Mallon essay on social construction

I just came across an article by Ron Mallon "A Field Guide to Social Construction," by Ron Mallon (in Blackwell's Philosophy Compass).  I believe there is direct and free access here:
http://www.blackwell-synergy.com/doi/full/10.1111/j.1747-9991.2006.00051.x

Toward the end of the paper, Ron makes the following point:
"But while nativist accounts of folk theories of race may compete with constructionist explanations, they might also be combined into accounts emphasizing a variety of factors -- both social and psychological -- leading to folk racial theorizing.  And in any case, there is no reason that human kind constructionism about race cannot be paired with alternate accounts -- including nativist accounts -- of racial theories.  Such accounts might hold that psychological predispositions contribute to the formation of racial social roles that have played an important role in racial oppression." (103)

I agree with Ron that in principle nativist accounts of race might be combined with constructionist accounts, but I am very dubious of empirical work (I'm most familiar with Hirshfeld's (1996), which I think is appallingly bad) suggesting "innate psychological propensities to categorize people" along lines that map the contours of social race.  Of course children recognize the difference between brown skin and pale skin as soon as they get color concepts, just as they recognize the difference between a red apple and a green apple.  But children have to be taught that sometimes green apples turn red, and they must be taught that brown skinned babies won't turn pale as they grow up, and that a pregnant pale mommy with a pale daddy will probably not produce a brown baby. (I know this from personal experience with my own family.) There are many complex issues here, but I don't think there is any reason, in fact, to think that a combined nativist + constructivist account of race is plausible.

What do others think?

--SH

Comments

[Josh Glasgow]

Hi Sally,

Fantastic blog, here! Just in case some haven't yet seen them, folks also might want to check out a couple of great papers that Edouard Machery and Luc Faucher put out in 2005, which get into the details of, and explore the possibilities of reconciliation for, the nativism/constructivism stuff. Machery has links to them on his Web page: http://www.pitt.edu/~machery/ .

Anyway, I agree with the outlines of your concerns about nativism: the presence of racial thinking even at an early age is not decisive evidence that race-thinking is to any degree native. And I agree with you that Hirschfeld (whose work did at least break new ground on these kinds of topics!) at times overstates his view. In addition to your point, he also says that ordinary race-thinking is always committed to the kind of essentialism that Paul Taylor would call 'classical racialism,' when, I believe, even his own data don't support such a strong conclusion. (A shameless plug: I've been doing a couple of studies with a psychologist that show, among other things, that ordinary race-thinking is more nuanced than that. If anyone's interested -- these manuscripts are more about the content of race-thinking than about how best to explain it -- email me at joshua [dot] glasgow [at symbol] vuw[dot]ac[dot]nz.)

I'm a bit curious: are you skeptical about the nativism+constructivism combination as a combination, or is it more a worry about the nativism side on its own?

And congratulations, again, on the excellent blog.

[Sally Haslanger]

Hi Josh, Thanks for your support for the blog and the references to the Machery-Faucher essays. I look forward to reading them (and more of your work). My worries, as you discern, are really about the nativism component of the nativism-constructionism combination. I grant that our constructions are responses to features of the world, but it would take a lot to convince me tha they are innate responses, or that there is no loopiness of the sort that Hacking and Mallon discuss.

I think people would be interested to hear more about your empirical studies, if you have a minute to share some of the data. --SH

[Ron Mallon]

In the Compass article, was drawing attention to the possibility of nativism + constructionism combinations in part because among psychologists, there is a very live debate about the extent to which human classifications have nativist explanations, and I think these kinds of theory should not be ignored by social theorists concerned with any sort of human category.
Any plausible explanatory theory of contemporary social categories will need to combine elements of both our psychological proclivities and our acquired culture (e.g. Machery and Faucher attempt to do this). For example, a classic social psychology approach that tries to understand racial or other classification preferences as a manifestation of in-group preference might hold that there is some relatively innate proclivity for in-group/out-group thinking that is employed in particular contexts to generate racial bias. What makes recent work different is that many recent theorists, including Hirschfeld, Machery and Faucher, think that the cognitive mechanisms are more specific in their function (though, in the case of race, none of them think that the cognitive mechanisms involved in racial cognition are for racial cognition because they don’t think there would have been selection for such mechanisms). So there’s a debate going on about what these mechanisms are for (for example, are they for picking out social kinds? ethnic groups? coalitions?), and there is a closely related debate about how they do or don’t explain various features of racial classification.
Sally writes: “I grant that our constructions are responses to features of the world, but it would take a lot to convince me that they are innate responses, or that there is no loopiness of the sort that Hacking and Mallon discuss.” Because I think any account of the cognitive mechanisms that give rise to racial classification will have to allow that culture plays a role, such an account will presumably be open to the possibility of “cognitive loopiness.” One way to put a common nativist idea in this area is that what’s innate is a proclivity to construe certain phenomena in certain ways (e.g. to construe certain kinds of difference as signaling the appropriateness of essentialistic reasoning), and this psychological “mode of construal” gets culturally elaborated in lots of ways.
This sort of explanation leaves unanswered lots of questions of interest to social and political theorists (noticeably absent is any account of racism, for example), but it does seem to me to be a part of the puzzle. (Dan Kelly, Edouard Machery and I have a work in progress that discusses some of these empirical questions, and tries to relate them to contemporary debates over race in social philosophy that can be found here: http://web.ics.purdue.edu/~drkelly/RaceNormativePsychKelly12-10-07.pdf .)
A quick quibble: Josh writes: “the presence of racial thinking even at an early age is not decisive evidence that race-thinking is to any degree native.” Lots depends on how strongly you read “decisive,” but Sally didn’t say anything so strong as that. If racial thinking in its adult form were really present from a very early age (which it is not), it would be powerful (if perhaps nondecisive) evidence for at least many components of racial cognition being innate. Conversely, the absence of such cognition at an early age (for example, if Sally is right that children must be taught that skin color is heritable) offers evidence against this view because the opportunities for acquisition of these beliefs from other sources expands considerably.
The explanation of racial classification and other kinds of classification is surely very complex. My own point in the Compass piece is that nativist and constructionists accounts of classification can agree on constructionist accounts of the categories thus produced. Because of the evidence from cognitive psychology (in which Hirschfeld played an important role) and because (as I note also in the compass piece), saying something is innate doesn't mean that it cannot be changed, this might actually be a desirable position to hold.

[Ron Mallon]

Oops: I meant to add: excellent blog!

[jj]

I'm a bit concerned about what one can draw from presence and/or absense in young children. Most of us don't get very far with numbers without a lot of schooling at a fairly late age, but that does not count at all against saying there's an innate number sense. Dehaene's work in this area provides a really enlightening model, I think. (Behavioral and Brain Sciences had a precis for his book, called, I think, The Number Sense that is very useful). Social kinds and facts are of course different, but the same story in a very general sense could be told. E.g., racial thinking, on such a story, would be the result of a complex coordination of capacities some of which are language dependant, and so won't be working well or at all in the early years. In fact, though I don't want to take sides on the specifics, one could start to appeal to "lower" animal's classifications as showing facts about what some of the components look like, etc.
Is the antipathy between cats and dogs a precursor of racism? Well, it is too serious a topic to be perhaps silly about, and we've had to do a lot of very thorough analysis before anything like that would look plausible, but still.... .

[Edouard Machery]

I think the first thing to acknowledge is that we know little about the ontogeny of the disposition to classify racially (and of racist value judgments).

But we know a little bit more than is mentioned by Sally Haslanger above:
- We know that there is a similarity between children's beliefs about race membership and their folk biological beliefs. (The extent of this similarity is however not clear.)
- Evidence suggests that 4-5 year-old children's beliefs about races (as elicited in experimental tasks) do *not* reflect their parents beliefs. This stands in contrast with older children's beliefs.
- Evidence from Hirschfeld suggests that the disposition to classify racially is not caused children's visual experience, in direct contrast to Sally's comment about color in her post.

Furthermore, we also know that there are striking cross-cultural similarities in the way races have been conceptualized. Work on racial classification in China and in Ancient Greece shows that racial classifications are not conceptualized in arbitrary ways.

I think that these and other phenomena should be taken into consideration when one wants to theorize about the nature of the cognitive systems underlying racial classification. (I should note that in contrast to Ron, I am reluctant to use the contrast between innate and learned; I have argued elsewhere that innateness is a confused notion.)

This is indeed a fantastic blog. Congratulations.

Edouard

[Sally Haslanger]

I'm finding the discussion here very helpful. A couple of quick points, though. First, in response to Edouard, I didn't mean to suggest that children classify racially based on color (or color alone), but that they notice and can tell the difference between different colored skin. This may seem too obvious to point out, but I wanted to acknowledge that there may be some basic innate capacities that are involved in racial classification such as the ability to distinguish colors and shapes. I don't think I really understand by the claim that the disposition to classify racially "is not caused [by?] children's visual experience." Do you mean that the disposition is not triggered by visual experience? What is it triggered by?

Second, Ron is right that I didn't want to claim that observations about what small children think is decisive about what's innate. But it should be noted that some of the work out there (such as Hirschfeld's) arguing for innateness does depend on responses by children to very very simplistic images. I've also found that research on this issue tends to realy on responses by children and adults in same-race families in communities where virtually all families are also same-race. This strikes me as methodologically problematic on several counts. I haven't been keeping up on the research like I should, though.

[Ron Mallon]

On Sally's quick points:
I think what Edouard has in mind (and he can say if this is wrong) is that the disposition to classify racially is not caused (developmentally) by merely observing visual difference (as it would be if it were simply generalizations based on skin color).
There are a few ways to see this: first, children seem to be engaged in group-based classifications and preferences before they learn to apply those on the basis of the physical differences we associate with race.
Here’s another way: while in the U.S. we think about race in terms of skin color, and our quite vivid array of different appearances, group-based classification and preference occurs historically among groups of people who look quite a bit more alike than our own racial groups do. One way to describe what is going on in these cases is that people are engaged in racial classification of persons, where the race concepts really don’t have anything much to do with appearance at all.

In truth, Hirschfeld’s simple drawings always draw chuckles when I show them to people. In what I think are some of the experiments you have in mind, he was trying to show that the physical markers of race in the U.S. (e.g. skin color and hair type) are used as cues to reason about growth and descent, and he contrasts these bases of inference with some others (body type, occupational uniforms).
A couple of concessive things to say about these experiments: (1) when dealing with children, the cues one uses have to be very simple, and the questions have to be very simple. And (2) how representational subject groups are in drawing “universal” inferences is a problem throughout psychology and philosophy. (Edouard and I have had occasion to note philosophers' penchant for generalizing from unrepresentative subject populations!). This persists for the obvious reasons that it is harder to get a broad range of data (though I think some cross-cultural data is now emerging about human group classification and inference).

In any case, attempting to understand features of cognition involves looking at data from a number of sources: data like Hirschfeld’s, along with lots of other data from other areas as well. To choose just one example, I think it’s quite striking that folk racial thinking resembles folk biological thinking in the way that kind membership is passed by descent. (Indeed, in both the species case and the racial case, having parents that belong to the kind seems to be treated as a necessary and sufficient condition for belonging to the kind). It seems to me that this is no accident. And because of this, it seems to me that the growing literature on folk biological essentialism (by people like Scott Atran, Frank Keil, and Susan Gelman) has something to teach us about racial thinking too, and some of this work has more cross-cultural confirmation.

Still, I think study in this area is still young, and there’s lots more information we’d like to have that we do not. It’s one thing to place one’s empirical bets, and another to think they have been vindicated beyond reasonable debate. I agree with jj that there’s room for resisting the kind of story on which racial thinking is innate (or as Edouard would perhaps prefer, canalized). There are no doubt more sophisticated empiricist models of racial cognition to be constructed and addressed, and there is, for example, a considerable body of evidence that tries to document cultural differences of racial classification that must be addressed to make this case. From the point of view of those interested in racial theory, though, I think the point is that this is a live rather than settled debate.

[jj]

Just a follow up to Ron's point about simplistic drawings. I don't know about research on the visual neuro-mechanisms involved in processing racial characteristics, but it is an very important general characterisitic of the human perception of faces that we can react to extremely simple cues. As a general point, human social interactions are very quick and don't take a lot of heavy cogitating. That's one reason why heavy cogitating can't substitute well, and so why even very bright ASD people can be at a social disadvantage.

In addition, I understand that simple facial features are processed by the more primitive of the visual pathways going to the occipital cortex and this also facilitates a very rapid processing that can in fact side-step any conceptualization and end up in the emotional centers of the brain. That's important because it means that there are valenced reactions to very simple and partial visual input from faces.

In sum: simple drawings may contain all the information that is available in early processing. It may well be later processing, which is partially top down, that gives us the fuller perception. It's pure conjecture on my part to say which is involved in perceiving racial characteristics, but I'd think it is an ok research hypothesis that there is some very early processing that is involved in the valenced reactions, If THERE IS AN INNATE COMPONENT that links with visual percepton.

Finally, if I remember Hirschfeld correctly, it isn't just any color difference in humans that sets off the race reaction. At least at some point he thought that it was to darker skin. That seems important, if it is true. It suggests that the sorting is not merely color sorting.

[jj]

AND PS: Thanks for earlier, generous comments, Sally, about the blog, Feminist Philosophers. Yours is certainly a wonderful blog.

[Edouard Machery]

Sally and Ron,

Ron explained the point I tried to make very well.

To answer Sally's question "What is the disposition to classify racially triggered by?" directly, Hirschfeld's evidence (tentatively) suggests that linguistic cues (the use of group names, maybe the use of generic sentences) are more important that the classification of people on the basis of visual cues in the ontogeny of racial classification.

Hirschfeld told children a story involving various characters (an old woman, a dog, an African-American man, etc.) and then asked them to summarize the results (the verbal condition). Other children were shown cartoons describing the same story (with the same characters) and they were then also asked to summarize the results (the visual condition). If my memory is correct, Hirschfeld measured how often children referred to the characters by means of racial, gender, (and maybe other) classes--for instance, he measured how often children described the African-American character as being African-American or Black.

If children acquire their concepts of race visually, they should be more likely to use racial predicates in the visual condition than in the verbal condition. But this is not what Hirschfeld found. Children were unlikely to refer to the characters by means of their racial membership in the visual condition, while they were likely to do so in the verbal condition.

There is of course much to be said about this particular experiment and, more generally, about the developmental evidence on racial classification. I share to some extent Sally's and Ron's concerns about Hirschfeld's and others' body of evidence. And I find it unfortunate that this work has not been pursued in a systematic manner.

Edouard

[Josh Glasgow]

Apologies for being away for so long...it looks like I missed some fantastic discussion here, but just to follow up on a couple of things:

Ron, I wonder if you could say more about this: "If racial thinking in its adult form were really present from a very early age (which it is not), it would be powerful (if perhaps nondecisive) evidence for at least many components of racial cognition being innate." First, right, let's read "decisive", well, decisively. But, also, I wonder what "powerful" means in this context. I mean, even if representations of race show up early, that doesn't by itself provide evidence that children aren't taught to think in racial ways at an early age, right? I guess part of my hesitation here is not just about what children can be taught when, but also about what the earliest is that we could be confident that a child's seemingly racial representation is an actually racial representation (or, put somewhat differently, when we can say with confidence that a child has mastery of racial concepts).

On a not unrelated note, Edouard writes, "we also know that there are striking cross-cultural similarities in the way races have been conceptualized. Work on racial classification in China and in Ancient Greece shows that racial classifications are not conceptualized in arbitrary ways." I'm not sure that cross-cultural representations are going to tell us too much about the nativism question (a question which Edouard was clear he wanted to avoid, so this is more relevant to those who are interested in that question). I always kind of liked a point made by Rita Astuti in response to Gil-White ("Are ethnic groups biological species to the human brain?", _Current Anthropology_ 41, p. 537), when Astuti noted that "Hirschfeld bases his empirical work on children for the simple reason that no amount of cross-cultural evidence of adult reasoning would support his claim that humans are endowed with an innate predisposition to essentialize human kinds." It's quite likely that I'm missing something crucial here, but at least at first glance that seems like a plausible response to me.

Also, Sally asked about the manuscripts I mentioned. One of them investigates the explicitly held beliefs participants report having about race (e.g., whether they think of it as biological or social, or both) and how those conceptions do or do not correspond with a couple of different kinds of racist representations. The other investigates the same kind of question, but instead of asking 'transparent' questions about explicit belief, we presented the participants with thought experiments and asked the participants to tell us the race of the people in the cases we gave them. Some of the cases were ones that are probably familiar to readers of this blog (e.g., a variation on Susie Phipps, and some cases inspired by the discussion at the end of Chapter 3 of Mills' _Blackness Visible_). Some of the results were more interesting than others. For example, above Ron wrote, in agreement with much of the literature, that "in both the species case and the racial case, having parents that belong to the kind seems to be treated as a necessary and sufficient condition for belonging to the kind". We actually found that in some extraordinary cases, many participants were willing to allow that one could change the race with which one is born and that one could have a different race than one's ancestors (including parents). Anyway, I'll avoid going into all the details here, but, again, if anyone's interested, please email me.

[Sally Haslanger]

So much material to process! I was confused earlier in the discussion of visual cues about whether we were talking about the etiology of the disposition or the triggering of the disposition. It does seem plausible that humans are disposed to group things, including other humans, and perhaps attribute essences to groups that are described in certain ways (though this brings in an important social element). But even so, there is still a long way to go to get races, so it is interesting to see what else happens to get races from this. That we are also highly sensitive to facial features is also clearly relevant.

I'm still vary wary. A friend of mine, Jackie Stevens (a political theorist who works on race: http://www.jacquelinestevens.org/ ) once made this point about Hirschfeld: suppose we showed simple black and white drawings of two adults with crowns on their heads and asked children which of the possible pictures of children were likely representing their kids; suppose we offered the subjects pictures of a kid with a crown and a kid without; and suppose that the children chose the kid with a crown. Would we conclude that the concept of monarchy is innate? Of course not. (For those unfamiliar with Hirschfield's work, some of his experiments are like this.) So why are we even tempted to draw this conclusion in the case of race? Does that say something about "us"?

Related to this, there seems to be a lot of controversy over what the concept of race even is, and on some views it would seem more plausible that we could get to it via the mechanisms indicated, but on other views it's hard to imagine how it would work (like 'monarchy')! Given the way the concept of 'race' (and the concept of 'innate') functions socially in the US, I'd be VERY WORRIED about the impact of research saying that racial classification is innate. This worry is relevant to deciding what the concept of race is, because I believe we should consider not just what people think about race, but the social function of the term.

It would help me to have a better sense of what concept of race is at issue, but I should read the recent papers that have been linked to the earlier posts to find out....(it's just hard to wait for that).

[jj]

SH, I think the crown construction story is really complicated. There are problems using thought experiments to rebut interpretations of empirical research. In this case, the fact that crowns are artefacts might be influencing our intuitions a lot.

I'm also struck by the fact that there seems to be a huge and very wide-spread preference for height; I'm forgetting all the details here, particularly whether it is just for men, or just for men-as-leaders. But, ironically, the kids might really like crowns and maybe they would be showing some appreciation of a connection here between height and power. I suppose there may even be a connection between wearing crowns and wanting to look taller. So this is all apriori, but I think philosophers' reactions to thought experiments might be less telling sometimes than they think, or indeed from another perspective, more telling.

(I say this as someone who spent about 30 years at various citidels of analytic philosophy and still has a tendency to think that science can't suprise me, thanks to Wittgensteinian tutors.)

[jj]

PS: I think maybe I shouldn't have left the point of the height story implicit. here's the explicitly put point: it might be that the better reaction would be to investigate whether the preference reveals anything. It's probably not a connection between artifactual clothing and government structure, which is possibly an irrelevant comparison if the question is whether children take race to be a natural kind. But it might be something else equally important that a philosopher might miss.

[Ron Mallon]

Josh- I suspect we do not really disagree with one another about the space of the debate. You wrote "the presence of racial thinking even at an early age is not decisive evidence that race-thinking is to any degree native." I wondered whether you were denying the logic of using early emergence as evidence for innateness, so I tried to provide a counterfactual case: one where racial thinking in its adult form was present "very early" (to be more clear now, let's say at birth). This is not the case, but were it so, it would provide powerful evidence of innateness. I was simply trying to insist that such early emergence could be decisive evidence that race-thinking is to *some* degree innate. (I thought your claim about (“to any degree”) was too strong. Why would this be so? Precisely because it means less time to be taught, to be corrected, to learn directly from experience, and so forth. It also means less time to develop "conceptual scaffolding." For example, if you think acquiring a complex concept requires requiring simpler others first, you should expect that the complex concept will come along only after the simple ones. Of course, these arguments are not demonstrative (and hence my hedging on "decisive"), but they are inferences to the best explanation that seem to me powerful. (Perhaps you disagree with that claim.) On the basis of your subsequent post, however, I now (think I) see that you were making the point that one can argue that early emergence is the result of teaching/learning. So, let me concede, in fact it's true that early emergence could, in principle, be the result of early learning/teaching. And as I said, I think the actual empirical questions here are far from settled.

On the issue of cross-cultural research: I think the right methodology to address these questions will be cross-disciplinary, looking at evidence from history, anthropology, and developmental work on children. Cross-cultural and historical evidence does not rule out the possibility that a trait just happens to be acquired universally. Developmental evidence can never fully rule out the possibility that the children we’re studying are a product of our peculiar cultural practices. And, indeed, cross-cultural research is standardly appealed to in the attempt to show the peculiarity and locality of our own U.S. racial practices, so it does seem relevant to any discussion about what components of racial classification might or might not be culturally produced.

Also, though Josh didn't explicitly criticize me for this, I shouldn't have written that in the racial case, the folk theory holds that having parents that belong to the kind is a necessary condition for belonging to the kind. There’s a long (and I think persuasive) literature to the contrary.

Both Josh and Sally raise a big and important question: what exactly counts as racial thinking? Perhaps whatever these studies show, they aren’t really about genuinely racial thinking. There are of course a number of questions here: semantic and pragmatic (a point Sally’s post as well as her work makes clear). I think a useful thing to do is to try to tell a descriptive story using an articulated vocabulary. George Fredrickson, for example, distinguishes early proto-racism (or racialism – I’m not sure I’m remembering correctly the terminology he used) biological racism in a way that (I think) avoids begging any of these questions and contributes to our understanding. Something like that distinction should be useful here, if it turns out Hirschfeld and others are on to something. E.g. Perhaps the thing to say would be something like “group-thinking” is innate, or that certain inference patterns are (e.g. the inference of membership in the kind from having two parents that are members of the kind). Another possibility for reconstruing the data is that Edouard and others like Paul Griffiths are right and we should forgo the use of “innate” all together, if it gives rise to widespread mistaken understandings of what is being argued for/against.)

[Josh Glasgow]

Thanks, Ron, that's very helpful!

[Jackie Stevens]

Hi Sally and friends, great blog!!! As the friend with the crown comment on Hirschfield's work, I thought I'd reply to the criticism of thought experiments with a thought that occurred to me while reading the thread up to this point.

This thought is that there is no mention of history in these discussions. Cognitive and behavioral psychology experiments are supposed to be designed to represent how humans generally behave, not to provide substitute Second Life versions of us. If we can find examples from society or history these are actually better evidence of who we are than the artifice of these experiments, especially because the experiments tend to find probabilities and not categorical results.

In the event, there is a large body of work suggesting that race was not a word or a concept until the fifteenth century. (With apologies for my poor memory, citations for this are in Ch. 5 of my *Reproducing the State* (Princeton, 1999).) Philosophers may recall Aristotle's *Politics*, where distinctions of natural slaves refer to their souls and not appearances. Indeed examples of what we might consider ethnic or national differences abound in ancient texts, but there are no references we might construe as racial.

Another reason I think the work using controlled experiments for trying to understand racial classifications is doomed comes from the critiques of probability studies by Karl Popper and Stephen Jay Gould.

If you take their critiques seriously -- Popper's of the problem with inferring knowledge from patterns and Gould's critiques of general claims based on the mean -- then the controlled studies are not producing knowledge about innate tendencies but merely data about the responses to the experimental conditions. None of them find differences among the subjects even close to those we associate with tendencies that are innate according to Popperian criteria.

That is, Popper (the kind of philosopher of science whose epistemology seems fair to apply to psychological studies) says that to count as knowledge one must generate a falsifiable hypothesis. If the hypothesis is not verified by the research, then a new one needs to be developed. It would count as knowledge, therefore, if I hypothesize the need to absorb oxygen is innate to humans. This claim could be falsified by finding one exception. As far as I know, the evidence suggests no human who does not absorb oxygen has ever lived. If such a person were found, then the hypothesis would be falsified and the claim that oxygen is an innate need would no longer be tenable. However, a variable that produces a hypothesized outcome 50% of the time is considered to be a robust result by most empirical researchers, and Popper finds this troubling.

If one considers how far the behavioral and clinical studies fall from Popper's criteria for knowledge it may make thought experiments based on historical evidence somewhat more attractive, i.e., the historical association of crowns with monarchies and a well-documented belief that the prerogatives of sovereignty have been episodically associated with hereditary privileges.

Thanks again, Sally, and everyone else for giving me a lot to think about. I'm working on a book that's really close to this topic and look forward to reading the works cited by those of you who've posted such interesting comments.

[Sally Haslanger]

Hi Jackie...welcome! I don't see you enough these days.
I'm even worried about the kind of inference pattern that Ron mentions: "the inference of membership in the kind from having two parents that are members of the kind". I know this is just anecdata, but let me give you an example that gives me pause:

When we lived in Ann Arbor, our kids were in a daycare in which the majority of families were interracial -- either by adoption or marriage. One of the few homogeneous families consisted of two white parents and two biological sons. The mother became pregnant with their third child, and one of the boys asked her please to have a brown baby. She said that it wouldn't be possible for her to have a brown baby, but he wouldn't take 'no' for an answer. For several days he wouldn't let go of the argument. Finally, the mother asked him why he thought that she could have a brown baby, and he replied, "Sally and Steve have a brown baby, and they look like you." (For those of you reading this who don't know, Steve and I are white and have two adopted African-American children.)

Ok, so there are problems with the example. In particular, I don't think the child had the concept of race, but was really just talking about physical appearances. (But this raises the earlier question: what concept of race are we talking about? On my view, the concept of 'race' is a lot more like 'monarchy' than it is like 'tall'.) The child also probably had very little understanding of reproduction. But it should also be noted that children have a very weak grasp of the distinction between biological parents and custodial parents until they learn a lot about how babies are made. The concept of a biological father is especially confusing.

Relevant to this, it should be noted that there are whole cultures which, according to anthropologists, don't have a biological account of reproduction. According to a feminist anthropologist friend of mine, "Some North Australians and Melanesians (Papua New Guinea and surrounding islands) are famous for not "recognizing" (or rather, not caring about?) biological paternity. They recognize that virgin women do not get pregnant, but on their view women are impregnated by matrilineal ancestral sprits; intercourse both "opens" a woman's body so spirit-fetuses can enter and semen nourishes the growing fetus, but does not contribute heritable substance ("blood") to its identity, which comes from the matriline." (email correspondence) For discussion of these issues, see, e.g., "Is there a Family? New Anthropological Views," Jane Collier, Michele Rosaldo, and Sylvia Yanagisako, in Rethinking the Family: Some Feminist Questions, Barrie Thorne and Marilyn Yalom, eds. 1982; and The Tiwi of North Australia, CWM Hart, Arnold R Pilling, and Jane C. Goodale, Thompson Learning, 2001 (third edition).

I don't think the anthro examples are considered controversial, but I expect there are critics. But if we allow that one's understanding of parent depends on social context, I can't see how the inference patterns about inherited kinds could be anything other than socially conditioned. Am I missing something?

[Ron Mallon]

Quick post:

As my comments above about sources of evidence probably make clear, I don't agree with Jackie Stevens that "examples from history or society are actually better evidence of who we are" in any absolute way. They are different evidence, and evidence that does not even approach the power of experimental settings to isolate causal elements. On the other hand, they do provide sorts of evidence unavailable in an experimental setting, and they are an important part of the story. I suppose it's also pretty clear that basically all experimental psychologists (nativist, empiricist, and otherwise) would want to resist the generally pessimistic global view of psychological studies that she suggests.

In response to her Aristotle point, as well as Sally's anecdote, I think the idea of Hirschfeld and others is that certain very specific kinds of thinking that are implicated in our racial practices are more ubiquitous than we might have thought. These get articulated in our cultural context in particular ways, for example, in ways such that color becomes associated with racial groups. As I said above, and Edouard said as well, there is evidence that that association only occurs in US children somewhat later.
So, its consistent with the Aristotle's case, since ancient Greeks did seem to divide humans into specific groups that reproduced themselves, biologically and culturally, even if they didn't think of color as important in this process. (See Benjamin Isaac's book The Invention of Racism in Classical Antiquity for this sort of line.)
And it's consistent with Sally's anecdote to say that the children had not yet learned to associate color with whatever sorts of "groups" (e.g. social groups, ethnic groups, coalitions, or whatever) this research posits. (As I said above, there are disagreements among even the "canalized" side of these cognitive researchers about what the nature of the capacity is that stake.)
Apropos of the earlier point about what counts as "race": maybe the right thing to say if this story is correct is that "race thinking" is acquired, because it does depend on social circumstances. (It sounds like Sally is tempted by this idea). You might be especially inclined to think this if you think that to count as "race" a classification must involve color. But (if the research is correct), the wrong thing to think is that it is acquired without also being structured by cognitive predispositions to categorize and infer in certain ways. In recognition of this idea, various kinds of "synthetic" approaches (like Machery and Faucher) attempt to explain cultural similarity and difference in their models.

As for the anthropological data: I'm not familiar with that exact case, but let's assume the details are exactly as Sally has them. The theoretical strategy would be to see if the case is an instance of folk essentialism, even if the essences are understood in radically different ways. For example, if members of different cultural groups outside this group count as members of that different group in part because their because their "matrilineage" is distinct, then this seems like a quite different cultural articulation of the essentialist idea that what you are depends not on what you choose but on how you were born. (Is that consistent with the case as you understand it, Sally?)
Still, much anthropology is invested in the idea that psychology does not have an explanatory role to play in understanding culture, so most of the anthropological evidence is going to emphasize the role of culture in cognizing human groups. That's not to dismiss it though: this extensive body of evidence needs to be addressed by these cognitive models. (I've been thinking recently about whether and how it can be.)

Okay, maybe that wasn't such a quick post… (the discussion has a number of interesting points!).

[Sally Haslanger]

So I think I'm finally getting the idea (sorry to be dense!). I thought that the nativist + constructivist model for race was supposed to have some aspect of the nativism specifically focused on race, though perhaps elaborated culturally. But now it seems that the nativism can be simply a propensity to attribute (non-chosen) essences to things, including groups of people. Then this propensity, in a given cultural context, can develop into racial classification. This seems much easier to swallow....but also explains much less about why we classify racially. Am I still missing something?

(The advantage of blogs is that you get to have these great conversations across time and space...the disadvantage is that I, anyway, tend to read them too fast during the middle of the term and don't always process as fully as I should!).

[Ron Mallon]

I think that's right. Different theories try to fill in the "why" a bit more, but there's not really anything approaching a consensus view.

Yes, this discussion has been really helpful! I've been thinking about these issues lately.

[Rob Wilson]

Thanks for all the comments on this thread--I've only just come to the blog in the last week or so. I'm currently working mostly on anthropological work on kinship, including some of the work that Sally referred to, and trying to make sense of the relationship between traditional and "revamped" work on kinship, divided roughly by David Schneider's work from 1965-1985. There is a lot of cross-over here with work on psychological essentialism in general, and I think on ethnicity and race in particular. I'll post something when I have something that might be of interest.

In the meantime, I guess I'll see some of you at the Phil of Social Science Round Table in Seattle this weekend.