Ludwig Wittgenstein (1953,1:66-67) first cracked the classical edifice by introducing the concepts of ‘family resemblances’, ‘centrality’, and ‘gradience’. He observed, first of all, that the members of some categories did not all share some properties in common (his most famous example is the category game) and concluded that what made such elements members of the category was not the satisfaction of a set of necessary and sufficient conditions, but rather that each possessed ‘enough’ of a set of properties that were characteristic (but not ‘necessary and sufficient’) of members of the category.

Think of the English category religion. There is a chain of family resemblances linking its members, and so long as a token has enough of these family resemblances (it doesn’t matter which), it will be a member of the class. This category includes such things as Shintoism, Shamanism and even Confucianism, at one end, and Islam, Tibetan Buddhism, and Zoroastrianism at the other. The promotion of a moral code, the performance of magic (or the manipulation of unseen forces with no clear physical explanation), stereotyped rituals, belief in supernatural entities, belief in the persistence of individual identity after bodily death (i.e. the soul), an expert class of practitioners (priesthoods), and an account of Creation are all characteristic of religion, but examples of members of the class which lack any of these can be found. Satanism, promotes an immoral code and Shamanism does not promote one either way. Zen Buddhism and Taoism perform no magic. Animism typically has no priests. Etc…There is no set of properties that all members of the class share. Notice, for example, the misgivings of an expert on Shamanism:

For a time it seemed to me best to write a book somehow without the words ‘religion’ or ‘shamanism’. ‘Religion’ seemed wrong for ideas and beliefs which are never set out as general theory and make use of relatively few abstract concepts, for which there is no holy founder, no organized institution, no moral dogmas, and no authoritative corpus of books. Above all, there is no tortuous justification of earlier beliefs enshrined in ancient texts.—Humphrey (1996)

This brings us to another of Wittgenstein’s contributions: centrality. The above quote reveals Humphrey’s intuitions (and mine, and probably those of most competent English speakers) that practices which are institutional, theological, historicized, and moral (in addition to being ritualized, magical, and supernatural) are more comfortably described as ‘religion’. Nevertheless, ritualized practices that attempt to influence, and codify reverence for, supernatural beings and forces (without the above additions, as, for instance shamanism) still strike us as somehow ‘religion’, if less comfortably so. Islam, then, is a ‘central’ (i.e. more typical) member of the category, whereas shamanism is not. Classical categories, as traditionally understood, did not allow for this because if every member equally satisfied all the defining criteria, then there wasn’t thought to be any basis on which to expect that some members would be "more equal than others."

Finally, Wittgenstein observed that at least some categories were characterized by ‘gradual’ membership. For example, how do you crisply separate the set of poor people from the set of rich people? You can’t, not crisply. The same goes for young vs. old, blue vs. green, tall vs. short, etc. Some people are clearly young (babies), and some old (nonagenarians), but the boundaries of these categories are not clear. To age is to gradually lose membership in the category young and gain it in the category old. Zadeh (1965) formalized these intuitions by generalizing set theory to include such ‘fuzzy sets’.

Eleanor Rosch pioneered experiments that could systematically investigate category structure, and synthesized advances by Wittgenstein and others into a theoretical framework that revolutionized psychology. She introduced the term ‘prototype’ to refer to the ‘cognitive reference point’ or most central member (or subcategory of members) of a category.

Rosch showed that a variety of experimental techniques involving learning, matching, memory, and judgments of similarity converged on cognitive reference points. She developed…experimental paradigms for investigating categories of physical objects. In each case, asymmetries (called prototype effects) were found: subjects judged certain members of the categories as being more representative of the category than other members. For example, robins are judged to be more representative of the category bird than are chickens, penguins and ostriches, and desk chairs are judged to be more representative of the category chair than are rocking chairs, barber chairs, beanbag chairs, or electric chairs. The most representative members of a category are called "prototypical" members.—Lakoff (1987:41)

Eventually, Rosch (1978) came to recognize that demonstrating prototype effects for a category did not in fact reveal its structure, for different structures are consistent with prototype effects. In particular, showing that there are typicality effects in people’s appreciation of various members of a category, or differential response-times in their identification of category members, for example, does not necessarily argue for a probabilistic categorical structure as in young people. The most forceful presentation of this argument is in Armstrong, Gleitman, and Gleitman (1983) where ratings of typicality and reaction times were examined for the category odd number. It was found that indeed some numbers, such as 3 or 7, were felt by subjects to be more representative of the category odd number than others, such as 109 and 2003, but "if subjects are asked directly whether typical odd numbers are or can be more odd than atypical ones, they will flatly deny it" (Keil 1989:30, my emphasis). In spite of this, as Rosch cogently emphasized, prototypical effects do constrain the range of possible category structures to those which can account for such effects. She also maintained that necessary and sufficient conditions (i.e. the classical view) cannot.

A representation of categories in terms of conjoined necessary and sufficient attributes alone would probably be incapable of handling all of the presently known facts, but there are many representations other than necessary and sufficient attributes that are possible.—Rosch (1978:40-41)

This last point is less defensible. I will try to convince you below that, despite their demonstrable prototype effects, natural kinds are classical categories, with necessary and sufficient conditions of membership.

How important is shared appearance in the construction of categories? Medin & Ortony (1989) note that Goodman (1972) and Watanabe (1969) have given formal proofs that, on the basis of shared predicates, all pairs of entities are equally similar. The reason is that an infinite number of things can be predicated about any entity, and infinity is not a number that can be subjected to relative comparisons such as "A has more predicates in common with B than it does with C." Thus, they argue that ‘similarity’ should be understood "not in terms of logically possible shared predicates but in the more restricted sense of represented predicates" (Medin & Ortony 1989:182). Certain things seem more similar to each other than to other things, because our brains represent them that way, effectively filtering out many logically possible predicates as immaterial to categorization. "For example, both tennis balls and shoes share the predicate ‘not having ears’, but it is unlikely that this predicate is part of our representation of either tennis balls or shoes" (ibid. p.182). They believe that surface similarity serves as a good heuristic to search for deeper or ‘hidden’ properties, which are the ultimate arbitrators of categorization.

I agree with M&O that our judgments of similarity must be theory-rich, and that "psychological similarity may depend on categorization rather than the other way around" (Rips 1989:22), but I think they may have overemphasized the extent to which such theory is a matter of ‘surface’, ‘visible’ properties vs. ‘deep’ or ‘hidden’ properties. With artifacts, for example, it is their relation to humans that brings an infinite number of possible predicates down to a few relevant ones. Their example is a case in point; tennis balls get thwacked by rackets in tennis courts, and shoes are worn on feet, and this brings to the fore the relevance of predicates such as "shoes have an opening (for sliding the foot into)", "tennis balls are brightly colored (to improve their visibility to the players who must thwack them)", etc. These are not ‘hidden’ properties; we usually have no problem identifying the uses of artifacts because people are using them all the time, and it is their uses which ground our standards of similarity (cf. Gelman 1988). M&O’s surface/depth distinction is better suited to our theories of natural kinds, as they themselves concede:

…psychological similarity is tuned to those superficial properties that are likely to be causally linked to a deeper level. This is particularly likely to be true with respect to natural kinds.—Medin & Ortony (1989:186)

Their theoretical proposal, ‘psychological essentialism’, is the idea that humans believe that natural kinds have constitutive and inalienable hidden ‘essences’. This theory has the advantage of explaining why dissimilar objects can be included in the same category (i.e. when they are believed to share the same ‘hidden’ essence).

None of this commits us to the belief that essences exist. One may certainly argue that, for some natural kinds such as ‘substances’ and ‘elements’, their chemical or atomic constitutions may qualify as a specification of a legitimately real essence; but the Darwinian revolution made it clear that biological natural kinds cannot be characterized as having real essences (Mayr 1964). This is because the nature of genetic variation and the evolutionary process are such that, for any given genetic locus, either (1) there is variation now, or (2) there will be variation at some point in the future, and such variation does not ipso facto force an unusual variant or mutant out of the category which names the population (in fact, typically it does not). This, however, does not deny that we may think in essentialist terms. ‘Psychological essentialism’, as this proposal has been baptized, is "not the view that things have essences, but rather the view that people’s representations of things might reflect such a belief (erroneous as it may be)" Medin & Ortony (1989:183; original emphases). It is a hypothesis about a ‘naïve ontology of natural kinds’—i.e. about how people think about natural kinds, and not about the ontological status of such things.

Perhaps the most interesting suggestion in M&O’s ‘psychological essentialism’ is the idea that a natural kind may have an ‘essence placeholder’. In other words, that with regard to a natural kind, we will assume the existence of an essence which makes the thing the thing that it is, even when we don’t have the first clue as to what that essence is. Thus, the essence placeholder may be empty or full, but we in any case always have the placeholder. According to Medin & Ortony (1989:184-185), the essence placeholder

…might be filled with beliefs about what properties are necessary and sufficient for the thing to be what it is. In other cases it might be filled with a more complex, and possibly more inchoate, ‘theory’ of what makes the thing the thing that it is. It might, additionally, contain the belief (or a representation of the belief) that there are people, experts, who really know what makes the thing the thing that it is, or scholars who are trying to figure out exactly what it is. Just as with theories, what the placeholder contains may change, but the placeholder remains.

To me it seems the placeholder either has predicates constituting a set of properties, or is empty. The belief that experts may know the essence is not an idea that is in the placeholder, but rather a belief that something is supposed to go in it, and that other people know what that is; or else that other people at least know when something has the essence (whether or not they can tell you what the essence is). Recall here the ‘fool’s gold’ example, where those of us who don’t know the first thing about chemistry accept that some people know why such a thing is not gold. What is paramount here is not that there be something in the placeholder, but that the existence of the placeholder––the assumption of an essence––implies a particular way of conceptualizing the category. The idea of less-than-full essence placeholders receives support from studies with children, for "Young children have impoverished scientific knowledge, yet firmly believe that [natural kind] category members have more in common than meets the eye" (Gelman & Markman 1986:199).

I think it is more useful to think of essence placeholders as ‘not full’, rather than empty. For whatever content is in our essence placeholder for a particular natural kind, our brains probably always expect that more hidden properties can be learned about it (this assumption about natural kinds has allowed us to develop sciences about them (Schwartz 1978:571-573 note: get this reference form Keil, mapping mind). Natural living kinds, in any case, do seem to come with essence placeholders that are not entirely empty (at least at the generic-species level), as revealed in children´s reasoning about development. Gelman & Wellman (1991) investigated children’s views of development by testing to see if they assume that common membership in a species implies a common, innate, developmental potential.

…a tiger cub has the potential to grow into something large and fierce, even though when born it is small and helpless. To explain developmental changes of this sort, we as adults often appeal to something like an intrinsic category essence that is responsible for how they grow…the belief in an essential nature or a determining but [at an early age] non-manifest predisposition.—Gelman & Wellman (1991:230)

The children were presented with examples of animals (e.g. tiger) that were raised with another kind of animal (e.g. horses) and in complete absence of their own kind. To questions concerning the form and behavior of these animals as adults, children relied more on category than on environment. For example, they answered that a tiger raised with horses (and never having seen another tiger after the early transfer) would display tiger traits and behaviors as an adult rather than horse traits and behaviors. Children seem aware that the nature of an animal is relatively impervious to the environment of development, and that adult traits and behaviors not present in the earlier stages of ontogeny are the product of the animal’s intrinsic developmental program rather than elicited by, or acquired from, an environment of conspecifics already exhibiting the traits and behaviors in question.

Remember The ugly duckling? Same story: categorical identity for members of a species is stable in the face of varying environments of rearing, and it implies that the adopted animal will develop to display the adult features and behaviors of the species-category from which it is biologically descended.

Rips (1989) gave adults a story about an animal he called a ‘sorp’, and which was described as having feathers, flying, making nests, etc.¾ all of the characteristics of a typical bird. This sorp came into contact with some toxic waste and underwent some transformations: shed feathers, developed transparent, membranous wings, a brittle, iridescent, outer shell, etc.¾ all of the characteristics of a typical flying insect. The transformed sorp eventually met a normal female sorp and mated, which led to the laying of eggs which in turn produced normal sorps. He calls this the ‘accident’ manipulation.

In a 1 to 10 scale where 1 is ‘insect’ and 10 is ‘bird’, subjects’ mean ratings were as follows: in the categorization condition, they rated the animal a 6.5 (i.e. most thought the transformed sorp was more likely to be a bird than an insect); in the typicality condition, a 4 (i.e. most thought the transformed sorp was more typical of an insect than a bird); and in the similarity condition, a 3.5 (ditto).

This shows two things. (1) The independence of similarity and categorization; and (2) The importance of what I believe is the necessary and sufficient condition for categorization in a natural living kind: biological descent. Notice: the transformed sorp is nevertheless the offspring of normal sorps, and produces normal sorps in reproduction, therefore, despite appearances, it is still a bird and not an insect. What I would like to see now is a modification of Rips’ experiment such that the transformed bird, in addition to all its other changes, lays eggs asexually and these, when hatched, resemble the transformed parent. I predict that respondents would say the transformed sorp is now an insect.

Although I think biological descent is necessary and sufficient for membership in a biological kind, I do not think this is tantamount to saying that biological descent from a given biological kind equals the kind’s ‘essence’. It seems to me that an ‘essence’ can only be made up of predicates constituting one or more properties, else we are no longer speaking English.

By this real essence I mean that real constitution of anything, which is the foundation of all those properties that are combined in, and are constantly found to co-exist with, the nominal essence: that particular constitution which everything has within itself, without any relation to anything without it.—Locke (1964; quoted in Keil 1989)

Certainly, one might say that ‘ability to reproduce the kind’ is perceived as part of a biological kind’s essence (and I would hazard that it is), but this is obviously different from saying that the essence consists in being descended from members of the kind, for this is not really a property but a statement of history. Thus, I believe biological descent from a given biological kind is taken as the necessary and sufficient condition for the possession (through inheritance) of that kind’s essence, and that one of the predicates assumed of any biological kind essence will be ‘ability to reproduce the kind’.

In a normal world, I submit, one has a given, biological-kind essence if, and only if, one is biologically descended from members of that kind. However, it is possible that in a fanciful world of weird toxic wastes, acquiring the essence some other way is an idea that can be entertained. This is why doing the above-mentioned modification to Rips’ experiment would be so relevant for my hypothesis: if people call the transformed sorp ‘an insect’ when it generates insect offspring in reproduction, but not when it produces normal sorps, this suggests that ‘ability to reproduce the kind’ is part of the essence, and that reproduction transfers the essence.

The research reported earlier was conducted in the summer of 1997. A replication was attempted in 1998 with an entirely new sample. The categories were crossed (the biological father was a Mongol, the adopters were Kazakhs), the order of the questions was randomized, and the sex of the child was changed to female. Not only did the earlier results replicate, the proportion of primordialists went up considerably (Gil-White, forthcoming). This is consistent with my hypothesis that the original study had a task demand where the order of the questions suggested to subjects that the ‘right’ answer (or the answer I, the researcher, wanted to hear, at any rate) was one where the environment of enculturation conferred ethnic ascription. For those subjects whose first chronological question is question 2 or question 3, the effect disappears and they are more comfortable giving answers in line with their beliefs.

In addition to the above manipulations, a sample of ‘hard’ primordialists (i.e. of those who, even in question 3, believed that the child’s ascription would depend only on the ethnicity of the biological parents) were asked an additional question. First, the details of question 3 were reviewed, and then I asked the following: "This adopted child, will he become exactly like the Kazakhs, or will he be somewhat different?" (Remember, in this second study, it is the Kazakhs who are the adopters.)

If they answered that the child would be somewhat different (a common locution was "No, he can’t/will not become exactly like the Kazakhs"), then I followed up with the question, "How will the child be different?" Almost invariably, the response to this was that the child’s features would be like those of a Mongol, revealing that subjects had understood the question as it was intended, in terms of similarity to Kazakhs as a group, and not to the adoptive parents in particular. Then I followed up with the question: "And how will the child behave? Will the child behave exactly like the Kazakhs or will he behave somewhat differently?" A majority of respondents replied that the child would behave somewhat different, that he would not behave quite like the Kazakhs, but somewhat like a Mongol. Note to brown-bag people: the table is not ready, I will present it in the talk.

The same was done with a sample of non-‘hard’ primordialists (i.e. those who did not give a primordialist answer in question 3). Most of these responded that the child would be exactly like the Kazakhs, but a few responded essentialistically. Thus, one’s views of ascription do not rigidly determine one’s answer to the essentialist question, though they certainly bias it heavily.

What these results reveal is that, in addition to having a biological ETAM (Ethnic Transmission and Acquisition Model; Gil-White 1999), Torguuds think that this carries meaning beyond just a label that marks your biological descent. Apparently, being a ‘Mongol’, or being descended biologically from Mongols, implies that a person will resist enculturation into an outgroup to a certain degree, even when such enculturation is the only one the person knows, and even when she has had no contact with the group she is biologically descended from. This is consistent with believing that a ‘Mongol’, by virtue of being descended biologically from other Mongols, has a Mongol ‘essence’ which confers an innate potential that will play itself out developmentally ‘on its own steam’, as it were (as in Gelman & Wellman 1991). The intuition that the adopted child will be somewhat different remains despite the fact that my subjects accept the extreme premise, namely, that the child in question 3 will have learned Kazakh language, norms, habits, and customs, and that he will have not the slightest bit of exposure to Mongols and will in fact be ignorant of having been born to Mongol parents.

As my fieldwork progressed, and I experienced people’s beliefs outside the rigid context of the interview format, I came to believe that my questionnaires were good tools for revealing people’s conscious theories, but not their intuitive ‘theories’. I use the term ‘intuitive theories’ to substitute for ‘theory’ or ‘naïve theory’ as the term is used in cognitive psychology, where it refers to the organizing and constraining (but subjectively unnoticed) content that underlies human concepts and categories. By ‘conscious theory’ I mean an elaborated belief that organizes knowledge and that the individual is aware of having. The responses my questions elicited, I submit, result from an interview context that forces people to use their conscious theories in a very explicit way. Such conscious theories are presumably the joint result of their cultural upbringing and their personal experiences. In a previous paper (Gil-White 1999) I called such conscious theories regarding the transference of an ethnic status ‘Ethnic Transmission and Acquisition Models’ (ETAM). That my questionnaire investigates conscious but not intuitive theories is not cause for despair, for they are far from irrelevant. Conscious theories are important, and a theory that human cognition is designed so as to naively process ethnies as natural living kinds makes the non-trivial prediction that a majority of cultures in the world will turn out to have blood-based ETAMs. However, if the design of the human brain is what we are after, conscious theories can be misleading, and I shall illustrate with two anecdotes.

My friend Tsoloo (Tsoh-lóh) is a bright 21-year-old Torguud who served as my guide in one of my brief forays into Kazakh territory. Although he was never formally subjected to my questionnaire (except for the essentialism questions), he developed a keen interest in my research; he witnessed as I administered it to many Torguuds, Oryankhais, and Kazakhs; and he shared his views of their answers with me. Tsoloo is what I call a ‘circumstantialist’, and a vociferous one. To him, the right answer to question 3 is that the child is a member of the adopting ethnie. He also insisted repeatedly that the right answer to my question "Will this adopted kid be exactly like the [adopting ethnie], or somewhat different?" is that the kid will be exactly like the adopting ethnie. How could it be otherwise? It bothered him considerably that many people give essentialist answers. To him it was all quite simple: if you have the customs of an ethnie and live with them, you are a member of that ethnie, and if you are raised by that ethnie from birth you won’t be any different from them no matter who your parents are. He is even a ‘hard circumstantialist’, that is, a rational-choice theorist, who maintains that adults can choose to make the ethnic shift by living with, and acquiring the customs of, an outgroup ethnie.

But this is when we were talking about my questions. When discussing other topics, he often seemed like an essentialist. For example, he is very afraid of the Uryankhai because he believes they can all cast curses. The Kazakh can also cast curses, according to him, but cannot be touched by anybody’s curses because of their strong religion, Islam (this makes them theoretically more dangerous than the Uryankhai, but in fact they are supposedly less prone than the Uryankhai to cast curses because they are not bad people). Now, of course, one could argue that this is all compatible with his circumstantialism because he may believe that being able to cast curses like an Uryankhai or a Kazakh, and being more or less well-disposed towards others, is a matter of being a good Uryankhai or Kazakh performer in cultural terms, and of being reared amongst people with certain views. In fact, he listed the impossibility of putting curses on the Kazakh as an example of them being buruu nomtoi [‘with the wrong book’, where ‘book’ is a euphemism for ‘culture’]. But our these conversations somehow never felt circumstantialist to me.

I was wondering about these things one day as Tsoloo, his friend Tulgai (also a circumstantialist), and I crept on our horses up to the barren and rocky mountaintops in order to cross back into Torguud country. To investigate the matter, I initiated the following exchange,

"Tsoloo, I have a question: if I learn the Uryankhai customs can I then put curses on people like an Uryankhai?"

"No, you can’t."

"Why not?"

He paused. "You have to be Uryankhai."

"But you said that if I had the customs of an ethnie I would be a member of that ethnie. So why can’t I put curses like an Uryankhai if I learn all their customs?" I was explicitly reminding him of the model he had resorted to when considering my questionnaire, and therefore stacking all the odds against him persisting with an essentialist answer.

"You can’t," he said.

"That’s right, you can’t," joined Tulgai.

"Well why not?"

Tsoloo paused for a second and said, "You…have to be born to an Uryankhai." They were both agreed on this point.

"So unless one of my parents is Uryankhai, I can’t put curses on people?"

"True."

"That means that if I was adopted at an early age by the Uryankhai, I would be just like them but I would not be able to put curses…"

"True."

"Is it the same with the Kazakhs?"

"The same."

At this point Tulgai rode off because he wanted to kill a marmot, and Tsoloo and I were left alone to continue our conversation. I paused for a second in order to consider what a tantalizing contradiction Tsoloo had just provided, and how completely unaware he was of it. Had it been any other circumstantialist it would have been striking but not this striking. Tsoloo had seen me administer the same questionnaire at least a good dozen times, we had discussed it together and thought a lot about it. And only a few hours before I had administered the questionnaire twice. Yet, he was completely unaware that his current answer was at variance with the position he had been explicitly maintaining all along, and the fact that my last question closely mimicked the question I actually use to probe essentialism did not clue him to what was going on. Tsoloo was not thinking theoretically, he was thinking intuitively. I was very excited but instead showed myself mockingly upset.

"Tsoloo, you lied to me," I said with a wink and a smile.

"Why?"

"Because you have been telling me all along that if a kid is adopted from birth into another ethnie he will be just like the adopting ethnie. But that is not what you think. Suppose that the child born of Mongol parents and adopted by Kazakhs was a Torguud Mongol. Well, then he would be a lot like the Kazakhs but not exactly like them, because he would not be able to throw curses. You see?"

"Oh…!" He looked down at his horse and took his free hand to his mouth in a thoughtful expression I had seen many times. An embarrassed smile crept over his face and then he looked at me, nodding with a smile that he accepted defeat graciously.

"You actually think that kid would be a little bit different, but that is not what you told me when we talked about my experimental questions."

"Yes…yes, I see. Aha…" He looked up the road with a curious expression. He seemed to be pondering the intricacies of his own mind.

"Let me ask you this: if an Uryankhai child was adopted by the Torguud, just like in question three, would he be able to cast curses or not?"

"Well…[a significant pause]…that child doesn’t know that his real parents are Uryankhai…so…he would be able to, but he wouldn’t know it."

"I see. If he knew, he might cast curses, but since he doesn’t, he won’t."

"True."

I paused again, then explained a difficulty, "This is a problem for my research, Tsoloo. You gave an answer based on customs when I asked you the experimental question, and said the kid would be exactly like the adopting ethnie. But that’s not what you really think. What if other people who answered like you also think like you? You seem to believe the adopted child will look and talk exactly like the adopting ethnie, and yet…and yet, you believe that the kid…inside…is somehow still different…"

"Yes, yes. That’s right!" He seemed excited that I had chosen just the right words to explain his thoughts. I had of course carefully chosen those words on purpose.

"Maybe some of the people who answered like you also think this way, that the kid is somehow still different inside…"

"That is what they all think," he said very convinced.

"You think so?"

"I am sure."

Notice that in Tsoloo’s mind, in order to have that thing which confers the hidden potential of an A, and which is ‘inside’, you have to be a biological descendant of an A.

My data on Kazakhs is of much lower quality than that for the Mongols because I didn’t spend much time with them and at first I didn’t realize one had to remind them explicitly that the child in question 3 would acquire both Kazakh culture and religion. However, I think I can safely claim that Kazakhs are by-and-large circumstantialists (as measured by my questionnaire) so long as they understand that the adopted child will become a Muslim in addition to the other aspects of Kazakh culture. This is consistent with Bessac’s (1965) work with Kazakhs in Xinjiang. Since the Kazakhs I worked with immigrated only a generation or so ago from neighboring Xinjiang, in China, I am not surprised.

However, here too I found a discrepancy between the conscious models my questionnaire elicits, and people’s intuitions. When I asked Kazakhs alone, the overwhelming majority of them were circumstantialists. On one occasion, however, at a bayr (party), a large group of Kazakh old men took interest in my research and I asked the questions publicly, to the whole group. The old men answered in chorus and they turned out to be heavily circumstantialist, and not at all essentialist. However, there was one lone dissenter, a young man, who protested furiously and gave contrary, primordialist and essentialist answers each time. As the old men went quietly back to their old guy things, I was pondering how interesting it was that the Kazakhs should be ascriptive circumstantialists, and not the least bit essentialists. However, these musings were immediately interrupted by a large circle of young guys, who were sitting to my right, and who began protesting furiously to me that what the old guys were saying was not true. Prominent among them was the lone dissenter from the earlier choruses and shows of hands. Apparently he was nowhere near being alone, merely the only one who had the guts to override the embarrassment of appearing publicly as a deviant. Another guy, who was also quite mad, was also doing a lot of talking. The argument was familiar: the girl is a Mongol because the biological parents are Mongol. The girl’s culture didn’t matter in the least. I turned to them and said,

"So I for instance can never become a Kazakh? If I stayed here, and learned Kazakh, and Kazakh customs, married a Kazakh girl, and became a Muslim, I would still not be a Kazakh?"

One guy spoke for all of them when he replied,

"Even if you do everything like a Kazakh, and everybody says you are a Kazakh, you still aren’t a real Kazakh because your parents are not Kazakh. You are different inside." And he pointed to his chest. All others were nodding. "That girl in your question, you know? Everybody would say that she was a Kazakh, but she would not be a real Kazakh."

My main interlocutor and I talked about other things for a while, and then I brought us back.

"Can I ask you another question?"

"Okay."

"You say that I can never become a Kazakh even if I convert and do everything like a Kazakh, because my parents are not Kazakh, right?"

"Right."

"But is this also true for the girl in my last question? Is it true for a girl who is adopted at a very young age?"

"Yes. People will say that she is a Kazakh, but she is not a real Kazakh."

"Because her parents are not Kazakh?"

"Right."

"She is also different inside?" I pointed to my chest.

"That’s right."

This time I had not been the one to supply the word ‘inside’. And this time too, being an A ‘inside’ was a matter of being descended biologically from A.

When I was back at Xurmet’s (my 22-year-old Kazakh hosts’s) yurt, I had him alone in the yurt for a second, so I asked him,

"Isn’t it interesting that the young guys thought so differently from the old guys?"

"Yes," he said with a quizzical smile.

"Tell me something: when I asked you my three questions the other day, you answered like the old guys. The young guys told me it was true that the adopted girl would be called a Kazakh, but did not agree that she would be a real Kazakh because she was different inside. What do you think? Do you think the girl will be a real Kazakh?"

He thought for a second and then said,

"This is what I think: I think that the girl is Kazakh ündecten like I said before…"

"But is she jinxene [real] Kazakh ündecten?"

"Yes. I think she is jinxene Kazakh…"

"And you think she will be exactly like Kazakh people?"

"Yes. She will be exactly like Kazakhs…but…I also think, you know…that she will be different inside." He pointed to his chest just like the other man had done. "Inside she is a Mongol," he concluded.

This was truly astounding. Xurmet thought she was a real Kazax but a Mongol ‘inside’. A contradiction? Perhaps. Apparently Xurmet has a conscious theory of what it takes to be a real Kazax. But this is not in line with his intuitions, which tell him that the girl in my example would still somehow be different ‘inside’, despite all outward appearances and ascriptive practices. These experiences convince me that the formal questionnaires probably significantly underestimate the degree to which people are intuitively inclined to think about ethnies in essentialist terms.

Assume it is true that shared appearances are typical of members in a natural kind, but nevertheless not determinative of membership. Assume also that what is determinative of membership is the ‘essence’, whatever it is. Finally, assume that ‘essences’ are cognized as ‘hidden’: "more than meets the eye," and therefore causative of a great many hidden properties and characteristics that are common to all category members but which are not immediately apparent. It seems reasonable then, that:

(1) Knowledge that an item is a member of natural kind ‘A’ will lead to the automatic assumption that it has an ‘A essence’.

(2) Learning that this item has hidden property P should lead to the assumption that P is either caused by, or a part of, the essence, and thus to the generalization that P is true of other A’s.

(3) A’s that don’t look like the target item will be thought of as having property P anyway, and non-A’s bearing a strong similarity to the target will nevertheless be thought of as lacking P (unless the property is also specified for the natural kind they do belong in).

In a study involving 3-year-olds Gelman & Markman (1987) provide evidence for the above propositions, and for G&M’s explicit hypothesis that natural kinds will favor categorical rather than appearance-based inductive inferences. Their findings are very similar to what they earlier found for 4-year-olds (Gelman & Markman 1986), and so the 1987 study pushes the developmental threshold back by a year. They tested both animate/biological and inanimate natural kinds. I illustrate the spirit of the experiment by describing the word-and-picture condition.

Children were presented with a member of a named category (e.g. ‘a cat’) which constituted the ‘target’, and were told something about it¾ e.g. "See this cat? This cat can see in the dark" (the children were pretested for naïveté concerning the properties they were informed about to ensure that they were in fact learning something new). Then they were shown other pictures: (1) similar to, and in the same category as, the target (e.g. a black cat with a white stripe down its back, just like the target, but sitting in a different position); (2) dissimilar to, but in the same category as, the target (e.g. a white cat in a different position from target); (3) very similar to, but in a different category from, the target’s (e.g. a black skunk sitting in the same way as the cat); (4) dissimilar, and in a different category from, the target (e.g. dinosaur).

For each of these pictures, the experimenter placed the picture on the table directly below the target and said, "This one’s an X [where X was the label, e.g. cat, skunk, or dinosaur]. See this X? Do you think it can see in the dark, like this cat?" After the child answered, the experimenter picked up the card that she had just talked about and placed the next nontarget card on the table—Gelman & Markman (1987:1536)

Where category membership and appearance supported opposite inductive inferences, children preferred categorical as opposed to perception-based induction.

Children drew more inferences to pictures of the same category but different appearance (e.g. from a leaf-insect ["bug"] to a black beetle ["bug"] than to pictures of similar appearance but a different category (e.g. from a leaf insect to a leaf). (…) This was true for both animal and inanimate domains, so it appears to be a fairly general assumption that children have by age 3.—Gelman & Markman (1987:1537)

It is important to note that children were not forced to choose between category and appearances, and could have made inductive inferences for both. Yet, they still preferred to rely on the categories; this supports proposition (3) at the beginning of this section.

Why do natural kinds promote categorical inductive inferences of non-obvious properties? Gelman & Markman (1986:184-185) observe that natural kind categories have rich, correlated structures, where a great many properties that correlate strongly are not obvious on first inspection, and therefore

extend far beyond our original categorization. For example, giraffes share a particular diet, life expectancy, gestation period, DNA structure, and so forth—attributes that are impossible to know by casual inspection. (…) The highly correlated structure of natural kinds suggests that new features learned about one category member will often be projected onto other category members as well.

A Darwinian unpacking of that statement might go like this. Any animal that relies heavily on learning will benefit by reducing the costs of the learning process. If we can reliably learn about whole suites of objects merely by examining one of them, then––by golly––evolution would have failed us if it didn’t provide mechanisms for doing so. Of course this argument applies to artifact categories (which are also characterized by rich, correlated structures) as much as to natural kinds. What makes induction in natural kinds special is that inductions are easily made for non-obvious (i.e. ‘hidden’) properties, and this is because members of natural kinds in fact do share many non-obvious properties.

Contrast this with, say, artifacts. If I showed you a ceramic pitcher and I told you that it brakes easily (a hidden property), you would be wrong to infer this as true for all pitchers, since they can be made of wood, stone, coconut shells, metal, etc., and, in modern times, out of unbreakable plastic (cf. Gelman 1988). The things members of an artifact category typically share in common is perceptually obvious: their parts and their interconnections are constrained to be similar because they must fulfill the same function; the hidden properties (such as the properties of the materials they are made of) can vary widely so long as the artifact itself fulfills the same function. Thus, if I show you that a ceramic pitcher breaks easily, you will generalize that to ceramic anythings (because ceramic is a ‘substance’); but you will not think of this as a ‘pitcher property’. Evidence for this can be found in Gelman & O’Reilly (1988:881)

Recent evidence (Atran et. al. 1997) supports the hypothesis that, where biological natural kinds are concerned, there is a privileged living-kind category level (or ‘rank’) for the purposes of making inductive inferences. This is the ‘generic-species’ level, which corresponds to the folk-species level in folk taxonomies (although it sometimes corresponds to scientific genera due to the fact that in most locales genera are monospecific and, when they are not, the species are very difficult to tell apart; Atran et. al. 1997:20-21). The generic-species level does not correspond to the ‘basic-level’ category in other taxonomical trees. The basic-level is the most informative on perceptual grounds (e.g. ‘chair’ rather than ‘furniture’, and rather than ‘kitchen chair’; Rosch et. al. 1976), and therefore privileged psychologically in salience, ease of processing, inductive potential, frequency of utterance, ease of learning, and representability. Note that there is a huge gain in information in going from, say, ‘furniture’ to ‘chair’ (the functions, shapes, and uses of chairs are considerably more restricted, and morphologically there are relatively very small variations across chairs, as against the same for different kinds of furniture), and a very small information gain in going from ‘chair’ to ‘kitchen chair’. On perceptual input alone, this also happens at the ‘life-form’ level in biological taxonomies (e.g. bird, tree, fish; see Rosch et. al. 1976). However, Atran et. al. (1997) found that the rank which acts as ‘basic-level’ for inductive inferences is the generic species. This seems to be the case because Atran et. al. tested inductive inferences for ‘hidden’ or non-obvious properties (e.g. susceptibilities to disease, the possession of particular proteins). Thus, whereas a domain-general mechanism does indeed seem responsible for the rank-similarity of the ‘basic-level’ in all sorts of taxonomical domains, the biological domain is different because of the nature of biological information, where countless non-obvious properties cluster at the generic-species level but not the life-form level. This discrepancy between biological taxonomies and other taxonomical trees suggests that there is indeed a privileged biological domain of cognition.

The idea is that universal, possibly innate, principles lead people to believe that visible morpho-typical patterns of each readily identifiable generic species, as well as non-obvious aspects of biological functioning, are causally produced by an underlying essence. The nature of this essence is initially unknown but presumed.— Atran et. al. (1997:37)

Atran et. al. (1997), by comparing the performances of the inhabitants of two very widely divergent cultures, and finding in both an absolute privilege of the generic-species level for the purposes of inductive inferences, support the hypothesis that the living-kinds or ‘folkbiology’ module is a human universal.

As in generic-species, there is great inductive potential in ethnic groups, for much of what one learns about one member of an ethnie can be extrapolated to the others. This may be counter-intuitive to anthropologists and sociologists, post-Barth, who have decided that ethnic groups and ‘cultures’ are not coextensive, so I elaborate.

What many take to be Barth’s (1969) views on the relationship between ethnicity and culture have become widely accepted. He is often presented by others as debunking the idea that ethnic boundaries organize culture; following his critique, anthropologists have become very skeptical that ascriptive boundaries closely correspond to ‘culture’ boundaries.

Such interpretations of Barth are in fact dead wrong. What Barth debunked was the idea that ethnic boundaries organize the totality of culture in a holistic way¾ i.e. that ethnic boundaries are closely coextensive with discontinuities in "trait inventories" arbitrarily compiled. But practically in the same breath he insisted that what ethnic boundaries do enclose is "ethnic organization" (Barth 1969:12), and that the cultural content relevant to ethnic groups is (1) the diacritical features that signal membership, and (2) "basic value orientations: standards of morality and excellence by which performance is judged" (ibid. p.14). So long as we agree that standards of performance are ‘culture’, Barth is not really subtracting all culture from ethnicity, as if ethnic ascriptions were truly arbitrary, but rather insisting on distinguishing the kind of culture that is most relevant to ethnicity: standards of performance, and the diacritica that signal membership inside a performance-norm boundary. As Barth himself has recently observed (Barth 1994:12), this part of his argument is often ignored.

. . .the issue of cultural content versus boundary, as it was formulated [by Barth (1969)], unintentionally served to mislead. Yes, it is a question of analyzing boundary processes, not of enumerating the sum of content, as in an old-fashioned trait list. But locating the bases of such boundary processes is not a question of pacing the limits of a group and observing its markers and the shedding of members. . .central and culturally valued institutions and activities in an ethnic group may be deeply involved in its boundary maintenance by setting internal processes of convergence into motion; and we need to pay special attention to the factors governing "individuals’ commitments to the kind of personhood implied by specific ethnic identities" Haaland 1991:158).—Barth (1994:17-18)

Lest we lose the thread, notice for now that concepts of personhood, as such phenomena get realized through the observation of performative norms with interactional consequences, are hardly things that are readily apparent on the casual inspection of an ethnic actor. So if Barth is right, this would mean that fully-socialized members of an ethnic outgroup (and even of the ingroup!) will have lots of strongly correlated ‘hidden’ properties. I shall come back to this question shortly.

It would indeed be remarkable if Barth were wrong. Explaining ethnicity in the absence of any real ‘substance’ that such mutual ascriptions organize would present an awful challenge. Surely, ethnic ascriptions occur for a reason. Barth’s answer (in fact, a commonly held view; e.g. Smith 1986:41, Deutsch 1953) is that they are there to demarcate ‘ways of being’.

According to him, we must define ethnic groups in terms of the actors’ own ascriptions because these imply commitments to certain interactional norms: "if they say they are A. . .they are willing to be treated and let their own behavior be interpreted and judged as A’s and not as B’s" Barth (1969:15). Coethnics therefore understand each other to be "playing the same game." Barth explains the maintenance of ethnic boundaries in terms of the punishments that accrue to those who fail to fulfill the normative expectations held by their coethnics. Such punishments enforce ingroup conformity and maintain normative differences between ethnic groups, which differences in turn¾ because actors assume they are signaled in ethnic diacritica and ascriptions¾ entail constraints on inter-ethnic interaction (Barth 1969:17-18).

Barth would probably hate having this pointed out but his argument for the maintenance of the boundary sounds very much like the feedback mechanisms of earlier functionalists (it is practically a Durkheimian social-solidarity argument). The only thing missing is the obligatory claim that it is adaptively functional for the group to enforce observance of its norms at a high level, which, although missing, is a quite reasonable proposition (which cannot be explored here). However, Barth says nothing about coordination costs, even though these suffice for boundary maintenance, and even though they supply a way out of the usual circularity of functional explanations by providing a very plausible, mechanistic explanation for the original emergence of ethnic boundaries.

Humans should be conformists because in this way they can maximize the number of potential interactants in their local community with whom to engage in mutually beneficial cooperative and coordinated endeavors (Robert Boyd, personal communication). Much evidence from psychology (e.g. the entire literature on ‘pluralistic ignorance’; Miller & McFarland 1991) bolsters this contention. The classical example is: when in Britain, drive on the left side of the road, because that’s what all the Brits do. This would be sound advice, of course, even if there were no ‘bobbies’ to inflict polite scoldings, and the wisdom it contains is routinely repeated as common sense in the universally known proverb "when in Rome, do as the Romans." Notice also that the proverb and example contain wisdom about what to do when one is a migrant. In an ancestral environment composed of wise conformists, migrants from one community into another would absorb the norms of the host community and this would prevent the blending-of-norms effect that migrants would otherwise have on cultural variation, ensuring in this way that at least some important interactional norms would regularly clump discontinuously across the human landscape. Thus, it seems plausible that posterior to the emergence of the conformist adaptation, the ancestral environment became populated with more or less well-defined, discontinuous norm-clumps. There are two reasons why thinking of such incipient ethnic groups as ‘natural biological kinds’ is advantageous in these circumstances:

(1) Interactions with those socialized into different interactional patterns and expectations will not as often be felicitous, and the costs incurred––in energy and time spent and wasted––will concede the evolutionary advantage to any mutants who manage somehow to discriminate (I am indebted to Robert Boyd, personal communication, for this argument). A naïve ‘living kinds’ theory of the social world, coupled with a self-similarity preference bias, would represent ego’s norm-group in her mind as one ‘kind’, and other norm-groups as different ‘kinds’, thus making the interactional choices favored by the self-similarity bias rather clear.

(2) To the extent that outgroup members do not become irrelevant, adaptation to a norm-clumped world presents the dual challenge of avoidance and prediction. It is certainly best to be able to predict as much as possible about outgroup members without having to obtain this knowledge through too much costly interaction with them. Since conformist behavior will ensure that members inside the same ‘conformist sampling horizon’ (i.e. inside the same norm-boundary) will show strong intercorrelations for many non-obvious behavioral patterns and expectations, inductive inference (from the one to the many) will usually be a good way of cheaply obtaining such valuable information.

When individuals, through trial and error, learned that marrying outside one’s norm-boundary carried enormous costs (because of different rearing, adoption, conjugal, and other practices; see Nave 1997), such incipient ethnic groups became endogamous. And because culture is acquired at a young age and tends to be developmentally stable, expert practitioners of the groups norms would also be those born into the group. When such groups began labeling themselves, it is only natural that category-based endogamy and descent-based membership primed the living-kinds module, for these two characteristics are highly diagnostic of biological species. Because the priming of this module would have had salutary adaptive consequences (prevent intermarriage, prevent costly interaction, promote efficient and informative inductive inferences of non-obvious properties) there was no selection to discourage such priming.

A good case can be made that the social psychology program which investigates social stereotypes, and which got its start with Gordon Allport’s The nature of prejudice (Allport 1954), has been investigating some of the pernicious effects of ‘natural kind’ discriminations and inductive inferences in the social arena. An often reported result in this literature is that what is learned about one individual of an ethnic outgroup member is then assumed to be true of the whole ethnie (bunch of references).