-theirabout germs and other contaminants, we would

else.Race: Race is a social forceand hence a cultural fo rm -

dren understand race and the way adults do. To explore therelationship between cootie lore and children's conceptionsof race, we need to understand the relationship between chil-dren and adult conceptions of race. A bare-bones descriptionof "adult" racial thinking in North America would includethe following three interrelated folk propositions. First, hu-man beings can be exhaustively partitioned into distincttypes based on their concrete, observable constitution. "Ob-servable constitution" means that racial types are embodied,natural, and enduring. Second, membership in a particulartype carries with it nonobvious and inner qualities as well asoutward ones. Third, the first and second propositions arelinked by a causal theory that posits that individuals haveboth the observable and nonobvious qualities of a particularracial type in virtue of having a particular racial essence. Thatthis conceptual edifice reflects the contingencies of the regu-lation of power and authority, which the system of racialthought serves. On this argum ent, race is about relations ofpower, power is about aggregate structural (political, eco-nomic, or cultural) relations no t m ental ones, hence th e con-cept of race is a function of existing structural relations (seeHirschfeld 1996).

This account is not a theory about the ontogenesis of ra-cial cognitions but a theory about its social construction. Thewidely accepted developmental account is even simpler (seeAboud 1988; Katz 1983). Children form racial categories byopening their eyes and looking. A slightly more subtle ver-sion might hold that in virtue of the contingencies of powerand authority, adults draw children's attention to race. Thenchildren w ould open their eyes and discover it. The idea tha trace's perceptual prominence contributes to its precociouslearning is consistent w ith a long-held view that children areconcrete thinkers, tethered to surface appearances. Youngchildren readily learn to distinguish race in much the sameway that they learn about other perceptually salient humanpropertysay, being tail or blond . The reason that childrenlearn about being tall, blond, or black before they learn aboutbeing Republican or French is because blacks and tall andblond people are more visually prominent tha n RepubUcansor French folk (Aboud 1988). The claim about perceptual sal-iency is in turn linked to a second claim about conceptualproperties. According to this argument, children n ot only at-tend to surface differences, they interpret these differences tobe superficial. For the young child, being black is initially nomore important than being tall, blond, thin, or brown eyed.Thus, while adept at forming racial categories, children aresupposed torelyon superficial appearance and when reasoningabout race they ascribe superficial and modifiable properties.8

Given how widely held these claims are, it is somewhatsurprising that so little research has actually tested the m. Theidea that perception drives children's racial cognitions is anassumption not an empirical discovery inference. For a vari-ety of reasonsnot surprising to anthropologists1 haddoubts about a realist in terpretation of race. 1 conducted aseries of studies to see if children really were as superficial intheir thinking.* Against received wisdom, 1 found that northernKuropean children's initial concept of race contains almost


12/17

62 2 American Anthrop ologist Vo l. 104, No. 2 June 2002no perceptual information. What little visual informationthere was typically was inaccurate and idiosyncratic (Hirschfeld1993).I also undertook several studies that explored youngchildren's reasoning about race. Previous research concludedthat young children expect that a person's race will change iftheir surface appearance changes. The work was based onasking children to reason about abrupt and unfamiliarchanges in appearance. To see whether the abruptness or un-familiarity of the tasks affected their judgments, I askedAmerican preschoolers about changes in appearance withwhich they were familiar; those involved in growth and fam-ily resemblance. Children know that as things grow their ap-pearances change. They also know that children resembletheir parents in some ways but not others. Using this knowl-edge as a base, I asked preschoolers whether certain changeswere possible. Even three year olds demonstrated that theyunderstand that a person's race will not change as she growsolder. In contrast to their beliefs about racial constancy, how-ever, three year olds anticipate that other aspects of physicalappearance, like a person's body type (hefty versus wiry), canchange during a person's lifetime. Similarly, young childrenbelieve that a child and his parents are necessarily of thesame race, although they did not believe that a child and hisparents would necessarily have the same body build.10

These findings demonstrate that children, in construct-ing and interpreting racial categories, pace the received view,go well beyond the information given. Even young childrenhave a sophisticated and adult-like cultural understanding ofrace, favoring a biological, abstract, and essentialist interpre-tation according to which race is conceptualized as immuta-ble, linked to family background, and diagnostic of groupidentity. In short, even very young children's conceptual rep-resentation of race is very much like the one of the adultswith whom she lives. In contrast to previous research thatclaimed the young children have only a superficial grasp ofrace, these (and other) studies evince a deep, theorylike,adult concept of race.Both race and cooties share a singular conceptual fea-ture, a naturalized interpretation of group differences, and asingular social feature, th e use of supposed naturalized differ-ences to signal and regulate differences in power. This latterfeature is uncontroverslal with respect to race. A naturalized

vision of race has frequently been linked to the contingen-cies of power and authority (Guillaumin 1980). Cooties, wesaw earlier, functions in a similar wayalthough the powerstakes may seem modest to adults, they are not to the chil-dren who experience them (the jacket cover for Ivy Green,Cootie Queen asks "Is Ivy's Life Ruined?"). Cooties of coursedo not create group-based social distinctions. Neither do coo-ties create the exclusionary practices that cause social rela-tions to be regulated through them . What cooties do Is natu-ralize the distinctions; and In this respect, cooties aredeployed In children's culture the way ran- Is deployed In theregulation of relations (>t power and authority In adult stxiety.The c(xitie-is-to-chlldn/n as race-ls-to-adnlts analogyside, the social and mental representations of race and coo-

ties change in the course of childhood. Consider the waythat preschoolers conceptualize the two social groupings.Susan Gelman, Oren Kosansky, and I found that kindergart-ners are familiar with cooties and use them to mark and regu-late social relations. Unlike school-age children, however,they do not expect cootie contaminations to involve thenaturalized transfer of particulate substance. With race, theinverse pattern obtains. Preschool-age children naturalizerace, see it as involving the transfer of a nonobvious essencefrom parents to children. However, race does not regulate so-cial relations or create social distance for preschoolers, as itdoes for slightly older children . The race of a child does notpredict with whom preschoolers will be friends. Thus, al-though cooties and race resonate, they do not follow thesame developmental course. Nor do they appear to be ver-sions of each o ther.

What then is the relationship between these various culturalforms and why do they so clearly resonate one to the other? Ioffer the following explanation. Cooties, racial, caste, andgender systems of classification are all expressions of a singlemechanism for conceptualizing human group differences.Cootie lore is not a juvenile, children's culture version ofthese other "more mature" social systems of categorization.All these systems are products of the same conceptual im-pulse to categorize played out in the cu lture of preadolescentAmerican children. Each is an independent instantiation of asingle dimension of cognition as it comes in contact withdistinct dimensions of the cultural environment. One doesnot cause the other to come into existence nor does it pro-vide a venue for practicing the manipulation of the structuralrelations inherent in the other. Rather, all emerge from acommon willingness or cognitive disposition to imagine theworld in a very specific way.

Elsewhere I have described what this very specific way is.I argue that these systems are products of a cognitive devicethat guides the development and elaboration of humankinds (Hirschfeld 1996). By this I mean that humans are en-dowed w ith a conceptual device or m odule specifically dedi-cated to processing information and guiding inference abouthuman collectivities.11 Virtually from birth, humans evincethe operation of this device. Infants discriminate between in-dividuals as members of different genders and language com-munities (Hirschfeld 1989a; Hirschfeld and Gelman 1997).From toddlerhood, and perhaps earlier, they sort humansinto kin-based, occupational, and racial groupings (Hirsch-feld 1989b, 1996). Moreover, they use these groupings to pre-dict nonobvious properties of their mem bers.

The existence of a cognitive device or conceptual modulefor reasoning about human groups would not be surprising.Much human activity is regulated by and mediated throughsocial collectivities. Yet human collectivities are not as easyto detect as, say, the social groupings of other species. Beesswarm, large social mammals move in herds, 8sh coalesce intoschools, birds migrate in flocks. We can set' these collectivities.


13/17

Hirschfeld Why Don 't Anthropo logists Like Children? 62 3Human social groups seldom present themselves in suchdear forms and not surprisingly researchers have found thateven functional groups like coalitions are difficult to detectperceptually (Stanton and Morris 1987). Most of our infor-mation about social groups is narrative not visual, yet, as dis-cussed earlier, the standard view is that children learn aboutthem by looking. The result is that there is a prevalent under-estimation in both anthropology and psychology of how dif-ficult it is to acquire a culturally-appropriate understandingof social groups (Hirschfeld 2001).

Still, it is obviously useful for people, including children,to know what groups there are in the social universe inwhich they operate. One of the ways that we come to knowwhat groups there are is to identify what their members arelike. Like adults, children recognize that some groups aremore important than others. Also like adults, relative impor-tance derives in part from the kinds of people there are, theirputative natures. A recurrent feature of thinking about groupsis that some ways of organizing people are more "natural"than others. This is evidently the case with gender, kingroups, age-grades, race, caste, etc. Conceptually, adults tendto naturalize these groups. In triguingly these are also the firsthuman groups that young children learn about and thegroups that young children are most likely to naturalize(Hirschfeld 1988, 1989b, 1995). Of course, children mightnaturalize these groups simply because adults do. But, as Ihave tried to show, the developmental evidence does notsupport this conclusion. Rather, children come to naturalizethese groups on their own initiative.

That children do this on their own initiative has impor-tant consequences for understanding adult beliefs as well.Children's representations , I argue, are grounded in a special-ized cognitive device for guiding learning about socialy children acquire knowledge about the world, it or-

This is not a logical necessity, of course. Adults could,

is, the more likely a faithful version of

of conventional wisdom. Children do not become whoare. Rather their elders become w hat th e child orspecifically w hat the architecture of the child's m i n d -emology of representations, on which this argum ent

Gender, kin, age, race, caste, and other naturalized

bound with the distribution of power and authority. This isnot a coincidence. According to conventional wisdom, chil-dren come to naturalize these particular groups because of thepolitical economic salience they have in the wider (adult) so-ciety. That is to say, according to conven tional wisdom, theyare early emerging categories because they are so sociallyconspicuous. There is a puzzling aspect to this account. Al-though, to my knowledge, no one has acknowledged it, thenotion that these categories are socially conspicuous is curi-ously coincidental with the notion that gender, kin, age,race, and other naturalized groupings are learned early be-cause of their marked perceptual correlates.12 On this argu-ment, as discussed earlier, children learn to recognize thesecategories because they are so physically conspicuous. As I ar-gued earlier, the idea tha t these categories are early-emergingconcepts because of perceptual salience is mistaken.

Race is important to the organization of power andauthority because of a conceptual proclivity children have tonaturalize categories like race. A specialized learning mecha-nism that renders social naturalization an easy "achievement"by making naturalization easily thinkab le is a precondition forthese processes of ideological distortion, explanation, andjustification. In a critical sense, power is racialized becausechildren find naturalizing social groups cognitively easy. Atleast in the case of race, the fact that the concept plays a fun-damental role in organizing power and authority is not thecause of its cognitive properties but derives from them. Racialpolitics sustain themselves not simply because they serveand define relations of power and authority, they are sus-tained because children make them easy to think (seeHirschfeld [1997] for a detailed discussion of this argum ent).

It is crucial tha t I make this poin t d early. I am no t sug-gesting that race is an innate concept or that children wouldacquire it without the support of the cultural emironment.Race is not an inna te concept. Indeed, given that it is histori-cally a relatively recent concept, it is difficult to see how itcould be. In the absence of a cultural environment in whichrace is a fundamental dimension of political, econom ic, andcultural organization, children would not learn it. This maybegin to sound like the view I am challenging, because it ap-pears that I am granting race's precocious acquisition to thecentral role it plays in organizing society. This is not myclaim. There are many ways to organize political, economic,and cultural life and no doubt many have been imple-mented. The question is how many are still with us? Rela-tively few. Race, it seems, is one of those that has managed tostick. Other organizing principles have not stood the test ottime. Others th at have, like class or nationality, are arguablybest Interpretable as versions of (or derivations from) racialthinking (Stoler 1995). Indeed, it has been argued, correctly Ibelieve, that race is less a central organizing principle ofAmerican society than a systematic distortion of that o rgani-zation (Wlnant 1994).

Race, 1 contend , has become so widespread because It isso easy to learn. Of the various dimen sions a round which po-litical-economic and cultural lite can be and has been organ-ized (or, as the cast may be, appears to be organized), the


14/17

62 4 American Anthropo logist Vol. 104, No. 2 June 2002

idea of race has "succeeded" because it is so easily learned. Itis so easily learned because of the relationship this particularidea has with a special-purpose program for learning aboutsocial entities. Specifically, the cultural idea of race meets theinput conditionstriggers the operationof a specializedmodule for creating hu ma n kinds and, hence, is readily stabi-lized and entrenched in a cultural environment. The humankind module is a developmental device, a mechanism forguiding the acquisition of knowledge.13 The idea of race con-tinues to be part of the cultural environm ent, and hence a di-mension of the environment which children "sample" intheir early mappings of the social world, because the idea ofrace "fits" well with the conceptual architecture that guideschildren's developing visions of the world. In this importantrespect, the child is the father of man (if you will excuse theovergendered phrase).14C O N C L U S I O NIt is worth repeating how curious is anthropology's aversionto children. Anthropologists certainly have lots of opportuni-ties to meet, talk to, and observe children. More importantly,children are theoretically crucial: anthropology is premisedon a process that children do better than almost all others,namely, acquire cultural knowledge. Nonetheless, the call tobring an anthropology of children into the mainstream hasbeen repeatedly made by, among others, Hardman (1973),Schwartz (1981), Toren (1993), Caputo (1995), and Stephens(1998). Still, a sustained, coherent, andmost criticallytheoretically influential program of child-focused researchhas not emerged. Rather than repeat, yet again, the sameplaint, 1 have tried in this essay to demonstrate the theoreti-cal relevance tha t a fine-grained understanding of knowledgeacquisition affords. Individuals living within a particular cul-tural environment seem to maintain relatively faithful ver-sions of each others ' cultural knowledge.15 To the extent thatanthropologists have tried to explain this, they have as-sumed a fairly simple pattern of elders instructingeither di-rectly through tuition or indirectly through modelingandnovices learning. The obverse idea that elders may behavend believe as they do because of the actions of their chil-ren has little currency in an thropology. Yet, as I have triedo show, there is considerable evidence that it occurs. 1 donot mean that as a general rule children shape adult behaviornd belief. One open and interesting question is how often

have realized.The issues 1 have emphasized here speak to a more generalhat makes something cultural? Understanding

process information. As Sperber (1996: 54) observes,widely distributed to explain why some repre-

Whatever else they may be, successful Ideas are gener-

available input, chances are the ideas will remain currentonly at significant cost.16 Learning isnot as simpleor ignor-ableas much of anthropology would have it, largely be-cause the mindits architecture, the ways it comes to formand use representations, and its natural historyis not assimple as much of anthropology would have it. Anthropolo-gists may not like children, but they should.LAWRENCE A. HIRSCHFELO Associate Professor, Anthropologyand Psychology, 1020 LS&A Building, University of Michi-gan, Ann Arbor, MI 48109-1382NOTESAcknowledgments. This article was w ritten wh ile I was a fellow ofthe Center for Advanced Study in the Behavioral Sciences, Stan-ford, CA. I gratefully acknowledge their generous support. Re-search review ed in this article was supp orted b y NSF award SBR-9319796 and a grant from the Office of Vice President for Research,University of Michigan. I thank Don Brenneis, Bambi Schieffelin,Ann Stoler, Robert Sussman, Kit Woolard, and several anony mou sreviewers for their com m ent s on earlier drafts.1. These claims should be doubly qualified. First, it is mainstreamanth ropolo gy that has evinced limited professional interest in chil-dren and sh own little concern to amend this absence. Second, thephrase "mainstream anth ropolo gy" here is m eant to capture a sta-tistical rather than theoretical regularity. I simply m ean the bulk ofresearch done by anthropologists. "Mainstream anthropology"makes little sense if the phrase is meant to pick out a convergentand unified b ody of th eory. It migh t be more accurate to talk aboutmainstreams in an thropology.2. My goal here is not to review, even briefly, child-focused re-search in anthropology, but simply to mention a few authorswhose works would be included on most anthropologists' l ist ofanthropologists wh o have worked with children.3. One anonymous reviewer suggested that one reason that a seri-ous interest in children has not emerged because, unlike womenand other previously excluded groups, there are no children an-thropologists, hence no constituency to agitate on their behalf.This seems reasonable, up to a point. The overwhelming majorityof anthropologists have children. And within the American mid-dle-class cultural tradition, our children are us. Ask most middle-class Americans how they are doing and they will almost certainlybegin to talk about their children.4. The rich corpus of work on language so cialization is exceptionalin this regard. I cannot comprehensively review this work or evenfairly characterize it here. Suffice it to say for the present discus-sion, the emphasis in language socialization research is largely onlanguage training, in particular the ways in which adults practicesshape th e way cultural and langua ge are jointly learned. The focusis on what adults do in their attempts to ensure that children ac-quire appropriate linguistic knowledge and cultural sensibilities.With the exception of Goodwin (1990), few anthropologists havebeen concerned with how children become socialized into child-hoodwhich, as will be suggested below, is arguably the primarytask facing the young child. Moreover, anthropological linguisticshas remained largely agnostic on the debate surround ing the ques-tion whether children learn language in virtue of the operation ofan innate language acquisition device or in virtue of domain-gen-eral algorithms for acquiring a broad range of knowledge. Argu-ably, this Is a crucial question about the child's own contributionto langu age learning, w hic h w ill be discussed b riefly later. Anthro-pological data are particularly relevant to settling on .in answer,and the most com prehen sive test of the uatiMst hypothesis is amassive cross-linguistic project in which several anthropologicall inguists pa rtklif ted (Slobin 1985).5. V Realization th eory rests, H arris (IJ98) argues, on an unhappyattribution (it causality to correlation. Parents w h o 'g ive plenty oflove and approval, set limits and entorc* them tirmlv but fairtv.


15/17

Hirschfeld Why Don't Anthropologists Like Children? 625don't use physical punishment or make belittling remarks" andwhose child Is "cheerful and cooperative, is reasonably obedient,...is neither too reckless nor too timid, does well in school, has lots offriends, and doesn't hit people without good cause" (Harris1998:17) are thought to be "good" parents. That is to say, their par-enting style caused their child to be a "good" child. It is just aslikely that children whose personalities are inherently consistentwith being "good" children cause their parents to engage in child-rearing practices that are consistent with being "good" parents. Wetend to discount this alternative interpretation because we tend todiscount the notion of "inherent" personality. Yet there is a greatdeal of research supporting the claim that siblings are as likely tohave similar personalities as two individuals chosen randomly(Harris 1998; Plomin and Caspi 1999; Sulloway 1996).6. 1 learned about engacho through personal communicationswith Yu Niiya and Yuri Miyamoto, participants in a seminar in theGraduate Program in Culture and Cognition, the University ofMichigan. Engacho came up during a discussion of cootie lore. Iam grateful to them for their help and particularly to Yu Niija forher efforts in searching for additional information and for prepar-ing a written account.7. Another interesting difference between attributions of groupcontamination under cootie lore and those that occur with in sys-tems racial and caste pollution is that, with the exception of gen-der, cootie lore lacks the explicit referencing of social groups thatare hallmarks of racial and caste thinking. Cootie lore typicallylinks social groups to pollution through implicit reference.8. In one study, Aboud (1988) dressed white children in Eskimocostumes and asked them whether they were now Eskimo. Theygenerally said "yes/ In another study, Semaj (1980) lightenedyoung black children's face with makeup and had them put onblond wigs. When these children asked whether they were nowwhite, they too answered "yes." Both findings admit multiple in-terpretations. Young children could easily be confused about theexperimenter's intent. Children call stuffed bears "bears." Theyknow that adults do the same. Rather than averring to a strong be-lief about race, the children in Aboud's and Semaj's studies mayhave simply been endorsing loose talk. Young children may alsohave been confused about how to answer questions about suddenand contrived changeslike a black child suddenly looking white.9. The discussion here will be limited to the results of studies ofchildren's reasoning about only two racial groups, namely, blacksand whites. However, the research on which this discussion isbased explored a wider range of groups including, depending onthe study, Hispanics, southeast Asians, and North Africans. Resultsindicate that in terms of the reasoning described here, children'sbelief systems are the same across groups of people of color.10. Body build was used as a comparison dimension for two rea-sons. First, it actually is fairly stable over the lifespan and acrossgenerations. Second, it is a reliable index of population of origin.11. This argument is detailed in Hirschfeld 1996, Jackendoff 1992,Furth 1996, and Gigerenzer 1997, who also argue for the existenceof a special-purpose faculty that guides social reasoning.12.

Documents

Why Anthropologists Don t Like Children