Você está na página 1de 14

IAN HACKING

HOW NATURAL ARE KINDS OF SEXUAL ORIENTATION?


(Received 20 August 2001)

Edward Steins book is an encyclopedic study of, in the words of its subtitle,The Science, Theory, and Ethics of Sexual Orientation.1 It is an authoritative history of twentieth century scientic approaches to sexual orientations, too often described simplistically as the scientic study of homosexuality. It is also a rich body of analysis, distinctions and arguments. I was reminded of late mediaeval works that explained, with a ruthless thoroughness, every conceivable argument on a topic, pro and con. The book is being most noticed for the question posed by its nal chapter, Should Scientic Work on Sexual Orientation be Done? To oversimplify, Steins short answer is, probably not. Or at any rate, the research will not prove to be a good way to change attitudes to the variety of sexual orientations. I shall not address this question directly except for a few words at the end, but I do share many of Steins reservations. In 1990 Stein published a valuable collection of theoretical papers about sexual orientation.2 The contributors roughly divided as essentialist and constructionist. Stein included his own paper about the nature of that division of opinion, where he suggested that the distinction itself may be wrongly or misleadingly made.3 Notice
Publishers Note: Due to a misunderstanding, this paper was published in Issue 21/1 of Law and Philosophy. We apologise for this mistake. We reproduce it here, as was intended, as part of the four Discussion papers of this issue. 1 Edward Stein, The Mismeasure of Desire: The Science, Theory, and Ethics of Sexual Orientation (New York: Oxford, 1999). 2 Edward Stein (ed.), Forms of Desire: Sexual Orientation and the Social Constructionist Controversy, 2nd edn. (New York: Routledge, 1992). 3 Edward Stein, Essentials of Constructionism and the Construction of Essentialism, in Stein (ed.), Forms of Desire: pp. 325359. Law and Philosophy 21: 335347, 2002. Kluwer Law International 2002.

336

IAN HACKING

that those two terms, essential and constructed can be taken in two ways: individual and collective. In terms of an individual, an essentialist holds that ones sexual orientation arises independently of the social milieu in which one lives, although the ways in which the orientation is expressed depends on history and social context. There is a tendency to treat essential properties as part of ones intrinsic biological nature; something one is born with. In terms of a collectivity, an essentialist holds that the range of personal sexual orientations within any society is constituted independently of the social arrangements within that society, although their expressions would depend on those arrangements. The individual and collective versions of essentialism do not appear to differ much, but there are great differences within the family of ideas called constructionism. In the most simple-minded constructionist version about individuals, ones sexual orientation is a matter of choice and taste conditioned by the social world in which one grew up or moved into. Notice how this contrasts with one collective version, often associated with the work of Michel Foucault. He proposed that the idea of the homosexual, as a kind of person, came into being only in the course of the nineteenth century in Europe. This kind was the product of a specic legal and medical body of thought and institutions. You could not be that kind of person whatever that means before then. This doctrine is entirely consistent with the idea that in many times, places, and civilizations, individual sexual orientations have arisen as matters of choice and taste conditioned by social arrangements. But whatever the orientation, practices, choices and behavior, those would not produce a certain kind of person, a homosexual, unless they lived in or after the nineteenth century. From a logical point of view, then, Foucaults constructionism is entirely consistent with a thorough-going essentialism. His analysis need not be revised even if genetics or some other branch of science were to prove that every individual in the history of the human race had a pre-determined sexual orientation. It would remain the case (if he was correct) that only at a certain time did the idea of the homosexual as a kind of person come into being. (Many types of collective constructionism other than Foucaults are also formally consistent with essentialism, but I single out Foucault

SEXUAL ORIENTATION

337

because one of his essays serves as the rst contribution to Steins 1990 anthology.) Long ago I sketched one subversion of Foucaults proposal in a paper of my own, which Stein also reprinted. Although the paper began briey with the example of homosexuality, my intention was to show how new concepts and modes of description, including selfdescription, create new ways to be a person, and create new kinds of choices for an individual.4 That was only one analytic philosophers way of learning from Foucault. There have been many other ways to learn from him. Constructionism is a narrow and unimaginative name for the sorts of thoughts that Foucault made possible. The subtlety, and in my opinion the interest, of Foucaults ideas makes it unnattractive to use the blanket label, constructionist, except as a sort of pointer.5

CAUTION ABOUT THE IDEA OF NATURAL KINDS

Steins new book is in many ways a continuation by other means of the topics introduced in his own paper in his anthology. In part he wants to undo or undermine the seemingly sharp contrast between essentialist and constructionist attitudes. One of his strategies is to introduce a theoretical framework taken from analytic philosophy, namely the theory of natural kinds. In brief, and far more crassly than in Steins elegant analysis: are sexual orientations natural kinds or not? This is only one aspect of the book. I have already mentioned the topic of the nal chapter, namely, should the research be done? That is the third of three concluding chapters on ethics. The previous nine chapters are grouped in two parts, Metaphysics and Science. Most of the discussion of natural kinds is in the rst part. One reason Stein thinks the research is poor is that questions are ill-formulated, in part because the scientists have not claried the classications they try to use. I was asked to comment on the
Ian Hacking, Making Up People, in T. Heller et al. (eds.), Reconstructing Individualism (Stanford: Stanford University Press, l986), pp. 222236. Reprinted in Stein, Forms of Desire, pp. 6988. 5 Ian Hacking, The Social Construction of What? (Cambridge, MA: Harvard University Press, 1999), chapter 3, explains why I am leery of the label construction.
4

338

IAN HACKING

book because I have spent some time thinking about classication. I shall take that as my mandate, to be a sort of technical consultant on natural kinds. Stein is hardly the rst philosopher to ask, Is homosexuality a natural kind? But he is the rst to make plain some of the confusions involved in posing a question in this way. It is a fundamental premise of Steins book that there are a great many different kinds of sexual orientation. Moreover it is not at all clear that homosexuality is a kind of anything, natural or unnatural. Hence the switch to a more complex question: are sexual orientations natural kinds? The diversity of sexual orientations is not, in my opinion, the end of the difculty. There is something profoundly anomalous about this way of organizing things. Stein is addressing what I call realworld questions. Is it the case that some same-sex-gender orientations are correlated with certain genetic markers, as is claimed by Dean Hamer?6 Is it the case that some sexual orientations are correlated with certain familial arrangements experienced in childhood?7 Such questions sound quite direct, and imply conjectures about the root causes of sexual orientations. The questions do not mention natural kinds. I can think of almost no other group of real-world questions, aside from questions that are quite explicitly about taxonomy, that prompts anyone to ask, is so and so a natural kind? Why has natural-kind talk seemed attractive in connection with issues of sexual orientation? There is one embarrassingly banal answer to this question. Homosexuality was once said to involve unnatural acts. What better way to discredit that language than to prove that homosexuality is a natural kind? Even contemplating the possibility seems to undermine the preconception that homosexuality is unnatural. Is that why natural-kind talk has been so attractive here? If so, we are being misled by a bad pun, for the natural in unnatural acts and nothing to do with the natural in natural kinds. I have already misrepresented Stein: his main analytical tool is not the idea of a natural kind but of what he calls a natural
Dean Hamer and Peter Copeland, The Science of Desire: The Search for the Gay Gene and thje Biology of Behavior (New York: Simon and Schuster, 1994). 7 For example, Robert Trivers, Parent-Offspring Conict, American Zoologist 14 (1974): 229264.
6

SEXUAL ORIENTATION

339

human kind. Before I examine that, I must make a number of rather skeptical remarks about the idea of a natural kind itself. We do not, in ordinary English, talk about kinds just-like-that. We talk about kinds of seal, the harp seal, the hooded seal. We talk about kinds of cabbage and perhaps even kinds of king. But we do not talk about kinds, full-stop. Yet we do talk that way in Englishlanguage analytic philosophy.8 The practice began in 1840 (William Whewell) and was cemented in 1843 (John Stuart Mill).9 The practice was derived from a real-life eighteenth century debate in natural history, the predecessor of biology. There was a great discussion about natural groups or classes. Natural contrasted with articial. An articial grouping was a classication made up in order to sort out complex material, not because the group was to be found in nature, but because it helped us organize our thoughts. In botany we construct a taxonomic tree with species, genus, family, phylum, right up to kingdom. One popular view was that the species are natural groups, while even the genus was articial, a classication that botanists introduced in order to create a convenient system of naming the vast arrray of species found in nature. Another view was that the genus was natural, but anything higher up, the family, the phylum, or whatever, was articial, that is, a classicatory convenience. This type of dispute continues today, in the tub thumping world of systematic biological classication. The problem of which classications are natural is not peculiar to biology. The French mathematician, physicist, and philosopher, A. A. Cournot, writing in 1851, used groups of stars as an illustration. The ancient Greek system of constellations is (he argued) articial.10 The stars in Cassiopeia or Orion are grouped together
There is no technical term meaning natural kind in any language other than English, but as the next note indicates, kind has been given a technical sense in English philosophy since 1840. 9 William Whewell, The Philosophy of the Inductive Sciences, 2 vols (London: John W. Parker, 1840), Bk. VIII, ch. I, 4. The title of this section is simply Kinds. John Stuart Mill, A System of Logic, Ratiocinative and Inductive; Being a Connected View of the Principles of Evidence and the Methods of Scientic Investigation (London: Longman, 1843), Bk. I, ch. VII, 4. 10 A. A. Cournot, Essai sur les fondements de nos connaissances et sur les caract` eres de la critique philosophique (Paris: Vrin, 1975), pp. 199208.
8

340

IAN HACKING

for the convenience of navigators, and for ease of recognition by observers. But the nebulae, which had been rst recognized by William Herschel, are natural groups of stars. Cournot was the rst philosopher to propose an explanation of the idea of natural classications in terms of causality. The members of any class will have some properties in common that serve to dene the class. The class is natural when the members not only share properties, but also when the reasons, the causes, why they have those properties are the same throughout the class.11 There is every reason to think that the stars in a galaxy have related causal histories that cause them to be in the galaxy. There is no reason to think that the stars in Orions belt have parallel causal histories. This may seem a very long way from our topic, but that is what technical consultants are for. First of all, I wanted to remind you that the idea of a natural kind began in a fully intelligible real-world non-philosophical way. The voyages of discovery and European imperialism were bringing endless plants to Paris and London. How should we classify the incredible diversity of species distributed across the globe? The terminology has changed a bit, but there continue to be, at this very moment, real-life debates about which taxa are natural and which are articial. Taxonomy is the home of the debates, and real-life questions about which kinds are natural seldom arise elsewhere. The appearance of the same question in connection with sexual orientation is, to say the very least, unexpected. A second point is that Stein denes natural kinds by taking one element from the rather rich theory that we owe to Hilary Putnam: Natural kinds play a role in scientic laws and explanations.12
Cournot thought that classes were dened by shared properties. But one can generalize to clusters of properties produced by the same underlying causes. See Richard Boyd, What Realism Implies and What it Does Not, Dialectica 43 (1989): 529. Boyd adds the requirement that the clusters of properties should be stabilized by causal homeostasis; that is, they should be kept clustered by some sort of feedback mechanism. An excellent summary of Boyds ideas can be found in Frank Keil, Concepts, Kinds, and Cognitive Development (Cambridge, MA: MIT Press, 1989), pp. 4247. Keil is a psychologist who studies cognitive development in children; he was a colleague of Boyds when he wrote this book. 12 Stein, Mismeasure, p. 81. Putnam presented his theory in many places, but the classic is his 1975 paper, The meaning of meaning , reprinted in Hilary
11

SEXUAL ORIENTATION

341

This is really not enough to get at Putnams intended idea of natural kinds. Since Martha Nussbaums comments on The Mismeasure of Desire precede mine, I have no need to develop the point, for she has given several counterexamples, both compelling and amusing.13 I cannot improve on them, but it is useful to repeat the source of the difculty. It is none other than the above-mentioned point made by Cournot a century and a half ago. Playing a role in laws and explanations is not enough. We need causes. In the strongest case, the underlying causes that make something of kind K should be the same, right across all the members of kind K. We require something like Richard Boyds explication of this imprecise idea, as cited in note 10. And of course that is precisely the point of the numerous research programs examined by Stein. Even an added causal criterion provides, together with then rest, at most a necessary condition and not a sufcient one, but I shall not pursue the matter. Note that Dean Hamer (note 6) is looking not only for a correlation between genetic markers and sexual orientation: he conjectures that there is a gay gene that causes people to be gay. A third point is that we understand a qualifying adjective only when we know what it contrasts with. Natural kind originally contrasted with articial kind. An articial classication was one introduced by artice, for our convenience. Stein alludes to a more recent concept, that of an artifactual kind. This is a completely different notion from that of an articial kind. It was introduced in order to name classes of artifacts, things made by people and not found in nature. Pencils have been used as an example.14 The class of pencils is not an articial kind there is nothing articial about grouping pencils together as a class but a class of things that are articial, or at any rate a class of artifacts. Stein, however, uses a different denition. An artifactual kind is, roughly a group of things that have a common property only in virtue of human intentions. 15 Stein writes that if (per impossibile)
Putnam, Mind, Language and Reality. Philosophical Papers, vol. 2 (Cambridge: Cambridge University Press, 1981) pp. 215271. 13 Nussbaum, pp. 318320 supra. 14 Steven Schwartz, Putnam on Artifacts, Philosophical Review 87 (1978): 566574. 15 Stein, Mismeasure, p. 79. He says he takes this explanation from Keil, Concepts, op. cit., note 11. Keil was struck by the apparent fact that children

342

IAN HACKING

there were in the Rocky Mountains a spring gushing forth a liquid that is indistinguishable in taste and in chemistry from Diet Pepsi, it would not be a diet soft drink until humans discover it and interact with it in certain ways. I nd this confusing. Compare gold, which has long been one of the philosophers favorite examples of a natural kind. Gold was certainly not a medium of exchange and a standard of monetary value until humans discover[ed] it and interact[ed] with it in certain ways. But that does not mean that gold fails to be a natural kind. Likewise, naturally occurring dietpepsi would be a natural kind. (I use lower case letters deliberately, as in kleenex, a proprietary name become generic.) Since dietpepsi does not occur in the table of elements, it would be a less fundamental kind than gold, but, by hypothesis, it would be just as natural. And if we got all our diet soft drinks from this and other springs, diet soft drinks would not form an artifactual kind, that is, a class of artifact, any more than standards of monetary value do. I apologize for mentioning such pedantic niceties. I do so only to illustrate the fact that what seems to be plain sailing, in Edward Steins introduction to the ideas of natural and human kinds, is full of shoals, sharks, and other submarine perils. I question the usefulness of the natural kind idea for Steins purposes. The idea arose in the history of biological taxonomy. It has been invoked in developmental psychology, where it has been conjectured that an ability to sort items into at least some natural kinds is innate.16 But the idea has not found much employment in any other real-life topic. Both Bertrand Russell and W. V. Quine opined that although the notion of kind might have some value early in the life of a child (or of the human race) the more we nd out about the world, the less important the notion of kind becomes.17 I could explain why it came
so readily distinguish living creatures from artifacts, as if the ability to make this distinction is innate. Stein cites Keil, pp. 4751 (the pages directly after those in which Keil presents Boyds causal homeostasis as a characteristic of natural kinds). Keil listed nine types of difference between natural kinds and artifacts, of which at most one, labeled The need for intention seems to resemble Steins denition. 16 Frank Keil, cited in notes 10 and 15, is an authority on this topic. 17 The doctrine of natural kinds is only an approximate and transitional assumption on the road to more fundamental laws of a different kind. Bertrand Russell, Human Knowledge: Its Scope and Limits (London: George Allen and

SEXUAL ORIENTATION

343

to matter in philosophy, mostly in connection with the problem of induction, for Mill, for John Venn, for Bertrand Russell, for C. D. Broad, for Quine. I could relate how it became part of the theory of meaning and reference in the work of Hilary Putnam and Saul Kripke. Those were all worthy philosophical projects, but they were not much directed to any pressing real-life controversy. Edward Stein has endeavored to apply this philosophical structure to real-life questions. I am not sure that he was wise to do so.

NATURAL HUMAN KINDS

Steins real interest is in what he calls natural human kinds, a novel label. It is the chief theoretical notion in the rst part of his book. Here I must cast modesty aside. Stein writes, Ian Hacking introduced the notion of a human kind to talk about groups of people. Expanding on this notion, I [Edward Stein] use the term natural human kind for a natural kind that applies to people . . ..18 In his terminology, the contrast type is social human kind. In fact Stein did not expand on this notion, but diverted it. I am not troubled by that. I sincerely subscribe to the view that once you have said something, any serious reader and speaker can make any use of your words consistent with the common decencies of intellectual life. And if the original author thinks he was misunderstood, that is usually his fault, not that of the consumer. I claim no ownership of the phrase human kind. I recently abandoned it because it was too seductive a phrase. People had started using it to mean all sorts of things. So I invented some less attractive nomenclature. But rst human kinds. I should begin with a minor qualm about what Stein calls essentialism. In footnote 10 of his Prcis he asks us to note that his way of using the term essentialism differs from the way some philosophers use it. This is rather an understatement. In his Prcis Stein writes that essentialism about sexual orientation is the view that sexual orientations are natural human kinds, while constructionism
Unwin, 1948), p. 461f. For a more amboyant dismissal of the notion of kind, see W. V. Quine, Natural Kinds, Ontological Relativity and Other Essays (New York: Columbia University Press, 1969), pp. 114138, on p. 138. 18 Stein, Mismeasure, p. 84.

344

IAN HACKING

is the view that they are not. 19 This does not seem quite right, for it appears to be grounded on the thought that essentialism about Xs is the view that Xs are natural human kinds. But take X = is infected with tuberculosis. Being infected with TB plays a role in scientic laws and explanations (Steins criterion for being a natural kind). There are denite causes for such infections, namely bacteria (the causal criterion I attribute to Cournot and Boyd). I can think of no philosopher who would claim that being infected with tuberculosis is an essential property of a tubercular person. But now I wish to take up what at at the outset is just another tedious terminological issue, but which leads into interesting territory. I used human kind to contrast with natural kind. Human kinds, as I intended to use the expression, are not natural kinds! Natural human kind would, in my parlance, have been a virtual self-contradiction. I no longer own the words, but I should explain what I was trying to do. What interests me about classications of people is the way in which they interact with the people classied. This research interest began 18 years ago with the very paper that Stein reprinted, Making up People. The dynamics of human kinds is fascinating. Let a new way to classify human beings emerge, and let people become aware of how they are classied, then they will often behave differently (not necessarily better or worse, but differently). The truths about that category of people will change because the people have changed. In consequence, the classications may themselves have to be modied, for what is being classied has changed. Certainly the knowledge that the classications are used to encode will change. That is, classications interact with the classied. I titled one long paper The looping effects of human kinds.20 There are many other types of dynamics. That is precisely why I wanted human kinds to form a contrast with natural kinds. To be schematic: there are no looping effects of natural kinds. Phosphorus does not change once we label something phosphorus. Of course once we know about phosphorus, we may change it and use
Stein, supra., and on p. 84 of Mismeasure. Ian Hacking, The Looping Effects of Human Kinds, in D. Sperber, D. Premack and A. Premack (eds.), Causal Cognition. An Interdisciplinary Approach (Oxford: Oxford University Press, 1995), pp. 351383.
20 19

SEXUAL ORIENTATION

345

it. But simply calling phosphorus phosphorus makes no difference to phosphorus. Phosphorus (in our anti-animist world vision) never comes to know that it is phosphorus. But being called homosexual, or coming to think of yourself as homosexual, has made a great deal of difference to many human beings over the past 150 years. To speak in a ludicrous way to convey my simple meaning: Phosphorus does not come out, homosexuals do. There have been endless looping effects. The concept homosexual, on my view of history, was once owned by the medical and legal professions. Then the people categorized as homosexuals took over the ownership of the concept, and changed names, changed meanings, changed the world. Homosexual became what I have called a self-ascriptive kind.21 That (in my opinion) is just one of the ways that people and the ways they are classied can interact. There are many different types of looping. My label human kinds was intended to introduce an analysis of this phenomenon. Some time ago I pretty much abandoned the phrase, and would happily jettison natural kind too. I say, pretty much abandoned, for of course labels dont evaporate quickly. The chief thing I notice about natural kinds, cosmic or mundane, quarks, phosphorus, or mud, is that they are indifferent. They could not care, let alone care less, about how we classify them. It makes no difference to quarks or mud, how we classify them, although of course it does make a difference what we do after we have created a classication and found out something about it. But it makes a big difference to people, how they are classied. Not just because we praise some or incarcerate others, but because they come to think of themselves in new ways, to see new choices for action, to be new kinds of beings. Some classications interact with the people classied, with associated looping effects. These classications are what I call interactive. In a nutshell, my present philosophy of kinds comes to this: the indifferent and the interactive.22 These are not exactly mutually exclusive, natural and non-natural, as it were, but I shall leave the matter there at present.
21 22

Ibid., p. 380. See Ian Hacking, What?, op. cit., pp. 103106.

346

IAN HACKING

MOVING TARGETS

I would prefer not to address the topic of sexual orientation in terms of kinds, natural, indifferent or whatever. But since we have been set in this direction by Edward Stein, I should conclude by suggesting how my observations link up with his question, Should scientic research on sexual orientation be done? He observes that such research often ends up with a medical model. Medicine is predicated on a dichotomy, health and ill-health, the normal and the pathological, well-being and deviation from the norm. Medical research is directed at pathologies. One investigates schizophrenia, and seldom considers why most people are not schizophrenic. So, Stein reminds us, a research proposal in the medical tradition will target the question, why are some people homosexual, and not ask why other people are heterosexual. Stein also observes that scientic research is usually taken to aim at discovering human universals. Anthropologists may examine what is specic to a society, but on the medical model one is concerned with human pathologies, qua human universals, even if they are distributed differently, for reasons of nutrition, climate, or inheritance, in different populations. Stein remarks that it is not obvious that all or even many societies attach much importance to sexual orientation. There may be no universal to investigate. These points are well taken. I would add one supplementary observation. The very words scientic research on human beings now suggest the discovery of underlying biological structures physiological, biochemical, neurological, genetic, or whatever. It would be of enormous interest if a research team found a genetic marker or biochemical substance that was strongly correlated with certain sexual orientations. I do not believe that is going to happen. I do not expect that we shall obtain results of this type that stand up to persistent critical inquiry and elaboration. But I do not a priori rule that out. I am surprised daily, astonished weekly, and astounded monthly, by advances in fundamental life sciences. If I were on a panel distributing scarce research resources, I would be unenthusiastic about a grant proposal in this area. I would use Steins chapters to encourage the applicants to get clear about what they thought they were investigating.

SEXUAL ORIENTATION

347

Despite this skepticism, I also know that we often do not understand results until after we have got them and have had prolonged, acrimonious, critical debate about what they mean. So I do not exclude the possibility of real results in this research area, even if we might not be able to clearly dene what they would mean until after we had got them. But I see a different difculty. What is called scientic research assumes that it is investigating universals that are coded by indifferent classications, classications unaffected by what we nd out. I have been suggesting that the wide range of sexual orientations constitute interactive kinds, for which the entire model of biologized scientic research is inappropriate. Biology aims at a xed target of indifferent kinds. We are here concerned with a moving and interactive targets that wont stay still. More power to them.
Collge de France 11, Place Marelin Berthelot 15005 Paris France

Você também pode gostar