How Thick Is Blood? The Plot Thickens - . .: If Ethnic Actors Are Primordialists, What Remains of The Circumstantialist/primordialist Controversy?
How Thick Is Blood? The Plot Thickens - . .: If Ethnic Actors Are Primordialists, What Remains of The Circumstantialist/primordialist Controversy?
How Thick Is Blood? The Plot Thickens - . .: If Ethnic Actors Are Primordialists, What Remains of The Circumstantialist/primordialist Controversy?
Francisco J. Gil-White
Abstract
An investigation of the cognitive models underlying ethnic actors own ideas concerning the acquisition/transmission of an ethnic status is necessary in order to resolve the outstanding differences between primordial and circumstantial models of ethnicity. This article presents such data from a multiethnic area in Mongolia that found ethnic actors to be heavily primordialist, and uses these data to stimulate a more cogent model of ethnicity that puts the intuitions of both primordialists and circumstantialists on a more secure foundation. Although many points made by the circumstantialists can be accommodated in this framework, the model argues that ethnic cognition is at core primordialist, and ethnic actors instrumental considerations and by implication their behaviours are conditioned and constrained by this primordialist core. The implications of this model of ethnicity for ethnic processes are examined, and data from other parts of the world are revisite d for their relevance to its claims. Keywords: Ethnicity; cognition; circumstan tialism ; instrumentalism ; primordialism; Mongolia.
1. Introduction I once asked a Mixtec Indian from Oaxaca, Mexico (a former governor of the State) whether one could become Mixtec if ones parents were not Mixtec. He looked at me as though celery stalks had suddenly begun sprouting from my head, and I do believe he feared for my intelligence. You can only be Mixtec if your parents are Mixtec, he said, what do you mean? Another day I asked an Ethiopian taxi-driver in Los Angeles, If
Ethnic and Racial Studies Volume 22 Number 5 September 1999 pp. 789820 1999 Taylor & Francis Ltd ISSN 0141-9870 print/ISSN 1466-4356 online
790
Francisco J. Gil-White
a child was born of Tigre parents, but was immediately adopted by an Oromo couple, subsequently grew up among Oromo, and was in every respect like an Oromo, would he be thought of as an Oromo? No, he replied, he would be accepted by the Oromo community, but the parents would think This is our Tigre child. He would still be Tigre. And recently, my Russian teacher, herself a Russian Jew, told me that to her I was a Jew, because I am descended from Jews. She will not budge, and maintains this view despite being aware that (1) I have to go back about four generations (perhaps more) to nd an ancestor who practised Judaism (after that they are all Roman Catholic); (2) I did not grow up with a Jewish identity; and (3) my parents and I did not even know that any of our ancestors were Jewish until I was about ten years old, when a genealogy buff in the family uncovered this information. I suspected my Russian teacher would scoff at the circumstantialist model of ethnicity which maintains that ethnies are constructed by rational actors who calculate their objective interests and then make decisions concerning association and/or political mobilization with others. I was right: she found it silly. I myself would be less harsh, especially considering that there is more to the circumstantialist view than the short caricature I gave her. Nevertheless, this article will ask and elaborate on the following question: if most people are like my Russian teacher, where does this leave the circumstantialist model? Recently the circumstantialist (aka instrumentalist) model of ethnicity vs. the primordialist model has been in the ascendant. However, the above anecdotes reveal that some people possess ethnobiolog ical and therefore primordialist models concerning the acquisition /transmission of ethnic statuses. This will not affect our views on ethnicity if such models are not common, but this study will present recent ethnographi c data from Mongolia that found primordialist models to be predominant there. These data point to very serious shortcomings in the circumstantialist model, and also to some under-recognized strengths in the primordialist view that can be formalized and operationalized in a cogent, scienti c model of ethnicity, which I shall attempt to do. With these insights, much data from other parts of the world, including some earlier put forth to support the circumstantialist position, can be shown instead to bolster these key primordialist points. I propose a more sophisticated and testable model, which ts the empirical data better and combines valuable insights from both primordialists and circumstantialists without succumbing to the excesses of either. Perhaps we can lay the controversy to rest. 2. Barth and the circumstantialist model Anthropologists once believed ethnic groups or cultures to be peoples in a unitary sense along various dimensions: ascriptive (labelling) ,
791
moral (normative) and cultural (linguistic and artifactual). An ethnic group therefore understood itself as such, was labelled by others in like fashion, had a particular and distinctive culture (including a dialect), and whose members preferred each other to non-members (that is, endogamy, discrimination, ingroup solidarity, etc.). This oversimpli es but still captures the culture area view of ethnic groups, predominant at one time as a result of the enormous in uence of British functionalism (which imagined societies as well-bound ed and functionally integrated organisms), and the adoption of Malinowskis model of ethnographi c work (Levine and Campbell 1972, pp. 81 84). Reactions against this view began with Edmund Leachs Political Systems of Highland Burma (1954), followed later by Moermans work among the Lue in Thailand (Moerman 1965, 1968) . These studies complained that ethnic identities did not map neatly to the distribution of cultural material, and proposed a shift from objective indicators of groupness, such as measurable discontinuities in the distribution of artifactual or ideational culture, towards a more subjective focus that relied heavily on the labelling processes of ethnic actors themselves. This view climaxed in 1969 with the publication of Fredrik Barths famous introduction to Ethnic Groups and Boundaries. The argument is simple: in order to have a social identity one must meet the conditions for being referred to by the linguistic expression [the label] that names the identity (Goodenough 1965, p. 21). Thus, the labelling processes of local ethnic actors themselves are the only guides to the limits of the group, for the [cultural] features that are taken into account are not the sum of objective differences, but only those which the actors themselves regard as signi cant (Barth 1969, p. 14; emphasis added) . Any aspects of culture not recognized by local ethnic actors as signi cant will not necessarily covary with different ethnic labels.1 Leach (1977, pp. 293 97) had reported that people in the Burma Kachin Hills sometimes switched ethnic identity. To him this was further evidence that the view of a society as a thing (that is, a bounded whole) was wrong. Barth, a student of Leach, documented similar behaviour in Swat, Pakistan, and gave a majestic theoretical framework for interpreting it. Thus, Barth had an enormous impact because his theoretical sophistication went well beyond a cogent articulation of the subjectivist approach. In the communities that Barth studied, some individuals born into the Pathan ethnic group were, later in life, labelling themselves Baluch as circumstances made this advantageou s. Similarly, some Fur in Darfur, Sudan, were taking up nomadism and calling themselves Baggara (Haaland 1969) . On the strength of these two studies, Barth argued that an ethnic status implies a particular kind of actor, who will therefore easily coordinate with others of a kind for reciprocal exchanges (Barth 1969, p. 15). Thus, if I am an A, but it is better for me to interface and
792
Francisco J. Gil-White
network with Bs, I shall acquire a B identity together with B ways of being so as to tap into the B network. Ethnic actors are rational actors who make choices about their ethnic statuses as (ecological/ economic/political) circumstances make this instrumental. With this argument, the circumstantialist aka instrumentalist school of ethnicity was born. Most scholars today would agree that Barth changed our views for ever. Following publication of Ethnic Groups and Boundaries, the idea that ethnies are in the rst instance collections of individuals sharing a common self-ascription, but with no necessary relation to any particular cultural content, became accepted by all. There is much controversy in ethnic studies, but that point, a full thirty years after its initial submission, is not contested, and on good empirical grounds, which makes Barths achievement that rarest of anthropological accomplishments: cumulative science. The debate now turns to the second part of his argument: are ethnic groups rational associations of self-interested actors, as he claims, or are they irrational primordial groupings governed by emotional attachments, as others maintain? Both sides of that theoretical coin, and their protagonists, will receive their full due below, but note here the glaring methodologic al gap (the awareness of which prompted the present effort): nobody seems concerned with investigatin g the cognition of ethnic recruitment and how it affects ethnic processes. This is remarkable, because we all apparently now agree with Barths subjectivist (self-ascriptive) approach which claims that ethnies and social identities more generally establish their boundaries from the conditions for being referred to by the linguistic expression that names the identity (Goodenough 1965, p. 21). If this is true, then whatever makes a particular group ethnic is that the conditions for using that groups label are ethnic (as opposed to being political, religious, club-like, etc.; cf. Nagata 1981, p. 111). If nothing distinguish es ethnic conditions of membership from those of other social identities, we should stop talking about ethnies immediately or risk making empty theories. But we have not investigated this in great depth, and certainly not with a methodology true to Barths crucial shift away from objective, cultural groups to subjective, self-ascriptive groups. However much Barth may have insisted on the importance of the subjective perspective, he did not investigate his informants ascriptive cognitive models, but chose instead to infer these from their behaviours. This may be a serious shortcoming given that the new perspective is self-consciously emic. Without data on the cognitive models we know neither how ethnies are thought of, nor what kinds of membership conditions are characteristic of ethnies. Neither do we know whether Barth documented switches of identity or merely of signalling. Suppose, for example, that none of his Pathans-turned-Baluch really
793
thought of themselves as Baluch, and that the Baluch did not think of them as real Baluchs either. If this were true, then the labelling changes might simply re ect the fact that some Pathans wanted to be counted, treated and evaluated as if they were Baluch indeed, because it was instrumental to network intimately with Baluchs and this required signalling that they accepted Baluch standards and expectation s without that changing their identity in their own eyes or in those of others. Unsolicited remarks from Barths and Haaland s informants suggest this may have been the case: Haaland was taken out to see Fur who live in nomad camps [that is, not really Baggara] and I have heard members of Baluch tribal sections explain that they are really Pathan (Barth 1969, p. 29). I shall treat these studies in greater detail below, but for now take note that such remarks suggest that perhaps if you are born a Pathan you will always be recognized as really Pathan even if you take up the Baluch way of life and for practical purposes wear a Baluch hat. In other words, both Pathans and Baluchs may have ethnobiolog ical models for the acquisition and transmission of ethnic statuses, which (in one sense at least) makes them primordialists . One may argue that, even if true, these are all mere window-dressing quibbles that subtract nothing from Barths rational-choice argument while conceding its most fundamental points. But I doubt that they are super cial quibbles, if Barths most original argument that ethnic statuses are interpreted by ethnic actors as signals of kind is in general correct, as I suspect it is. A natural kinds ethnotheory of ethnicity would probably think of such kinds as biologically inherited, and therefore inalienable, because this is how humans in general think about natural living kinds (Gil-White 1999) . If so, the general tendency would probably be to make it impossible, in most parts of the world, to signal that one is now a different kind of actor by appropriating an ethnic label one was not born with (for example, try to imagine a Croat waking up one day and deciding to call himself a Serb). If these arguments are good enough at least to provoke, then the study of ethnicity may bene t greatly from a cross-cultural investigation of the cognitive models that underlie peoples ideas for the acquisition and transmission of ethnic statuses. To underscore the last point, I shall now present cognitive data from Western Mongolia. It is necessary to bear in mind that the data below have few pretensions. They are not offered as a demonstration of anything beyond the study area itself that would require a cross-cultural replication of this methodolog y. But they nevertheless have several virtues: (1) they exemplify the kind of cognitive work that has been missing; (2) their nature seriously challenges the circumstantialist model as it currently stands, suggesting that it needs rethinking; and (3) they inspire a new model of ethnicity with fresh predictions amenable to empirical tests, which I explore after presentation of the data. Thus, these data are here to stimulate a new way of thinking about ethnicity that may
794
Francisco J. Gil-White
be fruitful, and to exemplify one way of investigating the questions, but they are entirely insuf cient as a con rmation of the ideas advanced further below. 3. A report on the Mongol cognition of ethnic boundaries
Field site
My study population consists in the main of Torguud nomadic pastoralists. Torguuds are a small Mongol ethnic group in Western Mongolia. They move around in the district of Bulgan Sum, in Hovd province, Republic of Mongolia. In winter they are not far away from the district centre, which includes a town lying by the Bulgan river. This town has ca . 2,500 inhabitants. The land is quite fertile on the banks of the river and the sedentary residents of the city grow all manner of fruits and vegetables in their small horticultural gardens. Another 2,500 people make a living as farmers beyond the town, and are considered as part of the district centre. Beyond these lies the steppe, where a total of about 5,000 nomads eke out a living. The Bangyakhan clan, with whom I worked, winters in one of the Bulgan rivers two large ood-plain valleys, some way from the town, and spreads out over an area of about 140 km 2 . During this time, male herders make regular two-week trips to the Gobi or to the nearby hills where the snow is less thick and the sparse grass more accessible. A few move around with their entire household. They also assist their livestocks diet with hay made in late August/September. In the summer months they move to the highlands in the Altai mountain range, changing their location constantly as pastures become depleted . They may make as many as ten migrations in a four-month period. These highland summer pastures are very green high-altitude forest-steppe, criss-crossed by innumerable glacial rivers and streams. It was in this lovely setting, while living with the pastoralists, that the research was carried out. The site is practically located on a double border: to the north lies the provincial border separating Hovd from Bayn lgii; to the West lies the internationa l border separating Hovd in Mongolia from Xinjiang in China. Apart from Torguuds, there are other Mongol ethnic groups in the area, as well as a large Kazakh population , the biggest local ethnic contrast being that between Mongols and Kazakhs. There are sedentary and nomadic individuals in all the local ethnic groups such that there is no sharp, ethnically based economic/ecological differentiation. Neither is there any noticeable ethnic socio-economic ranking as everybody is poor.
Methods
A short questionnai re was administered verbally to fty-nine subjects, chosen randomly in the sense that systematic patterns of inclusion were
795
consciously avoided. But it is a non-random sample because dif cult terrain, long-distances, and slow transportation (horseback) make proximity to the researcher and the likelihood of inclusion in the sample highly correlated. However, I doubt that spatial proximity to myself and ideas about ethnicity also correlated. A possible confound is that family members might tend to agree with each other, in which case the effective number of data points is less than what it seems. As it turned out, however, knowing how one individual responded was in general a poor predictor of how that persons close-kin would respond, which suggests that family background is not an important causal variable. So even though my sample is less than ideal, it is far from meaningless. The questionnaire was as follows: Question 1. If the father is Kazakh and the mother Mongol, what is the ethnicity of the child? Question 2. The father is Kazakh, the mother Mongol, but everybody around the family is Mongol and the child has never even seen a Kazakh, outside of the father. The child will learn Mongol customs and language. What is the ethnicity of this child?; Question 3. A Kazakh couple have a child that they do not want. They give it in adoption to a Mongol couple when the child is only a year old. Around the Mongol family there are only Mongols and the child grows up never meeting a single Kazakh. Since he was a baby when adopted, he knows nothing and thinks that his biological father and mother are the Mongol adopters. He grows up learning Mongol customs and language. What is the ethnicity of this child? 2 Simple, real-time, genealogy diagrams, whose logic was explaine d before asking the questions, assisted the representation of the questions. The procedure was the same for all interviewees (excepting idiosyncratic clari cations). The emphases shown above were used sociolinguistically in the verbal rendering. Whenever (1) people changed their sequence of answers under crossexamination (not very common but it did happen), (2) both the rst sequence and the second revealed consistent models, and (3) I was not highly con dent that I could discern which one they really believed , I erred against my preferred hypothesis by recording the less primordialist of the two sequences of answers.
796
Francisco J. Gil-White
Table 1. Ascription: appearance vs. descent Answer Child is Kazakh Child is Mongol Child is erliiz (half-breed ) Total
N = 59
also transmit ethnic ascription. The question was open in that they were not forced to choose among predetermined options (half-breed was their idea), but in another sense it was forced by presuming that children are born with an ethnic status. However, if actors are circumstantialists, perhaps they should object that it depends, and explain what it depends on. No such answer was ever given. The formulation of Question 2 presumes the opposite: ethnic ascription will depend on circumstances of enculturation which, in this question, are quite extreme. The but was highly emphasized sociolinguistically by raising my voice along with my index nger, while making big eyes that looked straight into the interviewee s in what I hope was an ominous expression. This was to draw close attention to a set of circumstances absent in the rst question that might make the answer to the second different, in fact implying that this was the answer I expected (because I wanted to create a bias against my favoured hypothesis). However, the overwhelming majority of respondents were unfazed by this implication and insisted that the child in the second question was Kazakh. Question 3 is perhaps the most extreme circumstantialist scenario possible. If respondents insist this child is also Kazakh, they will be saying that one can be Kazakh and not know it. More than half responded in this way. For them, apparently, the child will take the biological fathers ethnicity no matter what. This is interesting, because here the child has two fathers, one of whom he never knew. If the latter is the one that matters for ethnic ascription, the underlying model is extremely primordialist. A post-interview cross-examination always sought to discover why they had answered in the way that they had. This was to see if they had a rule, or could produce one, and also, where necessary, to point out what appeared to be an inconsistent rule at work (a practice which, on the assumption that peoples cognitive models are consistent, reveals to both researcher and subject that one or more of the questions have been misunderstood). I found a consciously held biopatriline al rule: a child takes the biological fathers ethnic status, often stated in so many words, and quite automatically. The majority had a rather strong version, answering
797
Kazakh, Kazakh, Kazakh. I followed this up with an especially strong challenge in the cross-examination. For example: Do you mean that culture and language dont matter at all? In Question 3 the boy doesnt even know he is Kazakh!, to which they would often respond with some variant of, Thats right, the only thing that matters is the ethnicity of the trcn aav (birth father). The kid may not know it, but he is still Kazakh. It doesnt matter. Mongol-Kazakh intermarriages simply do not happen in my eld-site area, though some occur in Ulaanbaatar. Thus, if my informants did already have a rule governing transmission/acquisition of ethnic statuses, they were applying it to questions they had perhaps never considered before, in much the same manner that one may apply grammatical rules to decode sentences one has never seen. The answers therefore probably reveal deep structure rather than the extraction of non-grammatical (non rule-like) phenomenological patterns extracted from memory. This has the disadvantage that one cannot test the relevance of the rule for actual behaviour by recording the ethnic statuses of the offspring of KazakhMongol marriages. However, intermarriages between two Mongol ethnies (such as Oriankhai and Torguud), though rare, do occur. I encountered two such intermarriages: both of them Torguud husbands with Oriankhai and Khalkha wives, respectively. In each case (1) the parents said the children were Torguud; (2) they agreed instantaneousl y; and (3) the answer was automatic, with no re ection given to the matter. Asked why, they did not respond that it was because the children were being brought up in a Torguud environment (which was certainly the case), but because the father is Torguud. This suggests that Mongols do, indeed, behave in a manner consistent with the answers they gave. Table 2 has family heads (ezen ) only. It is sex-biased, and a smaller sample. But a t-test by sex for the larger sample revealed no signi cant difference between the sexes.3 The ezen sample has the advantage that the effects of kin bias are greatly reduced since there are fewer close-kin among ezen . The trends are all the same, and the percentages quite similar (although the difference between those who answered Kazakh and Mongol to Question 3 is not signi cant with such a small sample).
Table 2. Household heads only Answer Child is Kazakh Child is Mongol Child is erliiz (half-breed ) Total
N = 15
Question 1 15 0 0 15 100%
100%
798
Francisco J. Gil-White
Finally, I must add that these data probably underestimate the numbers of primordialists. In the context of an interview, the effect is such that when interviewed alone, the interviewee s are more likely to give circumstantialist answers. I interpret this as a task demand: respondents, when alone, were intimidated by my overt implications that the answer should be circumstantialist. When asked in a group, on the contrary, they agreed with the right answer from their peers perspective. The second situation has greater ecological validity because ethnic ascriptions are public behaviours. People do not decide in the privacy of their minds to call someone a Mongol or a Kazakh, they commit themselves publicly to these labelling processes.
The models
The following simple taxonomy is useful: the sequence Kazakh, Kazakh, Kazakh corresponds to a hard primordialist Ethnic Transmission and Acquisition Model [ETAM]. These respondents assign a Kazakh ethnic status to the biological child of a Kazakh no matter what the circumstances. The sequence Kazakh, Kazakh, Mongol corresponds to a soft primordialist ETAM. For these respondents ties of blood are paramount but truly extreme circumstances allow them to bend the primordial criterion. The sequence Kazakh, Mongol, Mongol is that of a soft circumstantialist one who believes that childhood enculturation will affect ethnic status. A category that does not appear in the data because the instrument is not designed to pick it up is hard circumstantialist. This model allows for a fully enculturated adult Kazakh, say, to rationally decide to become Mongol, and to succeed on the basis of this decision. I did not concern myself with it because I was trying to see if a much weaker circumstantial criterion could operate. I reasoned that questions about rational choices would be moot if walking and talking like an A from a very early age plus being unaware of being of B descent did not change peoples opinions about their true ethnic status if they were informed about the biological facts. Hard circumstantialists are thus collapsed into the soft circumstantialist category, and I do not know how many of these are hard. But even if all of them were, this would still leave us with less than
Table 3. Ethnic Transmission and Acquisition Models [ETAMs ] Model Hard primordialist Soft Primordialist Soft circumstantialist Total*
*Total has rounding error
799
a fth of respondents espousing a non-primordialist position. If one adds the reasonable expectation that some will not be hard circumstantialists, this evidence looks bad for the rational-ch oice model. Cross-examination revealed that hard circumstantialists certainly do exist: at least two respondents t this description. But an accurate determination of this models relative frequency must await further research. The most important point is this. Given that most respondents have strict primordial models that is to say, nothing can change ones Kazakh status if the biological father is Kazakh an individual Kazakhs claim to Mongol status would fall on deaf ears. Two questions remain to be investigated: (1) are most ethnic communities around the world characterized by similar distributions of ETAMs?; and (2) does this limit peoples ability to make residential/interactional/signalling choices such that what Barth found in Swat is an interesting special case rather than the norm? I hypothesize that the answer to both questions is yes, and shall review further below indirect evidence to bolster my prejudice. 4. Broader implications of these results
800
Francisco J. Gil-White
(for example, the entire literature on pluralistic ignorance; Miller and McFarland 1991), it follows that the performative entailments of an ascriptive cognitive model enjoying a simple plurality will quickly get stabilized at a very high frequency (other things being equal). This is Barths subjectivist perspective with a twist. If most others think that it suf ces for me to decide my ethnicity and announce it, then circumstantialists of the rational-choice school will be correct. If most others think that it suf ces for me to learn some cultural habits and publicly display a few cultural markers, and these conceptions correspond to how outside observers parse the cultural world into dichotomous units, then the culture-area perspective on ethnic groups will be largely correct. On the other hand, if most others will accept an ethnic label X only for persons biologically descended from individuals so labelled, primordialists will have scored their most important point. The majoritarian model of membership criteria as such criteria are held for third parties de nes the social boundar y. A cognitive investigation of ethnicity is thus virtually demanded by the shift to subjectivism that Barth engineered . Social categorization theory and research have taught us that there is personality-based identity (personal identity), and categorical/ascriptive or group identity. Likewise, there are two kinds of attraction (Hogg 1992) : personal attraction (caused by the personality of another), and social attraction (caused by the group-membership of another, through the prototyped stereotype attached to such membership). However, this theory takes the (contextless) process of categorization itself to be decisive (Turner et al. 1987) , and proposes no theory about the different kinds of groups with which people can identify. There are many kinds of group identities : ethnic, kinship, political, religious, gender, class, racial, regional. Each creates an ingroup (where ego is member) and an outgroup, and therefore a social boundary, by stipulating certain conditions that members must satisfy in order to be such. But psychologists have in general treated all group identities and ingroup/outgroup cleavages as resulting from the same, general, stereotype-formation process (for example, Bar-Tal et al. 1989; Leyens/Yzerbyt/Schadron 1994). 4 However, we are probably equipped with specialized psycholog ies (plural) to process the different kinds of social boundaries that recur in the social world. Group identities are probably domain-speci c, like most other aspects of human cognition (Symons, 1992; Tooby & Cosmides 1992). If so, membership conditions in one domain (political groups) may well be different from those in another (ethnic groups). If ethnic categories conceptually delimit what our cognition sees as natural kinds, I would expect the categories to require necessary and suf cient conditions of membership even if governed by Roschian properties on the inside (Gil-White 1999; see Lakoff 1987 for an extended discussion of Eleanor Rochs research). In the case of Jews, for example, having a Jewish mother seems to ful l both necessary and suf cient
801
conditions for being a real Jew (see Chervyakov et al. 1997, for evidence that Jewish descent is both necessary and suf cient). One may still be a bad example of a Jew, if one is an atheist, but nevertheless a real Jew (contrast this with a man who has converted to and practises Judaism but whose parents are Irish; would he still be a Jew if after some time of practising Judaism he became an atheist?) Of course, social-behavioural outputs certainly almost always rest on psychologies more complex than mere categorical structure. However, if categorical thinking in a particular domain is important, then this is the place to begin, for the broad categorical features of cognitive models will constrain the hypothesis-space and the perception -space of actors (Quine 1960; Bloom 1993), making them more susceptible to the acquisition of certain ideas than others concerning the objects so categorized (Boyd and Richerson 1985; Tooby and Cosmides 1992; Boyer 1994). On these basic biases, and through the selective acquisition of ideas, more elaborate cultural models to reason about these objects tend to result. The deepest implications of this perspective on things may be stated thus: (1) If the nature of a social boundary (i.e. the particular type of conditions members must satisfy in the contrasting groups) implicate the insider and outsider as particular kinds of agents, then these perceptions probably in uence behaviour such that interactions within and across one type of social boundary are different from those within and across another type. (2) Furthermore, if there is something recurrent about certain human agglomerations everywhere that makes us identify some groupings and lump them together as ethnic the world over, then there must be some constant psychological features leading humans everywhere to organize themselves into groups with ethnic criteria of membership. (3) Finally, if we perceive the behaviour of such units to have some deep similarities regardless of where they occur, then the psychology which produces these social units must also signi cantly in uence the behaviours of their members. The results reported above should make us uncomfortable with circumstantialism as it now stands and desirous of a cross-cultural cognitive empirical effort. We need to know whether ethnic actors themselves perceive the ethnic world through a primordial (circumstantial) lens, and how this affects their behaviour. If in most communities the majoritarian cognitive model is primordialist (circumstantialist), and such models signi cantly affect behaviour , then it follows that ethnic groups and the processes underlying them will be signi cantly primordialist (circumstantialist). Given that understanding the distribution of ETAMs is the key to many of the issues under contention , and practically mandated by
802
Francisco J. Gil-White
Barths conceptual innovations, it is remarkable that barely any cognitive work in this area exists.
803
Like Thompson, many believe that there is something to the primordialist view, and that if only primordialists could summon more faith in the possibility of systematizing (or at least clarifying) their intuitions, perhaps we could investigate what that is. But he is too sweeping; at least Van den Berghe (1987) has produced an explanation which, whether or not one agrees with it, quali es as a bona- de theory. Circumstantialists have responded with harsh criticism. Primordialists are charged with defending an unscienti c, unsociological and even racist concept (Eller and Coughlan 1993, in an especially super cial article). However, some disagreements are straw-man fabrications by the circumstantialists, who caricature primordialism and make it harder to explore its strengths. For example, they often (wilfully) fail to distinguish between what an ethnic group is to its members psychologically , and the objective reasons why such groups may form (cf. Grosby 1994). As a result, primordialists (who often do make this particular distinction rather clearly) are caricatured as maintaining that ethnic groups are objectively primordial and therefore eternally permanent and impervious to modi cation by circumstances, as well as having impermeable boundaries (e.g. Bonacich 1980; Lemarchand 1986, p. 188; Eller and Coughlan 1993). In the same breath, primordialists are often charged with naturalizing ethnicity (for example, Eller and Coughlan 1993; Jenkins 1996). Such criticisms are in a sense unfair because they do not attempt to put the opposing views on their best footing before debunking them [cf. Tilleys (1997) criticism of Eller and Coughlan (1993)]. In the rst place, they attempt to paint primordialists as analytical naturalizers rather than analysts of naturalizers, and this is pure rhetoric. To insist that actors perceive co-ethnics as sharing biological descent is to describe the manner in which individuals cognize the ethnies they participate in. It does not say that new ethnic groups cannot arise in place of old ones which disappear, nor is it incompatible with this idea. (It does, however, commit its defenders to quali ed statements about the maximum thresholds on the rate of ethnogenesi s and ethnic boundary change. That is, if ethnic actors believe that membership is a matter of shared biological descent, changes in the boundaries of ethnic groups will happen on inter- rather than intragenerationa l timescales; Van den Berghe 1987, p. 27). And if primordialists have implied that there may be something natural about ethnic group formation, it is unclear how this automatically disquali es their views; natural and wrong are not synonymous, and the data will judge. Besides, to advance self-interest as the motivation behind ethnic group formation is no less naturalizing: it merely posits a different kind of nature. That said, perhaps primordialists deserve these criticisms, after all, since they have failed to operational ize adequately and test their views, which are neither unoperation alizable nor untestable the way many primordialists themselves believe. The results reported here suggest a
804
Francisco J. Gil-White
reasonable scienti c basis for a defensible primordialism, which I formulate as a few hypotheses:
Hypothesis 1: a majority of ethnic actors will possess ethnobiolog ical ETAMs. Hypothesis 2: if 1 is true, this will affect ethnic actors perceptions of their interests as well as their motivations for pursuing them vis-vis interests and motivations evaluated with respect to other kinds of group boundaries. By extension, it will affect behaviour. Hypothesis 3: a basic primordial model is part of our innate psychology (Gil-White 1998). Hypothesis 4: if 3 is true, it will act as a built-in bias affecting the cultural/developmental elaboration of the basic model (in terms of culturally transmitted rules about what can and cannot be done with ethnic statuses). There will thus be a non-random distribution of ethnotheories of ethnicity, with a tendency towards rather in exible primordial ETAMs (the Mongol case reported here supports the hypothesis but, obviously, it does not suf ce).
805
explicitly and uncritically advocat ed by Banton (1994). Statements such as these: . . . ethnicity is an essentially modern phenomenon, primarily rooted in urban settings, and intimately tied up with the processes of change introduced by economic and political modernization. Ethnic ties are forged in the competitive struggle of modern politics. In so far as they express traditional attachments, these are constantly rede ned in the light of changing conditions and in response to the political exigencie s of the moment (Lemarchand 1986, p. 188; cf. Wallerstein 1960; Young 1976) reveal the preoccupation with ethnic mobilization as the be all and end all of ethnicity which is so common among modernization theory (or structural) circumstantialists. But explaining the behaviour of, say, a few particular dogs, in some places, is not the same thing as explaining what dogs as a species are or why they emerged evolutionari ly, although the two questions are obviously related. Just as a particular dog is a dog whether or not it has learnt to fetch or walk bipedally on its hind legs, particular human aggregation s may be ethnic whether or not they presently exhibit ingroup solidarity and overt con ict with outgroups. 5 If we believe that African and Asian pastoralists, African and Amazonian rainforest hunter-gather ers and horticulturists, North American and Siberian hunter-gathere rs, New Guinea horticulturists, Australian Aborigines, organize in ethnies as the literature suggests we do then we shall need a broader theory both of ethnic groups and their mobilization. This is because in their cases (and others) urban settings were (and sometimes even are) absent, and the arrival of modern state structures has often ended rather than initiated ethnic mobilization, as in the case of the Nuer and the Dinka (Kelly 1985) and the New Guinea ethnic groups. One has also to consider the predictions the circumstantialist model makes about behaviour. If ethnic actors are instrumentalists, then new ethnic groups should follow shifting interests, arising and disappearing as suddenly as do purely political or territorial alliances; people should spontaneously switch ethnic identity when it becomes convenient; and it should be more common for new ethnicities to spring forth around changing material interests and concerns, than for ethnicities to persist in spite of costs to their members interests. However, if most ethnic actors are ethnobiological primordialists, and if to ally with particular ethnic actors one must usually share membership, then many kinds of alliances of interest will be precluded, given that switching ethnicity cannot be arranged instantaneously and at will. It is also quite plausible that a primordial model of the ethnic world will affect peoples conception of their interests. Much evidence supports primordialist rather than circumstantialist
806
Francisco J. Gil-White
prejudices. Individuals quite often sacri ce economic self-interest for ethnic group goals. For example, in unranked, polyethnic state after state, 6 it has been the case that ethnic parties, rather than class-based parties, have developed , often with perfectly unreconcilable aims. Nigeria, by the time of independen ce, already had Ibo, Hausa and Yoruba-dominated parties; Guyana and Trinidad, Creole and East Indian parties; Sri Lanka, Sinhalese and Tamil parties; and Malaysia, Chinese and Malay parties (Horowitz 1985, p. 9). More illuminating, perhaps, is the failure of explicitly and avowedly non-ethnic parties to remain such in unranked , polyethnic states, especially parties of the left who have it as their central doctrine to be universalistic, advocating the interests of the working class regardless of ethnic origin. Thus, it is striking that The Communist Party has been dominated by Ansaris in the Sudan, by Sinhalese in Sri Lanka, by Greeks in Cyprus, and by Chinese in Malaysia (ibid, pp. 9 10). Similarly, the Communist Party has been captured by different ethnic groups in the different states of India. By the Sikhs in Punjab, by the Ezhava in Kerala, by the Bengalis in Assam, and in Tripura, rst by the indigenous hill people, then by the Bengalis, but never both simultaneously (ibid, p. 10). This is similar to the failure of Communist Parties in Europe to transcend ethnonation al boundaries, nally resorting to nationalistic appeals in order to become electable. Socialists, for their part, have fared no better. In Guinea they were Fulani; in the Ivory Coast, Bt; and in Congo, Mbochi (ibid, p. 10). As Horowitz goes on to detail, many revolts and insurgencies ostensibly inspired by class ideology, have sometimes derived their impetus from ethnic aspirations and apprehensions instead, such that their sole or main participants were members of one particular ethnic group. Such data embarrass extreme circumstantialist predictions such as Pattersons (1975): Where a plurality of allegiances involves a con ict between class interests and other interests, individuals . . . will choose class allegiances over all other allegiances, including ethnic allegiance . He did not test this with cross-cultural data, however, but instead offered a single case-study which is in line with his predictions. Less extreme instrumentalists may argue (with Hechter 1992, p. 273) that class loyalties do not usually win against ethnic loyalties because the problem of collective action is more easily solved in primordial than in instrumental groups, given that in the latter there is greater monitoring and sanctioning by peers. But punishment by third parties (necessary to get rational individuals to incur costs altruistically for the sake of collective action) is a public good that is, punishment requires the punisher to incur certain costs; if others are punishing, one should free-ride on their efforts (if one is rational; see Boyd and Richerson 1992). Thus, here again we must explain why ethnic actors forgo individual interests to coerce each other for the purposes of collective action. They are clearly not acting as straightforward instrumentalists la Economic Man.
807
All of this is not to belittle the circumstantialist concern with structural variables in ethnic mobilization; it is an important topic because ethnies do not always mobilize, and not always in the same ways. Circumstantialists are certainly correct that perceived costs and advantage s to individuals underlie much of their behaviour, and that ethnic statuses are used strategically and politically. But in the light of the evidence (including the data reported here), the idea must be taken seriously that a primordialist view of the ethnic world, by ethnic actors themselves, may constrain the space in which they consider it legitimate to advance their self-interested goals, and may also in uence their choice of goals such that they are not all self-interested. Thinking of the instrumentalism of ethnic actors as both framed and constrained by a eld of possibilities that a primordialist psychology delimits may be a more reasonable and accurate perspective on the problem. The strategic management of multiple identities is one area in which a revised circumstantialism might offer a more theoretically cogent solution. Hogg (1992, p. 94), speaking of groups in the broadest sense of the word, remarks that the relevant social group shifts with circumstances. Ethnicity would appear to be no exception . Several authors have pointed out that people often have more than one ethnic identity, where such identities are organized in a concentric arrangement. Ethnicity . . . is a set of descent-based cultural identi ers used to assign persons to groupings that expand and contract in inverse relation to the scale of inclusiveness and exclusivene ss of the membership (Cohen 1978, p. 387). Many authors echo this sentiment and add that ethnic actors will give salience to whichever of these identities is most relevant in particular circumstances, where relevance is a matter of segmentary opposition la Evans-Pritchard (1968) (for example, Moerman 1965; Horowitz 1975, pp. 118 19, Keyes 1976, pp. 2067). Others argue more generally that available ethnic statuses are used strategically or politically, but avoid a narrow focus on segmentary opposition (Royce 1982, pp. 184 215; Nagel 1993). Ethnic actors, according to this view, will put on the ethnic hat best serving their purposes in particular circumstances. For instance, in contexts of opposition to other Mongol ethnies, Torguuds may want to emphasize their Torguud identity. But Torguuds are also Mongols, and in contexts of opposition to Kazakhs they may want to give salience to the more encompassing Mongol identity. Clearly, such manipulations do take place, but if one can only choose primordially available hats from the ethnic rack, we must revise our views of just how manipulable and instrumental ethnic statuses can be (cf. Nagata 1981). In this manner, circumstantialism and primordialism can be merged into a more sophisticated view of ethnicity. We must also resist the temptation to speak of all group identities as if they were the same thing. True, if I am a Catholic Mexican there may be circumstances where I want to emphasize my Catholicism (for
808
Francisco J. Gil-White
example, with an Irishman) because there is no ethnic identity link that I can establish. But the fact remains: if I am a Catholic Mexican, I can become a Mormon Mexican if I wish. I can unbecome Catholic. But can I unbecome Mexican? Obviously people do not usually discard and reacquire religious identities lightly, but there are other group memberships which are easily discarded. It is easy enough to switch from being Democrat to being Republican and vice versa, as circumstances and ones own changing ideologies make this advantageou s. More than one elected of cial has, in fact, made the switch while in of ce and nobody appears to think of them as wolves in sheeps clothing (Newt Gingrich, for instance, used to be a counterculture Democrat). If people cannot lose and acquire ethnic statuses the way they might other kinds of statuses or identities , the politics of ethnicity will be qualitatively different from other kinds of politics.
809
thus: This was not the way one should be received by a nomad (Haaland 1969, p. 70). For the sake of argument, if we assume the two informants are representative, the fact nevertheles s remains that the offending nomads in question were not called Baggara; there was merely a complaint that they should behave like good nomads if they were going to live that way. By choosing to focus on standards of performance evaluation (certainly of immense relevance to ethnicity), Haaland perhaps unwittingly ceased to make the issue identity, which was ostensibly the point of his investigation (see above quote). He tells us that any similarities between just-nomadized Fur and the Baggara resulted from adaptive constraints in the nomadic lifestyle; these Fur nomads had assimilated neither Baggara culture nor language, and apparently considered themselves Fur (certainly they retained and practised Fur standards of performance). Other Fur apparently agreed. Haaland rst learnt of these new nomads when one of his informants among the sedentary Fur asked him whether I wanted to visit Fur people who live in nomad camps (Haaland 1969, p. 68, emphasis added). There is no evidence of ascriptive change here, which suggests an af rmative to Haalands last question in the quote above: where ethnic ascription is concerned, only in the next generation can a lineage make the transition from the Fur to the Baggara. Barth argued as though Pathans were choosing to become Baluch. Perhaps this is merely a failure to distinguish membership in an ethnie from other social structures. Some sections of some Baluch tribes have traditions of incorporation exible enough to admit new members from other ethnies. Many Pathans had been incorporated into the Marri tribe, whose sections all had such traditions. But note: Of the three main branches of the Marri, the Ghazani contains subsections of various origins, the Loharanis are half constituted by the Shiranis, tracing descent from the Pathan tribe of that name, and the Bijaranis are regarded as predominantly of Pathan origin. Among the Bijaranis the Powadhi section has had the most proli c recent growth . . . their leader is totally virtually autonomous in his relation to the Bijarani leader; and the growth . . . has taken place so predominantly through incorporation of Pathans that it is referred to as Pathan by other Marris, though it is uniformly Baluch-speaking (Barth 1963). This shows that members of an ethnic group can join en bloc the political structures of another, and acquire their culture, without losing their original identity. This should not have misled Barth who elsewhere had noticed Kohistanis incorporating themselves into Pathan political structures (as serfs to the Pathans) without ceasing to be Kohistanis (Barth 1956). That unsolicited remarks by putative Baluchs to the effect that they were really Pathan were forthcoming (Barth 1969, p. 29) should
810
Francisco J. Gil-White
underscore the point: there is both inside and outside recognition of their Pathan status no change yet in ascription or identity. Barths contention that, given the incompatibilities between Pathan and Baluch social structures, incorporated Pathans have a need to absorb Baluch norms, abide by them, and signal this to others is convincing (Barth 1981, pp. 110 11). The public use of the Baluch label could serve the latter purpose, but none of this requires a change in identity, and his own data make me doubt that it happened . It is of course possible, even likely, that several generations after the switch people will forget the original ancestry (or discount it) and regard descendants of incorporated Pathans as really Baluch, making movement of personnel across the ethnic boundary a real possibility. But this is because everybody forgot the origins, there has been intermarriage; or because several generations of ancestors who publicly call themselves Baluch means, in fact, that they are descended from Baluch. In either case, a primordialist model is active, preventing true switches in ethnic identity in the rst generation . Leachs (1977[1954]) work in Burma, which in uenced Barths theoretical focus, exhibits similar primordial nuances in his reported cases of ethnic absorption (even his choice of words seems at variance with the points he was supposedly making). 7 I shall illustrate with one example (pp. 293 97). On the Nogmung, Leach, relying on Barnard (1925), relates that they had adopted the dress and language of the Shan, although some still spoke their native Jinghpaw. Leach closes this case by reference to the (then) current conditions where the Nogmung might be taking up Jinghpaw again, and refers to this process with the phrase The Nogmung are probably becoming Jinghpaw again. His grammar suggests a change of ascription: The Nogmung become Jinghpaw again (which implies that for a while they had been Shan). But in fact the Nogmung were never Jinghpaw to begin with: they were Nogmung, and they spoke Jinghpaw. There is a Jinghpaw tribe that also speaks Jinghpaw, but this is not the case of the Nogmung. Nogmung is the name for a tribe of what, according to Leach, is the Kachin ethnic group (which also includes the Jinghpaw tribe and many others). Perhaps Leach wrote as he did because for many outsiders Jinghpaw and Kachin are synonymous, and he wanted to contrast the Kachins (hill peoples of different sorts), with the Shans (wet rice cultivators who live in the valleys) . From this perspective he is saying that the Nogmung (who are one kind of Kachin) became Shan, and then became Kachin again; his evidence is that they rst acquired Tai (the language of the Shan), then reacquired Jinghpaw. But linguistic change and identity change are not the same; the Nogmung cannot become Kachin simply by speaking Jinghpaw (or Shan, for that matter, by speaking Tai) any more than the Irish became English when they nally adopted the language of their conquerors. And if it is true, as Leach
811
suggests, that after the collapse of the feudal system that made the Nogmung serfs of the Shan these former serfs wished again for their ancestral language, then one should be sceptical about their ever having lost their Kachin identity. His other examples are equally based on linguistic or cultural change, and are silent about identity change or suggest the opposite. Only one case appears to involve unequivocal identity change. It involves Assamese slaves who became Kachin. Two points bear special mention: (1) Leach thinks that there was intermarriage with their Kachin masters; and (2) from the time that they still called themselves Assamese, in 1824, when they were made slaves by the Kachin, to the time when a different source reported that their descendants were called Kachin (a crucial point), in 1925, it is a full century and three or four generations later (depending on life-expectancies and average reproductive age). Another case that might involve identity change concerns Kachins of the Khaphok and Kha-lang tribes who may have become the lok hka Shan serf class. Here too, between the observation of the Kachin dependants of the Shan (1828), to the observation of the lok hka (1925) a full century has gone by. Cases of ethnic absorption reported elsewhere similarly reveal the need for intermarriage and/or several generations. Dinka absorbed by the Nuer are not real Nuer they have a special name: Jaang Nuer (Kelly 1985), and Jaang (or Jieng), incidentally, is what the Dinka call themselves. Likewise, Jok Jok, a Dinka friend and anthropologist, tells me that Nuer who marry into the Dinka ethnie are not considered real Dinka either and are called Nuer-da, which means our Nuer. However, and this is illuminating, their children (who will be descended from at least one Dinka) will be real Dinka. He also knew of a case where a Nuer couple came to live in a Dinka village under the auspices of the chief. Despite the fact that they were accepted as full residential members of the community, they were always Nuer-da and, most importantly, so were their children, for they had no Dinka blood. To my questions, he commented revealingl y, Some identities you can only get by blood. In Yauri, Northern Nigeria, it is common for Gungawa to make the transition into Hausa, but this never happens in one generation , nor do the Gungawa accept foreigners into their ethnie, though they will readily accept as Gungawa the children of foreigners who intermarry (Salamone 1974, pp. 109, 117, 236 37). Some Turkana in the Isiolo area, Kenya, routinely become Samburu, but this is an arduous process which requires rst transitioning through the category Ilgira. The required amount of time for a complete assimilation of Turkana into Samburu then seems to be at least two generations (Hjort 1981, emphasis added). Arabs in Malaysia successfully assimilated themselves into a Malay identity, but only by forcefully pointing out the several generations of intermarriage that had already taken place between them and Malays (Nagata 1981).
812
Francisco J. Gil-White
I have found only one example of complete ethnic absorption in the rst generation . Interestingly, it concerns Kazakhs in Xinjiang near my own eld site. Bessac (1965, pp. 378 79) reported that his informants often stole children from Mongols and Tibetans and adopted them but, in the opinion of the thieves, these children became fully Kazakh upon absorption of Kazakh customs and initiation into Islam. Note, however, that even here this is still not a case of adults making rational choices. If one cannot complete an ethnic switch for oneself, but may ensure it for ones children by providing them with actual blood ties through intermarriage, this places an upper limit to the speed with which ethnogene sis and ethnic dissolution (in its various forms; see Horowitz 1975), can take place. This would make ethnic groups different from other kinds of groups, particularly the political coalitions to which circumstantialists believe ethnic groups are so similar. None of this denies the instrumental manipulations of ethnic statuses that Barth documented. For his Pathans-turned-Baluch, being really Pathan was obviously of minuscule practical signi cance. But it is nevertheless important to clarify the cognitive framework within which these boundary crossings are taking place: ethnic actors have primordialist ETAMs. In Swat, considerable playing with the ethnic labels is allowed, so those who wish to articulate with particular networks may use the labels that signal the expectation s associated with them. But if ethnic actors are in general primordialists, and also essentialists, (that is, ones ethnicity implies an inalienable essence, Gil-White 1999), then I would expect the Barthian exible signalling system to be the exception, and to nd that in most times and places, one cannot simply grab a new ethnic label and begin interfacing with another ethnic community as a full member. Barth himself is equivocal about the question of primordiality, admitting that . . . ethnic membership is at once a question of source of origin as well as of current identity (1969, p. 29). Horowitz (1975, pp. 113 14) makes a stronger primordial point, but also equivocally. He argues that although ethnic identity is generally acquired at birth, it is a matter of degree. Some ethnic groups, he says, change their boundaries quickly, deliberately, and noticeably. He concedes that usually changes in identity require a generation or more to accomplish by means of intermarriage and procreation but believes that Linguistic or religious conversion will suf ce in some cases. Then he says, There are ctions about, and exceptions to, the birth principle for most ethnic groups. Ethnicity thus differs from voluntary af liation, not because the two are dichotomous, but because they occupy different positions on a continuum. What exceptions? Horowitz does not give any examples or citations. Moreover, he does not honour his analytic de nition in practice: he claims elsewhere in the piece that Sikhs started out as a religious sect but
813
became an ethnie as membership requirements became ethnobiological. This concedes what practically everybod y, circumstantialist or primordialist, has assumed ever since Weber (1978, p. 389) rst offered a de nition: members of an ethnie will represent themselves publicly as a descent group. This is important. Circumstantialists and primordialists may disagree on the interpretation, but they apparently agree with the description of ethnies. The literature shows that circumstantialists believe ethnies are characterized by public ideologies of common descent.8 So both camps at least agree on how to pick out ethnic groups in the world. What ethnies are, then, is not merely a semantic but a scienti c issue. If scholars were merely divided over the proper referents of a given label here: ethnie the matter would be trivial and easily solved by adopting a technical de nition that eliminated the ambiguities. But the referents are not in question: all scholars know one when they see one: ethnic groups are agglomerations of people who, at a minimum, represent themselves as vertically reproducing historical units implying cultural peoplehood. What divides scholars is how much attention they pay to such public representations, and the analytical interpretation they have for the agreed set of referents. 9 Is the primordialism that everybody sees in ethnic actors own public representation of the ethnie merely an instrument of their mobilization? Or is it a prism for viewing social life that frames that life, constraining the actors instrumental choices? This is the crux of the debate. What comes rst, mobilization or categorization? My Mongolian data if representative of the world make me sceptical that instrumental choices concerning interaction/coalition can easily precede ethnic categorization. Coming back to Horowitz, ags should go up when in his very own discussion Sikhs go from a religious sect to an ethnie as membership gets linked with descent. This suggests that in Horowitzs intuitions vs. his academic de nitions membership in ethnic groups is not, in fact, sometimes a matter of religious conversion. Rather, the extent to which religious groups abandon conversion for descent as a criterion of inclusion is also the degree to which he feels inclined to think of them as ethnies. The intuitive de nition of an ethnie implicitly used by Horowitz (and scholars everywhere ) is therefore not fuzzy at all it is not ethnie: groups which may recruit members through descent, and/or religious conversion, and/or linguistic assimilation, and/or voluntary af liation, and/or political integration. What is fuzzy is the membership of any particular object in the category ethnie. This will depend on the proportion of ethnobiolog ical primordialists in the group and in relevant outside groups. Obviously, ethnic groups have not existed since the time of Adam. An incomplete process of boundary change therefore entails that, of those persons enjoying, say, group ascription A, less than the overwhelming majority walk around with biological models of membership
814
Francisco J. Gil-White
in their heads, and this group therefore may be thought of as having less than 100 per cent membership in the category ethnie, and thus as partially ethnic, or as a group undergoing ethnogenes is. The birth principle in the conception of ethnies appears to me quite unequivocal. Otherwise, why are Jews controversial as an ethnie? If some ethnies really honoured linguistic or religious conversion for membership, as Horowitz contends, then Jews would be one of these groups. They are controversial as an ethnie precisely because religious conversion can apparently gain an individual entry into the group. Ethnic actors themselves sometimes use overlapping labels that do not clearly delimit easily recognizable and bounded ascriptive groups, and much has been made of this (Levine and Campbell 1972, pp. 89 99). These are typically places where processes of ethnogenesis, ethnic dissolution, or ethnic absorption (or all of them) are taking place. The world is not static. If ethnic boundaries change, this does not mean that ethnic groups are fake, nor does it make ethnic actors non-primordialist. When processes of boundary-change are under way, people may have more than one simultaneous identity, because they not only use their new labels but they also retain the old ones. These processes are incompatible with a primordialist position as advanced here only if they take place on intra-genera tional timescales the timescales in which political organization of various sorts typically waxes and wanes. By the same token, if these changes take place only on inter-generationa l timescales, they are not compatible with a circumstantialist view that sees ethnic groups simply and merely as one manifestation among many of political organization.
Hypothesis 5: Real changes in the ethnic status of a lineage will require either blood ties through intermarriage, several generations of instrumental labelling, or both. Individual actors will not be able to effect such changes in status merely by changing their behaviour and purposefully displaying a new label.
5. Conclusion I have argued that what distinguish es ethnic groups from other kinds of groups is that ethnic actors conceive membership in terms of categorical descent: biological descent from those possessing a label implying a given cultural essence or peoplehood . This is distinct from kinship (pace Van den Berghe 1987), which is neither necessary nor suf cient for ethnicity because (1) with unilateral ETAMs two people sharing common descent by kinship (for example, same grandfather) could have different ethnic statuses; (2) there are cases of people joining the kinship structures of ethnic outgroups without their obtaining membership in the ethnie
815
(Barth 1963; Hjort 1981); and (3) the ethnic rhetoric of kinship is usually that of a founding father myth or end-point common-origin myth, not with the tracing of actual genealogies (that happens in what should properly be called kinship: lineages and clans). The rational-ch oice version of the circumstantialist school amounts to saying that people who do not really believe themselves to share common descent will nevertheles s participate in collective self-delusion because pretending to share such descent is conducive to their common mobilization, which is desirable as it serves common, objective interests, rationally identi ed. The primordialist position states that (1) whatever notions of common descent ethnic actors already have will constrain and guide their behaviour , so that they will not easily invent new myths of common descent, where there were none, in order to mobilize with relevant others whose interests they believe they share; and (2) ethnic actors will perceive common interests with those with whom they already assume shared descent. I close in the same spirit as Nagata (1981, p. 111). At the risk of taking a position on the fence, it would seem that both the circumstantialist and the primordial approaches to ethnicity can be accommodated. If the truth is up on the fence, I say perch on it. Much of what circumstantialists focus on is important to ethnic processes. But to discover the degree and the ways in which primordialists and circumstantialists are right or wrong, we need to carry out cross-cultural research on the ETAMs, and see whether such cognitive models affect behaviour. This was necessary all along: to focus on the microprocesses by means of which interests are identi ed, statuses assigned, and behavioural choices made (Bentley 1987, p. 26). This is the challenge: to see how a psychology of primordially de ned ethnicity (if there is one) produces boundaries to the eld of action in which individuals may rationally calculate their self-interest (to the extent that they do), and/or produces biases to acquire irrationally certain kinds of ideas (if it does), and how different historical and social contexts tip social processes one way or the other. This is less tidy and elegant, perhaps, than saying people are either primordialists or circumstantialists, that they are either emotional or rational. But if human beings are not tidy, we should not tidy up after them in our theories which does not argue against ltering the noise out of our theories, but merely against pretending that real causes are noise just because they complicate our work. Acknowledgements I am grateful to Robert Boyd for planting the seed that made this article possible, and to his students in the Ethnicity and Evolution seminar for valuable discussions. Allen Johnsons suggestions also proved most
816
Francisco J. Gil-White
helpful. I thank my friends Tsoodol and lzii, with whom I discovered the methodology employed here, for their kind hospitality and friendship. And, nally, I thank my respondents in the Torguud Bangyakha n clan for their patient understanding and cooperation with my research. Notes
1. The subjective emphasis has since been explored by other anthropologists, though not always in the same ways. For example, Mitchell (1974) used criteria of subjectively felt social distance, and this is not the same thing as label-using criteria. Okamura (1981) calls it the cognitive dimension of situational ethnicity. However, no work on peoples cognitive models concerning ethnicity (the basic structure of their reasoning about it) has been done. 2. NOTE: The Mongolian word ndesten, translated as ethnie or ethnicity, is one that locals regularly apply for such groups as Mongols and Kazakhs. It is a good match to our own usage and even has similar polysemic ambiguities. For example, some informants used it for the Mongol tribes, but others preferred to refer to these with yastan which roughly translated is tribe or sub-ethnic group. The applicability of ndesten to the Mongol tribes generated some controversy among them, but not when applied to the Kazakh/Mongol contrast. The word may sometimes be used to denote groupings that in the West are called races but which are also uncertainly distinguished from ethnic groups. 3. Question (2) dummy means: males = .18, females = .16, p =.86; Question (3) dummy means: males = .43, females = .39, p = .75 4. This affects not only the experiments they conduct, but their generalization s. For example, Rodkin (1993, p. 633) discusses race and limits himself to the white-black race relationship in the US, but believes it should generalize to the construction of ethnic and gender differences. 5. Despres (1975, p. 196) provides the concepts of ethnic population and ethnic group, where the latter corresponds to an ethnic population politically organized. However, even this is insuf cient because one can conceive of diffuse mobilization s that occur without a political centre (for example, many forms of discrimination, selective raids directed more often against outgroup members in pre-state environments, spontaneous urban riots) which do not then cease to be ethnic mobilization. 6. A ranked, polyethnic state has different ethnic groups occupying different structural positions (different castes or socio-economic classes). In an unranked, polyethnic state the various ethnic groups are well represented throughout the socio-economic structure (Horowitz 1985). 7. With respect to identity, Leachs use of categories is confusing, to say the least. First he uses the phrase apparent change of cultural identity (p. 40; emphasis added) and refers the reader to the appendix (pp. 2937) for documentation. However, the appendix (which relies on other sources) is entitled Some documented cases of linguistic change (emphasis added). Finally, the subheadings for the different cases are not linguistic but essentialist. Thus we see Jinghpaw become Shan and Assamese become Jinghpaw, rather than such-andsuch speakers acquired such-and-such language . Such essentialism suggests ethnic identity switches but, in fact, the data show linguistic and cultural change and are either silent about identity change or suggest the opposite. Leach wrote as though he thought cultural practice, language and ethnic identity were coextensive, but the point of his book seems to have been, among other things, to dispute this! All this must make the interested reader wonder whether circumstantialists are reading Leach and Barth, or merely ritually quoting citations that they blindly assume contain proof of the received circumstantialist wisdom. 8. For example, Bonacich (1980), after debunking at length the notion that ethnic groups are objectively primordial, concedes, True, they are social phenomena which call upon primordial sentiments and bonds based upon common ancestry. Likewise, Eller and
817
Coughlan (1993), though they clearly think little of the primordialist position, nevertheless state, In many parts of the world, but perhaps best documented in Africa, new ethnic identities and groups are being created which claim [emphasis added] . . . primordial status. Patterson (1975), who presents the most extreme, individualistic, self-interest-maximiser model of ethnicity, nevertheless concurs that, should members subjectively assume the existence of such mythical [primordial] bases, the salient condition of ethnicity is met. Finally, Barth originally admitted that . . . ethnic membership is at once a question of source of origin as well as of current identity (1969, p. 29). 9. The problem is thus akin to the controversy in the biological sciences over what a species is. All biologists agree what organismal aggregates they want to call species, but providing a scienti c de nition of such units has been dif cult and continues to divide biologists (Ridley 1993, ch. 15). However, an attempt to provide an analytical description of species has given biologists insights into the natural world, and also into the human brains handling of natural categories. I expect the same will be true of an even moderately successful effort to de ne ethnicity from the analytical and causal point of view.
References
BANTON, M. 1994. Modelling ethnic and national relations, Ethnic and Racial Studies, vol. 17, no. 1, pp. 119 BARNARD, J. T. O. 1925 The history of Putao, Journal of the Royal Central Asian Society , vol. XV BAR-TAL, D. et al. (eds) 1989 Sterotyping and Prejudice, New York: Springer-Verlag BARTH, F. 1956 Ecologic relationships of ethnic groups in Swat, North Pakistan, American Anthropologist, vol. 58, pp. 1079 1089 1963 Ethnic processes on the Pathan-Baluch boundary, in G. Redard (ed.), IndoIranica, Wiesbaden (ed.) 1969 Ethnic Groups and Boundaries: The Social Organization of Cultural Differences, Boston, MA: Little Brown & Co. 1981[1969] Pathan identity and its maintenance, in F. Barth (ed.), Features of Person and Society in Swat. Collected Essays on Pathans: Selected Essays of Fredrik Barth, Volume II, London: Routledge & Kegan Paul 1994 Enduring and emerging issues in the analysis of ethnicity, in H. Vermeulen & C. Govers (eds), The Anthropology of Ethnicity: Beyond Ethnic Groups and Boundaries, Amsterdam: Het Spinhuis BENTLEY, G. C. 1987 Ethnicity and practice, Comparative Studies in Society History, vol. 29, pp. 24 25 BESSAC, F. B. 1965 Co-variation between interethnic relations and social organization in Inner Asia, Papers of the Michigan Academy of Science,Arts, and Letters, vol. L, Ann Arbor, MI: University of Michigan Press BLOOM, P. 1993 Language development, Handbook of Psycholinguistics, Academic Press Inc. BONACICH, E. 1980 Class approaches to race and ethnicity, Insurgent Sociologist, vol. 10, pp. 924 BOYD, R., and RICHERSON, P. 1985 Culture and the Evolutionary Process, Chicago, IL: University of Chicago Press 1992 Punishment allows the evolution of cooperation (or anything else) in sizable groups, Ethology and Sociobiology , vol. 13, pp. 17195 BOYER, P. 1994 The Naturalness of Religious Ideas, Berkeley, CA: University of California Press CHERVYAKOV, V., GITELMAN, Z., and SHAPIRO, V. 1997 Religion and ethnicity: Judaism in the ethnic consciousness of contemporary Russian Jews, Ethnic and Racial Studies , vol. 20, no. 2, pp. 280 305
818
Francisco J. Gil-White
COHEN, R. 1978 Ethnicity: problem and focus in anthropology, Annual Review of Anthropology, vol. 7, pp. 379 403 DESPRES, L. A. 1975 Toward a theory of ethnic phenomena, in L. A. Despres (ed.), Ethnicity and Resource Competition in Plural Societies, The Hague: Mouton, pp. 187208 ELLER, J. D., and COUGHLAN, R. M. 1993 The poverty of primordialism: the demystication of ethnic attachments, Ethnic and Racial Studies, vol. 16, no. 2, pp. 185202 EVANS-PRITCHARD, E. E. 1968 The Nuer, a Description of the Modes of Livelihood and Political Institutions of a Nilotic People, New York: Oxford University Press GEERTZ, C. 1963 The integrative revolution: primordial sentiments and civil politics in the new states, in C. Geertz (ed.), Old Societies and New States, New York: Free Press GIL-WHITE, F. J. 1999 Are Ethnic Groups Species to the Human Brain? Essentialism in our Cognition of some Social Categories, unpublished manuscript GOODENOUGH, W. H. 1965 Rethinking status and role: Towards a general model of cultural organization of social relationships, in P. Kunstadter (ed.), The Relevance of Models for Social Anthropology , London: Tavistock Publications GROSBY, S. 1994 The verdict of history: the inexpungeable tie of primordiality a response to Eller and Coughlan, Ethnic and Racial Studies, vol. 17, no. 1, pp. 164 71 HAALAND, G. 1969 Economic determinants in ethnic processes, in F. Barth (ed.), Ethnic Groups and Boundaries, Boston, MA: Little, Brown and Co HEATHER, P. 1996 The Goths , Cambridge, MA: Blackwell HECHTER, M. 1986 Theories of ethnic relations, in J. J. Stack (ed.), The Primordial Challenge: Ethnicity in the Contemporary World, New York: Greenwood Press, pp. 13 24 1992 The dynamics of secession, Acta Sociologica, vol. 35, pp. 267 83 HENRICH, J., and BOYD, R. 1998 The evolution of conformist transmission and the emergence of between-group differences, Evolution and Human Behavior, vol. 19, no. 4, pp. 215 41 HJORT, A. 1981 Ethnic transformation, dependency and change: The Ilgira Samburu of Northern Kenya, in J. G. Galaty and P. C. Salzman (eds), Change and Development in Nomadic and Pastoral Societies, Leiden: E.J. Brill HOGG, M. A. 1992 The Social Psychology of Group Cohesivenes: From Attraction to Social Identity, Washington Square, NY: New York University Press HOROWITZ, D. L. 1975 Ethnic identity, in N. Glazer and D. P. Moynihan (eds), Ethnicity: Theory and Experience, Cambridge, MA: Harvard University Press, pp. 111 40 1985 Ethnic Groups in Con ict, Berkeley, CA: University of California Press ISAACS, H. R. 1975 Basic group identity: the idols of the tribe, in N. Glazer and D. P. Moyniham (eds), Ethnicity: Theory and Experience, Cambridge, MA: Harvard University Press JENKINS, R. 1996 Ethnicity etcetera: social anthropological points of view, Ethnic and Racial Studies, vol. 19, no. 4, pp. 807 22 KELLY, R. 1985 The Nuer Conquest: Structure and Development of an Expansionist System , Ann Arbor, MI: University of Michigan Press KEYES, C. F. 1976 Towards a new formulation of the concept of ethnic group, Ethnicity , vol. 3, pp. 202 13 KHAZANOV, A. 1994 Nomads and the Outside World , (2nd edn), Madison, WI: University of Wisconsin Press LAKOFF, G. 1987 Women, Fire, and Dangerous Things: What Categories Reveal About The Mind , Chicago, IL: University of Chicago Press LEACH, E. R. 1977 [1954] Political Systems of Highland Burma, London: London School of Economics and Political Science LEMARCHAND, R. 1986 Ethnic violence in tropical Africa, in J. F. J. Stack (ed.), The Primordial Challenge: Ethnicity in the Contemporary World , New York: Greenwood Press, pp. 185 205 LEVINE, R. A. and CAMPBELL, D. T. 1972 Ethnocentrism: Theories of Con ict, Ethnic Attitudes and Group Behavior, New York: Wiley
819
LEYENS, J. P., YZERBYT, V. and SCHADRON, G. 1994 Stereotypes and Social Cognition , London: Sage MEDIN, D. L., and ORTONY, A. 1989 Psychological essentialism , in S. Vosniadou and A. Ortony (eds), Similarity and Analogical Reasoning, Cambridge: Cambridge University Press MILLER, D. T. and MCFARLAND, C. 1991 Why social comparison goes awry: the case of pluralistic ignorance, in J. Suls and T. Ashby (eds), Social Comparison: Contemporary Theory and Research, Hillsdale, NJ: L. Erlbaum Associates MITCHELL, J. C. 1974 Perceptions of ethnicity and ethnic behaviour: an empirical exploration, in A. Cohen (ed.), Urban Ethnicity , London: Tavistock publications MOERMAN, M. 1965 Ethnic identi cation in a complex civilization : Who are the Lue?, American Anthropologist, vol. 67, pp. 1215 1230 1968 Being Lue: Uses and abuses of ethnic identi cation, in J. Helm (ed.), Essays on the Problem of Tribe, Seattle, WA: University of Washington Press MOTYL, A. J. 1997 Reimagining Primordialism: Constructing an Alternative to Constructivism, unpublished manuscript NAGATA, J. 1981 In defense of ethnic boundaries: The changing myths and charters of Malay identity, in C. F. Keyes (ed.), Ethnic Change, Seattle & London: University of Washington Press NAGEL, J. 1993 Constructing ethnicity: Creating and recreating ethnic identity and culture, in N. Yetman (ed.), Majority and Minority: The Dynamics of Race and Ethnicity in American Life, Boston, MA: Allyn and Bacon OKAMURA, J. Y. 1981 Situational ethnicity, Ethnic and Racial Studies, vol. 4, no. 4, pp. 453 65 PATTERSON, O. 1975 Context and choice in ethnic allegiance : a theoretical framework and a Caribbean case study, in N. Glazer and D. P. Moynihan (eds), Ethnicity: Theory and Experience, Cambridge, MA: Harvard University Press, pp. 30549 QUINE, W. V. O. 1960 Word and Object, Cambridge, MA: MIT Press RIDLEY, M. 1993 Evolution , Boston, MA: Blackwell Scienti c Publications RODKIN, P. C. 1993 The psychological reality of social constructions, Ethnic and Racial Studies , vol. 16, no. 4, pp. 633 56 ROOSENS 1994 The primordial nature of origins in migrant ethnicity, in H. Vermeulen and C. Govers (eds), The Anthropology of Ethnicity: Beyond Ethnic Groups and Boundaries, Amsterdam: Het Spinhuis ROYCE, A. P. 1982 Ethnic Identity, Bloomington, IN: Indiana University Press SALAMONE, F. A. 1974 Persistence and Change in Ethnic and Religious Identity in Yauri Emirate, North-Western State, Nigeria , New Haven, CT: HRAFlex Books, Human Relations Area File, Inc SHILS, E. 1957 Primordial, personal, sacred, and civil ties., British Journal of Sociology , vol. 8, no. 2, pp. 130 45 STACK, J. F. J. 1986 Ethnic mobilization in world politics: the primordial perspective, in J. F. J. Stack (ed.), The Primordial Challenge: Ethnicity in the Contemporary World, New York: Greenwood Press, pp. 1 11 SYMONS 1992 On the use and misuse of Darwinism in the study of human behavior , in J. H. Barkow, L. Cosmides, and J. Tooby (eds), The Adapted Mind: Evolutionary Psychology and the Generation of Culture, New York and Oxford: Oxford University Press, pp. 137 62 THOMPSON, R. H. 1989 Theories of Ethnicity:A Critical Appraisal, New York: Greenwood Press TILLEY, V. 1997 The terms of the debate: Untangling language about ethnicity and ethnic movements, Ethnic and Racial Studies, vol. 20, no. 3, pp. 497 522 TOOBY, J. and COSMIDES, L. 1992 The psychological foundations of culture, In J. H. Barkow, L. Cosmides and J. Tooby (eds), The Adapted Mind: Evolutionary Psychology and the Generation of Culture, New York and Oxford: Oxford University Press pp. 62737
820
Francisco J. Gil-White
TURNER, J. C. et al. 1987 Rediscovering The Social Group: A Self-Categorization Theory , Oxford and New York: Blackwell VAN DEN BERGHE, P. 1987 [1981] The Ethnic Phenomenon , Westport, CT: and London: Praeger WALLERSTEIN, I. 1960 Ethnicity and national integration in West Africa, Cahiers dtudes Africaines, vol. 3, no. 1 WEBER, M. 1978 Economy and Society: An Outline of an Interpretive Sociology , Berkeley, CA: University of California Press YOUNG, C. 1976 The Politics of Cultural Pluralism, Madison, WI: University of Wisconsin Press
FRANCISCO J. GIL-WHITE is a PhD candidate in Anthropology at the University of California, Los Angeles. ADDRESS: Department of Anthropology, Box 951553 , University of California at Los Angeles, 405 Hilgard Avenue, Los Angeles, CA 900951553, USA. ([email protected])