32 WHO ARE "THE FOLK" IN FOLK TAXONOMIES?: COGNITIVE DIVERSITY AND THE STATE Roger Sanjek Department of Anthropology Queens College, City University of New York New York, New York Part I In what has come to be called cognitive anthropology or ethnoscience, anthropologists have been concerned with the (sometimes "elegantly" logical, sometimes "psychologically valid") structure and content of culturally recognized domains of experience - color, disease, kinship, plants, animals and so forth. A widely used and universally acceptable name for the goal of this research is folk taxonomy or folk classification. For some examples we might turn to the recent issue of the American Ethnologist (1976) entitled "Folk Biology." There the classifications of such domains as plants, mammals, birds and body parts are examined among such folk as the Aguaruna of lowland Peru, the Ndumba and Rofaifo of high- land New Guinea, the Romanian-Saxon-Gypsy villagers of Vingard, Transyl- vania, and an American hiker. If we were to look a bit farther we would find that folk taxonomies have been collected and analyzed among as many sorts of societies as any political or evolutionary anthropologist might wish to identify. "Folk" taxonomies have been studied among some very different kinds of folk - from neolithic horticulturalists to the chronic alcoholics and bird watchers of American industrial society. The points I wish to make in this paper begin from this observation. I will propose that there can be important differences between tax- onomies in egalitarian and in state societies, or to put it another way, in non-literate and in literate (cf. Goody and Watt 1963) societies. These differences may be found in (1) the processes by which new terms and concepts enter and circulate within current taxonomies-in-use in literate state societies, and in (2) the nature of what cognitive diver- sity we might discover or expect in such societies. I will suggest that the word "folk," though serviceable up till now, has outlived its usefulness as a general term for natural taxonomies-in-use in human cultural systems. While we might continue to speak of "folk classifica- tion" in non-literate socieites, the analogous taxonomies which anthro- pologists record within villages, among occupational groups, in market- places and on street corners in literate state societies should be dis- tinguished with a sejiarate general term. Here I will refer to them as "popular taxonomies." I will argue that popular classifications often bear special (and eminently researchable) relationships to "official" or "legal" taxonomies; these are written, and quite regularly promulgated or even enforced by the state apparatus.1 In fact, folk taxonomies and popular taxonomies are not analogous at all. 33 Aside from its use in "folk taxonomy," the principal use of the term "folk" in American social anthropology has been in Redfield's (1947) conceptualization of "the folk society.",2 It is unnecessary to review all the characteristics of Redfield's model, but it is germane to recall those which relate to cognitive content and its transmission: (1) "[I]n the ideal folk society, what one man knows and believes is the same as what all men know and believe (297)."3 (2) "Nor is there any habitual exercise of classification, experi- ment, and abstraction for its own sake, least of all for the sake of intellectual ends (300)." (3) There is "little communication with outsiders" and "intimate communication among the members of the society (296)," (4) There is an "absence of books. The folk communicate only by word of mouth . . . among neighbors, within the little society itself . . . [O]ral tradition has no check or competitor (296),"t Now it would be sophistry to imply that the use of the term "folk" in "folk taxonomy" carries with it acceptance of Redfield's model of the "folk society." Cognitive diversity has been documented by several an- thropologists working in societies Redfield would have referred to as "the tribal groups that still remain around the edges of expanding civilization (306)," groups he saw as "the small remainders of this primary state of living (306). " (See Berlin and Berlin 1975; Heider 1972; Sankoff 1971.) And the work of Levi-Strauss (1966) and Horton (1967) certainly calls into question the absence of classification and abstraction for intellectual ends in non-literate societies. The final two points of Redfield however - the lack of communication with out- siders, and the absence of writing - bear closer attention. Redfield considered several alternative wordings but chose the term "folk" "because, better than others, it suggests the inclusion in our comparisons of peasant and rustic people who are not wholly inde- pendent of cities (293n)." My point of departure from Redfield and from the general use of "folk taxonomy" is precisely that the presence of, regular contact with outsiders (and with their cognitive classifications) - through markets at very least (cf. Skinner 1964) - and the presence of writing places peasants, "rustic people" and all other members of liter- ate state societies into a qualitatively different existential situa- tion from members of non-literate "folk" societies. Goody and Watt (1963; see also Goody 1968) have given careful atten- tion to the differences between non-literate and literate societies, and to the consequences of the transition from restricted to widespread literacy. In non-literate societies where the "transmission of the ver- bal elements of culture by oral means can be visualized as a long chain of interlocking conversations between members of the group (Goody and Watt 1963:29)," they see "elaboration of the vocabulary (29)," that is, development of folk taxonomies, to be in accord with daily life: 34 [T]here is an intimate functional adaptation of language in non-literate societies (29). One of the most important results of this homeostatic tendency is that the individual has little perception of the past except in terms of the present. . . . [T]he elements in the cultural heritage which cease to have contem- porary relevance tend to be forgotten or trans- formed (33-34). Goody and Watt's theory does not rule out cognitive diversity: "we are denying neither the occurrence of social change nor yet the 'survivalst which it leaves in its wake (31)." Terms may be introduced, spread within local groups and fall out of use. Goody and Watt's posi- tion is compatible with an expectation too of diversity along age and sex axes, within specified domains. They also acknowledge the role of mnemonic devices (31) in preserving cultural content (quite obviously in a differential fashion). But the overall picture remains one of a "homeostatic" relationship between taxonomies and immediate social activ- ities. Language is viewed as practical and adaptive in a direct sense. With writing the situation becomes quite different. The relationship between local classifications and local activities need not have the same "'close fit (Goody 1968:5)"1 at all. "[W]ith the possibility of written communication (not to speak of the more overt influence of national government and institutions), the physical community no longer limits the field of socio-cultural interaction (8)." Writing also enables geographically dispersed groups to carry on coordinated activities in terms of common understandings. In the case of Imperial China one written language united a literate class speaking several mutually unintelligible dialects, and even permitted communication across lan- guage borders beyond China to Japan, Korea and Vietnam (cf. Wang 1973:55). The oral mode of communication continues of course, but within a new intellectual and political context. In literate societies individuals may transfer information from written discourse to oral, and vice versa. An individual may use one or the other, or various combinations of oral and written communication in different arenas of activity. An individual's cognitive map of a particular domain may expand in content rapidly, merely by reading a book; habitual reading may make mazeway reorganization a continuous process. At the same time, portions or fragments from the written body of taxonomic information may be transmitted orally (as in doctor-patient consultations) and this may in turn affect cognitive content and organ- ization among others through oral or even written communication (writing a friend about what the doctor told you). The implications of these processes are considerable for the study of cognitive diversity in societies where literacy is widespread. [T]he mere size of the literate repertoire means that the proportion of the whole which any one 35 individual knows must be infinitesimal in com- parison with what obtains in oral culture. Literate society . . . prevents the individual from participating fully in the total cultural tradition to anything like the extent possible in non-literate society, . , , [T]he literate in- dividual has in practice so large a field of personal selection from the total cultural rep- ertoire that the odds are strongly against his experiencing the cultural tradition as any sort of patterned whole (Goody and Watt 1963:57-58) . But Goody and Watt (1963:58-60) also point out that the written mode of communication by no means imposes itself uniformly. Not every- one reads; those who do, read selectively, and in varying amounts; no two individuals are likely to have read precisely the same things, or to have read the same things in the same sequence. What are the implications of all this for different semantic do- mains and for different social groups within literate societies? Where are there likely to be pockets of cognitive sharing and agreement amid the structually determined diversity which Goody and Watt correctly make so much of? What are the precise channels of information exchange between written and oral modes? What are the social types of women and men who mediate such exchange, and are they more or less effective in different domains? If "writing . . . encourages private thought (62)" and if "such coherence as a person achieves is very largely the result of his personal selection, adjustment and elimination of items from a highly differentiated cultural repertoire (63)," then is not sharing and agreement the exceptional situation in literate societies, and cognitive diversity the expectable norm? In literate societies we may easily record any number of individual propriospects or cognitive maps for a semantic domain but when should we begin to sketch out the "cul- tural pool" (Goodenough 1971) or the "larger cultural informational system" (Roberts, Chiao and Pandey 1975) of which they form a part? Part II I would now like to turn to some enthnographic instances in which either state policy or official-legal classifications interact with or alter popular taxonomy. I shall choose my illustrations from several domains. It should be obvious that none of what I am about to discuss occurs in non-literate societies. My first example is the terminology for prices and monetary denom- inations in use among the residents of the Adabraka section of Accra, Ghana, where I carried out research in 1970-71 and 1974. (See Sanjek 1972; Sanjek and Sanjek 1976.) Adabraka people had two sets of terms by which they spoke about money and prices. The off'icial terms, those created and enforced by the state, appeared upon coins and paper notes, in government publications, and on the posted prices of items in shops 36 and markets as required by law. These terms were "new cedi" and "new pesewa," 100 new pesewas (np) making up one new cedi (NO). Adabraka residents certainly used these terms, and spoke of earning "35 cedis" per month, paying "5 new" for the newspaper and so forth. Along with this classification an alternative popular terminology was also in use. The terms in it included "penny," "pence," "shilling," "bob," "pound," and "guinea." We quickly learned to use this termin- ology along with the official one: that N02 equaled one pound, that 60np equaled six shillings, and that we need pay only 5np when told an item costs "sixpence." The Ghanian government tacitly recognized the popular system by minting an oddly-shaped coin worth 2inp and providing thereby a "threepence" piece. We did not collect systematic data on whether sex, occupation, education or other social characteristics were associated with using either of the two sets in the same situation (diversity) or using them both alternatively in different contexts (variability). Hypotheses along these lines could certainly be tested (though I am certain there are more interesting cases of popular/official taxonomic coexistence in which to do so.) Now in this case it is clear that today's popular taxonomy is yes- terday's official-legal classification. But I am not certain that things are always so obvious. Foster (1953) and Marriott (1955) have argued that, in other domains, what appears to be the "folk" system within a village might bear important relationships to past or present official- legal categories codified in written texts. In the Ghana case we see state policy officially replacing previous state policy and leaving a popular taxonomy as residue. In other domains and in other cultures one might find a less imme- diately identifiable contrast between popular and official-legal classi- fications. And one might also find cases in which individual terms from one system become incorporated or replace terms in the other. Ghanaians would not speak of "20np and sixpence" as the price of anything (it would be "25np" or "two and six") but the incorporation of Spanish color terms into the color classification of Aguaruna Indians who have been exposed to Mission education has been recorded by the Berlins (1975). As the Aguaruna pass from non-literate integrity to illiterate disad- vantage, their folk taxonomies are becoming popular taxonomies in relation to the official-legal classifications of the Peruvian national literate culture and government. I would now like to mention another approach to the co-presence of popular and official-legal classifications, one drawing upon Geertz's concepts of force and scope (1968:111-113). Geertz developed this dis- tinction in comparing religious belief over time and among social groups in Morocco and Indonesia, but he recommends it as a useful tool in the analysis of "any symbol system men use to construe experience (111)." Force concerns the "thoroughness with which such a pattern is internalized in the personalities of the individuals who adopt it, its centrality and 37 marginality in their lives (111)"; and scope is "the range of social contexts within which" a set of symbols is "regarded as having more or less direct relevance (112)." In Adabraka one would probably find that both the force and the scope of the popular "shilling-pound" classification exceeds that of the official "pesewa-cedi" system for most if not all informants. Problems of operationalizing the two concepts of course remain; Geertz is char- acteristically reticent about procedures of measurement. These problems, however, are solvable ones. I believe it would be rewarding for cognitive anthropology to pursue questions about the scope and force of alternative taxonomies in literate societies. My next pair of examples comes from the application of ethnoscientific methods to problems of health and disease. Harwood's research (1971) on hot-cold food classification among Spanish-speaking residents of the South Bronx identified a popular taxonomy quite different from the offi- cial classification of types of food used by medical doctors serving the ,area. The results of Harwood's work have encouraged doctors to prescribe diets in accord with the popular taxonomy and to avoid prescribing diets which might fit the official classification but would not be followed by the patients. Through the written record, here the Journal of the American Medical Association, popular taxonomy has come to be used in conjunction with the official classification, at least among those doctors who have read Harwood's article. Using Geertz's concepts, both the scope and force of the official classification have been diminished among these physicians. A reverse process is also possible. No doubt there are Puerto Rican New Yorkers for whom the scope and force of the hot-cold taxonomy is neg- ligible, and who attempt to plan meals, in sickness or in health, according to their personal knowledge of classifications of food by official nutri- tive and medical value. As I suggested earlier, anthropologists might study the social contexts of the interaction between popular and official taxonomies. In the case of medical knowledge we tnight find that popular systems, though of major force and scope in determining the behavior of their holders, could be understood as partial and incomplete reflections of official classifications. In Manning and Fabrega's discussion (1976) of their research on med- ical knowledge in Chiapas, they found Subjects . . . did not show complete agreement regarding the folk medical knowledge evaluated in this study .. . There is thus considerable variability [what I would call diversity] within small samples as to the agreement on the corelates of given medical problems, their relationships to each other as types, and their relationships to cul- tural conceptions of 'health', 'illness', and 'treatment'(43). 38 On this basis they speak of "difficulty in the ethnoscience approach to information gathering (43)." I suggest the difficulty arises from their conceptualization of the problem and not from "the ethnoscience approach." I would argue that they are not, in a small town within the literate society of Mexico, studying a "folk" system at all. They state that "Disease definitions, symptom patterns, and treatment options arise within a socio-cultural unit (family, neighborhood, friends) which is actively coping with them. The illness episode works itself out in a particular family with a given history and residue of past experiences with illness and in a given social and temporal location (47-48, my emphasis)." Have they not defined their field of study too narrowly? Elsewhere they report that, when ill, a family might "visit a curandero and a physician" after "various measures, e.g. herbs, were used initially (49-50)." It is expectable that differential experience with physicians (and with curanderos for that matter) would alter the "folk" medical knowledge of particular families, individuals, and perhaps their neighbors and friends, especially when the advice of the health specialists led to recovery. These inputs would lead in turn to consider- able diversity in popular medical knowledge. Furthermore, if any of the informants were literate, or had obtained information from literate acquaintances, then medical knowledge from newspapers, farmers' almanacs, Selecciones, and other sources would increase diversity even more so. It is probably a conceptual mistake to speak of "the 'folk theoryt of illness which prevails in this region (50)." Rather there are the written official classifications of disease and treatments at the access of physicians, and the popular oral taxonomies of curanderos and of or- dinary Ladinos and Tzotzil and Tzeltal Indians (perhaps varying by village), There might even be no popular systems shared by all non-specialists in particular ethnic categories; logical and psychologically valid cognitive organization might exist only at the level of family traditions or indi- vidual propriospects. The final area of cognitive research I would like to mention is folk, popular and official-legal classifications of ethnicity - the taxa used to describe the physical and cultural human differences recognized with- in a society and among its neighbor societies. This important and inter- esting domain has received little attention from cognitive anthropology (but see Kay 1975; Sanjek 1971, 1977.) Less formal treatments of what Schwartz calls "ethnoethnography" in such non-literate societies as the Admiralty Islands (Schwartz 1975) and the Nuer (Evans-Pritchard 1940:125- 135) suggest that the general "homeostatic" processes Goody and Watt identify within non-literate society apply in this folk domain. In state societies the separation of popular and official-legal versions of "ethnoethnography" is especially necessary. A full treatment of this point will not be attempted here, but I would like to call atten- tion to two examples which bear further investigation with the popular and official-legal distinction, and the room for interplay between the 39 two, in mind: (1) the pre-conquest Ashanti case, where a legally- enforceable ban on discussion of ethnic origins occurred within a context of considerable numbers of individuals of non-Ashanti slave and captive origin becoming absorbed as Ashantis (cf. Rattray 1929); (2) the evol- ution of ethnic categories over the last four and a half centuries in Latin America, where the interaction between legal ethnic status, pop- ular terminologies, and changing state policy has been complex (cf. Diggs 1953; Morner 1967). In a companion paper to this one (Sanjek 1977), I have called attention to the factor of government policies and taxonomies in interpreting diversity at the popular level of ethnic categories in Adabraka. Part III In conclusion, let me return to the issue of cognitive diversity. I have attempted to make the point that the kinds of cognitive diversity which may be encountered in non-literate society are considerably com- pounded by social mechanisms producing cognitive diversity in literate state societies. I have argued that state policies and the presence of written records may affect popular taxonomies-in-use in several ways: by replacing or competing with popular classifications, by suppressing them, and by permitting the introduction of new bits of information or modes of classification through channels not present in non-literate societies. Within this context, a sample of individual cognitive maps of a domain should not be expected to provide data for the construction of a single coherent set of dimensions of contrast or mode of organiza- tion. I do not mean to assert that such is the case for every social group in every domain within literate societies. People do talk to each other and are often understood. Perhaps it is in those domains in which the oral mode is central in the learning process that agreement and sharing would be most expected. Domains such as color, body parts, household furniture and modes of transportation are learned orally by children in American culture, with the same process of "direct semantic ratification (Goody and Watt 1963:29)" characteristic of non-literate society. In these cases questions about the relation between popular and official- legal taxonomy may be sterile. But when literacy intervenes or replaces oral transmission in the learning process we may expect considerably more diversity. A group of anthropologists may all agree that an arm is an "arm" and a leg is a "leg." or that one color chip is "blue" and another is "yellow." But imagine what diversity would reign if they were asked about how many specializations and subfields are included in anthropology, and what exactly they are. There is no reason to shy away from the challenge of research on "popular" taxonomy in literate society, nor to abandon the methods which have been used in non-literate societies, but such research cannot 40 continue with the same set of assumptions underlying work in "folk" taxonomy. In fact, I think we need a good deal more ethnoscience in literate state societies. Recognition of the special characteristics of cognitive diversity in such societies will bring ethnoscience into closer relationship with general anthropological thinking about his- torical development, about social structure, about inter-locality exchange and about power. ACKNOWLEDGMENTS This paper was presented at a symposium on "Theoretical Implications of Cognitive Diversity - Some Whys and Therefores" at the University of Alabama in Birmingham, October 15-17, 1976, and later read at a Wednesday Bag Lunch at the Language-Behavior Research Laboratory, Univer- sity of California, Berkeley. I would like to thank Jane Christian and Peter Gardner for the invitation to prepare the paper, and Brent Berlin, Marshall Durbin, Gene Hammel, bagmeister Paul Kay, Dena Lieberman, Pierre Maranda and Gillian Sankoff for their comments and suggestions. NOTES lThere may be several classifications at each of the two levels - popular and official-legal - at any particular time. Popular classifications might vary with ethnicity, region, occupation or class within a state society; several official-legal classifications may simultaneously be in use in different branches of the state apparatus. In a comment fol- lowing a presentation of this paper at Berkeley, Gene Hammel pointed out that this latter situation now obtains with respect to classifications of "race" in use among different American educational and governmental agencies. 2"Folk" of course is also a central concept in the study of folklore, both literary and anthropological (see Dundes 1966). Definitions of "folk" in "folklore" vary (Dundes 1965:1-2), but one important approach in American anthropology has been to define folklore as unwritten, orally- transmitted literature: "forms of artistic expression whose medium is the spoken word . . . .folklore may never be written, even in a literate society, and it may exist in societies which have no form of writing (Bascom 1972; see also Herskovits 1972)." Dundes (1965; 1966) for quite cogent reasons rejects exclusively oral transmission as a charter for what to include in the study of "folklore"; he would include some written materials, and also directs attention to the exchanges of "folklore" between oral and written modes of transmission. Dundes approaches folk- lore substantively, in terms of its forms (1965:3; 1966:238). His list does not include cultural domains such as color, animals, kinship, ethnic categories, diseases, or most other topics dealt with in studies of "folk taxonomies," or discussed in this paper. 3Redfield continues, "In real fact, of course, the differences among in- dividuals in a primitive group and the different chances of experience 41 prevent this ideal state of things from coming about. Nevertheless, it is near enough to the truth for the student of a real folk society to report it fairly well by learning what goes on in the minds of a few of its members (297)." REFERENCES CITED American Ethnologist 1976 Special Issue: Folk Biology. Volume 3, number 3. Bascom, William 1972 Folklore. In Standard Dictionary of Folklore, Mythology, and Legend. Maria Leach, ed. P. 398. New York: Funk and Wagnalls (original edition 1949). Berlin, Brent and Elois Ann Berlin 1975 Aguaruna Color Categories. American Ethnologist 2:61-87. Diggs, Irene 1953 Color in Colonial Spanish America. Journal of Negro History 38:403-427. Dundes, Alan 1965 What is Folklore? In The Study of Folklore. A. Dundes, ed. Pp. 1-3. Englewood Cliffs: Prentice Hall. 1966 The American Concept of Folklore. Journal of the Folklore Institute 3:226-249. Evans-Pritchard, E.E. 1940 The Nuer. Oxford: Oxford University Press. Foster, George M. 1953 What is Folk Culture? American Anthropologist 55:159-173. Geertz, Clifford 1968 Islam Observed: Religious Development in Morocco and Indonesia. Chicago: University of Chicago Press. Goodenough, Ward 1971 Culture, Language, and Society. Reading: Addison-Wesley. Goody, Jack 1968 Introduction. In Literacy in Traditional Societies. J. Goody, ed. Pp. 1-26. Cambridge: Cambridge University Press. Goodyg Jack and Ian Watt 1963 The Consequences of Literacy. Comparative Studies in Society and History 5:304-345. (As reprinted in Literacy in Tradition- al Societies. J. Goody, ed. Pp. 27-68. Cambridge: Cambridge University Press.) 42 Harwood, Alan 1971 The Hot-Cold Theory of Disease: Implications for Treatment of Puerto Rican Patients. Journal of the American Medical Asso- ciation 216:1153-1158. Heider, Eleanor Rosch 1972 Probabilities, Sampling, and Ethnographic Method: The Case of Dani Colour Names. Man 7:448-466. Herskovits, Melville 1972 Folklore. In Standard Dictionary of Folklore, Mythology, and Legend. Maria Leach, ad. P. 400. New York: Funk and Wagnalls (original edition 1949). Horton, Robin 1967 African Traditional Thought and Western Science. Africa 37:50-71, 155-187. Kay, Paul 1975 Tahitian Words for Race and Class. Working Paper No. 40. Berkeley: Language Behavior Research Laboratory. Levi-Strauss, Claude 1966 The Savage Mind. Chicago: University of Chicago Press. Manning, Peter K. and Horacio Fabrega, Jr. 1976 Fieldwork and the 'New Ethnography'. Man 11:39-52. Marriott, McKim 1955 Little Communities in an Indigenous Civilization. In Village India: Studies in the Little Community. Chicago: University of Chicago Press. Morner, Magnus 1967 Race Mixture in the History of Latin America. Boston: Little Brown. Rattray, Robert S. 1929 Ashanti Law and Constitution, London: Oxford University Press. Redfield, Robert 1947 The Folk Society. American Journal of Sociology 52:293-308. Roberts, John, Chien Chiao and Triloki N. Pandey 1975 Meaningful God Sets from a Chinese Personal Pantheon and a Hindu Personal Pantheon. Ethnology 14:121-148. Sanjek, Roger 1971 Brazilian Racial Terms: Some Aspects of Meaning and Learning. American Anthropologist 73:1126-1143. 43 (Sanjek, Roger) 1972 Ghanaian Networks: An Analysis of Interethnic Relations in Urban Situations. Ph.D. dissertation. Anthropology Department, Columbia University. 1977 Cognitive Maps of the Ethnic Domain in Urban Ghana: Reflections on Variability and Change. American Ethnologist 4:603-622. Sanjeck, Roger and Lani Morioka Sanjek 1976 Notes on Women and Work in Adabraka. African Urban Notes 2(2):1-25. Sankoff, Gillian 1971 Quantitative Analysis of Sharing and Variability in a Cognitive Model. Ethnology 10:389-408. Schwartz, Theodore 1975 Cultural Totemism: Ethnic Identity Primitive and Modern. In Ethnic Identity: Cultural Continuities and Change, George De Vos and Lola Romanucci-Ross, eds. Pp. 106-131. Palo Alto: Mayfield. Skinner, G. William 1964 Marketing and Social Structure in Rural China, Part I. Journal of Asian Studies 24:3-45. Wang, William S-Y. 1973 The Chinese Language. Scientific American 228(2):51-60.