In recent years several philosophers of biology have proposed a pluralistic approach to science. This move away from monism has been largely motivated by attempts to wrestle with “the species problem”—the challenge of providing criteria for species membership (e.g. Kitcher 1984, 1989; Ereshefsky 1992; and Dupré 1993). In The Disorder of Things, John Dupré argues for a version of pluralism that encompasses more than just a pluralist solution to the species problem. As an anti-essentialist, Dupré advocates what he calls “promiscuous realism,” which is the claim that “there are many equally legitimate ways of dividing the world into kinds” (Dupré 1993, p. 6). As an anti-reductionist, he espouses pluralism with respect to the reality and causal efficacy of entities at many different levels of organization. Pluralists of all breeds must deal with a familiar class of worries that are routinely expressed at the suggestion of any rejection of monism. One such worry is that pluralism is a relativistic position in which “anything goes” in science. If there is no one privileged classification of objects into kinds, then are all classificatory schemes equally legitimate? That is, could a pluralist evaluate one classificatory scheme as better than another? Does anti-reductionism imply that any level of explanation is as genuine as any other? In the problem’s most general formulation: is there no way to distinguish “good science” from “bad science”? Dupré responds to such concerns with the reassurance that he has several tools at his disposal to “try to sort the scientific sheep from the goats” (Dupré 1993, p. 263).
In this paper I examine Dupré’s proposals for saving his pluralism from the much-feared overly permissive relativism. I use a case study to illustrate the type of evaluation of science he has in mind. This study of the debate surrounding the use of anthropomorphic language in sociobiological accounts of rape shows Dupré’s tools for distinguishing between good and bad classification in action. This particular example has been chosen not only because it is a useful example of criticism of scientific classification, but also because it provides the opportunity to discuss an aspect of anthropomorphism that is usually ignored in philosophical and scientific discussions about the use of descriptive vocabulary in science. This aspect is the extent to which anthropomorphic language poses moral, social and political problems, as opposed to purely scientific or epistemic problems. Thus, this paper has two primary goals. First, in section I, my aim is to explain and illustrate Dupré’s method for distinguishing good science from bad science. Second, in section II, I attempt to defend Dupré’s mode of evaluation against some pressing objections.
How does the promiscuous realist distinguish good classifications from bad ones?
--A case study in anthropomorphism
Towards the end of The Disorder of Things, Dupré writes, “It would strike me, for example, as a fatal flaw in my position if it led to the conclusion that nothing could be said in explanation of the epistemic superiority of the theory of evolution over the apparently competing claims of creationists” (Dupré 1993, p. 242). In order to avoid this devastating defect in his position, Dupré proposes certain means for determining which accounts of the world have epistemic authority. The means at his disposal are what he calls epistemic virtues. Among the class of possible and actual epistemic virtues, Dupré includes the following: “sensitivity to empirical fact, plausible background assumptions, coherence with other things we know, exposure to criticism from the widest variety of sources, and no doubt others” (Dupré 1993, p. 243). These virtues are tools for evaluating forms of knowledge production. For example, Dupré thinks that the sophisticated argumentation, accumulation of empirical support and years of theoretical criticism that have generated evolutionary theory show that evolution possesses many more epistemic virtues than does creationism (Dupré 1993, p. 243). Dupré, therefore, responds to the relativistic worry by proposing that “Such an approach [based on epistemic virtues] would at the very least have the capacity to capture the rich variety of projects of inquiry, without conceding that anything goes” (Dupré 1993, p. 243). Those things that do not “go” are practices deficient in epistemic virtues.
Epistemic virtues are not the only tools of which Dupré avails himself. In various parts of The Disorder of Things, and especially in the last chapter, he argues that there are certain cases in which it is legitimate to evaluate science on purely moral or political grounds. Epistemically unobjectionable conceptualizations of the world can be, on this view, attacked for their moral and political implications. For instance, particular economic definitions of the concepts of “household” and “income” have come under attack for the ways in which they can devalue the contributions of women and form the basis of social policies that fail to promote the well-being of women and children (Dupré 1993, p. 251). Dupré thinks that, in the absence of any evidence of factual mistakes on the part of the economists, it is legitimate to object to their work for political reasons. This then is another way that the pluralist can avoid an anything-goes position. Those scientific practices that do not “go” are those that have significantly negative social and moral implications. Dupré, therefore, has two types of standards against which to measure science: epistemic virtues and moral/political virtues. As Dupré puts it, “…scientific practices vary from the highly plausible to the quite incredible and, often in a parallel fashion, from the extremely valuable to the entirely pernicious” (Dupré 1993, p. 262).
In two recent papers, Dupré uses these two methods of evaluation to attack reductionist explanatory programs like evolutionary psychology and rational choice theory (Dupré 1994, 1998). For instance, Dupré criticizes evolutionary psychology for its failure to provide adequate empirical support for hypotheses. While Dupré does not use the terminology of epistemic virtues in these papers, his argument is essentially that evolutionary psychology lacks the epistemic virtue of sensitivity to empirical support. Dupré also argues that evolutionary psychology and rational choice theory can be morally pernicious. His recent work, therefore, illustrates how Dupré qua anti-reductionist can evaluate scientific practices using the epistemic and moral virtues as standards.
In the rest of the paper I want to explore the other side of Dupré’s pluralism: his promiscuous realism. The promiscuous realist claims that “there are many equally legitimate ways of dividing the world into kinds” (Dupré 1993, p. 6). While this pluralist recognizes a wider variety of valid classificatory schemes than the monist does, the promiscuous realist does not accept every scheme as equally valid. Just as Dupré’s recent papers show how the anti-reductionist brand of pluralism evaluates explanatory projects, this paper demonstrates how the promiscuous realist evaluates classificatory schemes.
The controversy surrounding anthropomorphic use of language in sociobiology provides a useful case study. One general way to characterize this debate is that worries about anthropomorphism are concerns about the unsound use of language to describe non-humans in cases where that language has its primary meaning in human contexts. This use can be unsound in that it makes a category mistake or makes unwarranted inferences about the mental lives of animals or, as has been less frequently argued, because it has negative moral consequences. The charge of anthropomorphism has been leveled against the descriptive language employed by sociobiologists. In particular, a controversy has arisen over the use of “rape” to describe the behavior of non-human animals. In what follows I will summarize the use to which sociobiologists have put this term, and will present two types of criticisms against this use. Each of these types of attack rely on a different one of Dupré’s methods of evaluation. The first type of criticism attacks the sociobiologists’ classificatory scheme on the grounds that it fails to live up to epistemic standards. The second type of criticism launches a moral criticism of the anthropomorphic language. My claim is that Dupré’s epistemic and moral virtues are standards according to which critics can attack, and have in the past attacked, a classificatory scheme.
Sociobiological classification and rape
Several sociobiologists have described what they call “rape behavior” in non-human animals (e.g. Abele & Gilchrist 1977, Barash 1977 and Thornhill 1980). I will focus on Randy Thornhill’s work with scorpionflies (Mecoptera: Panorpidae). In an ambitious 1980 article, Thornhill not only argues that male scorpionflies engage in rape, but he also interprets this behavior in light of sexual selection and proposes a hypothesis that predicts the occurrence of “true heterosexual rape” in species that meet certain criteria (Thornhill 1980, p. 52). Thornhill describes rape in scorpionflies as follows:
A rape attempt involves a male without a nuptial offering (i.e. dead insect or salivary mass) rushing toward a passing female and lashing out his mobile abdomen at her. On the end of the abdomen is a large, muscular genital bulb with a terminal pair of genital claspers….He then secures the anterior edge of the female’s right forewing in the notal organ (Mickoleit 1971), a clamp-like structure formed from parts of the dorsum of the male’s third and fourth abdominal segments. Females flee from males without nuptial gifts. If grasped by such a male’s genital claspers, females fight vigorously to escape….The male retains hold of the female’s wing with the notal organ during copulation, which may last a few hours in some species.
Thornhill goes to some length to defend his classification of this behavior as rape. He defines rape as “forced insemination or fertilization” (Thornhill 1980, p. 52). Heterosexual rape, on which his paper focuses, is further defined as involving “a male inseminating an unwilling female or fertilizing an unwilling female’s eggs without insemination as in species with external fertilization” (Thornhill 1980, p. 52). This definition supplies two key criteria that must be met before any behavior can be classified as heterosexual rape (hereafter, rape). First, the female must be unwilling. Second, the female must be inseminated or have her eggs fertilized. In species with internal fertilization, intercourse without ejaculation does not count as rape even if the female is opposed to the intercourse.1
While it is possible to ascertain when a female scorpionfly had been inseminated using an experimental procedure (removal of the sperm-storage organ and examination of its contents under microscope), trying to determine whether a scorpionfly, or any other non-human animal for that matter, is unwilling is far from simple. Not only are we limited to behavioral evidence as a guide, but the observed aggressive grabbing by the male and the struggling movements of the female admit of more than one interpretation. Thornhill puts the problem as follows: “Courtship often involves components of aggression (Bastock 1967; Thornhill MS); thus, what appears a heterosexual rape may be merely a male successfully overcoming the discriminatory behaviour of a female by aggressive courtship” (Thornhill 1980, p. 52). What looks like a female resisting rape may instead be a “coy” female testing the strength of her potential mate to determine whether he is likely to provide her offspring with traits adapted to life in a dangerous environment (Mitchell 1996, p. 139). Thornhill’s definition of rape, therefore, presents a problem to the investigator—how to distinguish rape from female coyness. Thornhill’s solution is that when it is not in the female’s reproductive interests to mate with the aggressive male, then she can be considered to be unwilling to be inseminated by that male. In such cases, if the insemination occurs, it is an act of rape. Thornhill uses his observations of the behavior and ecology of scorpionflies to argue that when a female is approached by a male who offers no dead insect or salivary mass to her, it is not in her reproductive interests to mate with that male (Thornhill 1980, p. 55-6). Therefore, when aggressive courtship between such a male and female is observed, it is appropriate to infer that the female is unwilling and that it is an act of rape. As Sandra Mitchell puts it “Thus the analog to unwillingness [in Thornhill’s definition of rape] is reproductive disadvantage” (Mitchell 1996, p. 139).
Thornhill’s classification of rape involves one additional criterion that any putative act of rape must meet. He stipulates that “In order to demonstrate male rape behaviour it is necessary to…show that males that rape enhance their own fitness” (Thornhill 1980, p. 52). If the behavior is not in the reproductive interests of the males, then it does not count as rape. Thornhill does not explain the purpose of this additional criterion, but I think Mitchell probably diagnoses it correctly when she says:
What is the motivation for these criteria? The descriptive content of the term “rape” must be derived from its use in human social contexts….[In these contexts] there is an essential intentional component. For it to be rape two psychological conditions must be true, the female must be unwilling to engage in sexual intercourse, and the male must be willing (Mitchell 1996, p. 138).
If Mitchell is right, Thornhill introduces this criterion of male reproductive advantage because he wants to draw on the common usage of “rape” as something that the rapist must be willing to engage in. Again the effect of the behavior on the individual’s fitness determines whether or not one can say that the individual was willing to engage in it.
In summary, a behavior must meet the following three conditions in order for it to count as rape according to Thornhill: 1) the female is inseminated, or in the case of species with external fertilization, the female’s eggs are fertilized, 2) the female is unwilling to be inseminated or have her eggs fertilized, where “unwilling” means that it is against her reproductive interests, 3) the male is willing to inseminate the female or fertilize her eggs, where “willing” means that it is in his reproductive interests to do so. On the basis of these criteria, Thornhill argues that male scorpionflies rape females. Thornhill also frequently refers to human rape in this 1980 paper and his later works (Thornhill & Thornhill 1983; Thornhill & Palmer 2000). Thornhill, therefore, takes these criteria to be providing a general characterization of rape, one that allows for the classification of both humans and non-human animals. It is this classification that is the subject of various attacks on sociobiological conceptions of rape. The first attack illustrates the use of Dupré’s epistemic virtues in evaluation of classificatory schemes, while the second attack relies on moral virtues as standards.
Criticisms of sociobiological use of “rape” The epistemic sins of Thornhill’s classification scheme
One way to evaluate a classificatory scheme is to take stock of its epistemic virtues. Anthropomorphism has routinely been viewed as an epistemological error. For instance, Philip Kitcher has noted that “Some critics have viewed anthropomorphism as the original sin of pop sociobiology. The sin lies in neglecting to investigate the kinship of forms of behavior that are superficially similar” (Kitcher 1985, p. 184-5). To be guilty of anthropomorphism can, therefore, mean that one has not provided adequate evidence to justify describing human and non-human behaviors as similar in certain respects. A classification that commits this crime can be charged with lacking the epistemic virtue of sensitivity to empirical fact. Sandra Mitchell has leveled this charge against Thornhill’s work on rape. Her criticism is an example of how epistemic virtues can be brought into play in evaluating classificatory schemes.
Mitchell claims that Thornhill’s work is, “an example of the illegitimate grouping together of disparate behaviors under one descriptive category” (Mitchell 1996, p. 136). The main target of her argument is Thornhill’s evidence that the scorpionfly behavior meets his criteria for inclusion in the category of rape. Mitchell disputes Thornhill’s claim that his evidence is sufficient to distinguish female coyness from rape. She also argues that Thornhill does not eliminate another competing explanation of the female’s behavior:
Furthermore, given the aggressive nature of the interactions with conspecifics, heterospecifics, and predators in that environment, the female’s behavior may have nothing whatsoever to do with her willingness or unwillingness to copulate with a given male. Rather than struggle indicating an instinct to avoid copulation, it may be part of a different behavioral set. She may be avoiding being grasped in general, a behavior much to her advantage (Mitchell 1996, p. 139).
Mitchell points to the fact that “rapist” males do not emit the pheromone by which the flies recognize members of their own species. This fact helps to make her alternate explanation a live option. A female grasped by an individual that she does not recognize as a conspecific may struggle as a result of an adaptation to a dangerous environment. Since Thornhill does not provide evidence that eliminates this alternate explanation, he has not, according to Mitchell, sufficiently shown that the behavior he describes in scorpionflies is rape.
The crux of Mitchell’s critique is designed to show that Thornhill’s use of “rape” does not meet scientific standards. Superficial similarity is not sufficient to support scientific classificatory schemes. Estep and Bruce also make this point in their attack on sociobiological use of “rape”: “We argue that to apply a human label to the behavior of non-humans does not necessarily make the events the same….The loose application of such words as rape is imprecise and does not further the goals of science” (Estep & Bruce 1981, p. 1273). Classificatory schemes that do not possess certain epistemic virtues, like sensitivity to empirical fact, have often been attacked on the grounds that they do not “further the goals of science.” These attacks, therefore, illustrate one way that Dupré’s virtue epistemology can distinguish between good and bad classificatory schemes. A second way for the promiscuous realist to evaluate classificatory schemes is through moral or political criticism of science. The next section shows how Thornhill’s definition of rape has been rejected on moral or political grounds.
While the previous critique emphasizes the epistemic failings of anthropomorphic use of language, these are not the only grounds on which critics have objected to anthropomorphism. Several authors have argued that sociobiological anthropomorphism can create serious moral and social problems (e.g. Gowaty 1982, Fausto-Sterling 1985 and Kitcher 1985). These criticisms are examples of the second way I suggested that the promiscuous realist can evaluate classificatory schemes.
Patricia Adair Gowaty launches an attack based on the social dangers of this use of “rape.” One of Gowaty’s main reasons for objecting to Thornhill’s classification is that the use of “rape” in a sociobiological context is “jargon with possible social repercussions” (Gowaty 1982, p. 630). She claims that the sociobiological definition of “rape” deviates so widely from the common usage that it essentially turns “rape” into a term of art. According to Thornhill’s criteria, the impact on the fitnesses of both the victim and the rapist must be taken into consideration when determining whether an act constitutes rape. Common usage makes no such demand. As Gowaty says, “This [sociobiological concept] is a different concept from the one used commonly for humans to mean any act of forced intromission or sexual assault” (Gowaty 1982, p. 630). The sociobiologists have, therefore, attached a new connotation to the term “rape.” Why should we be troubled by this turn of events?
Gowaty’s concern stems from her analysis of the potential social dangers resulting from the discrepancy between common usage and Thornhill’s definition of “rape.” To understand the source of this discrepancy, imagine the following scenario taken from Gowaty’s own experience of an exchange at a scientific meeting. You are asked, “Is it rape when a post-menopausal woman is forced to intercourse?” How should you respond? Clearly the answer you give will depend on which usage you adopt. If you are a typical English-speaker, your answer will likely be that the scenario describes an act of rape. Since common usage classifies acts of forced intercourse as rape, the post-menopausal woman has been raped. However, if you are a Thornhillian sociobiologist, then your answer might be different. Since the rapist will produce no offspring from the intercourse, his fitness will not be increased and the act will not count as rape.23 Consider another question that reveals a similar gap between standard usage and sociobiological classification, “Is it rape when penetration without ejaculation is involved?” Again, I would argue that common usage is that “rape” applies to any act of forced intromission. In contrast, Thornhill’s definition requires that the term “rape” not be attached to this behavior since it fails to meet the first criterion of insemination or fertilization. These examples point to areas of mismatch between the common and sociobiological conceptions of rape. This mismatch occurs when the folk system classifies a behavior as rape, while the sociobiological scheme classifies the behavior as non-rape.
The common usage of rape is closely tied to “the traditional, psychological, sociological, and legal ways we think of rape” (Gowaty 1982, p. 630). For example, the legal system affords redress to women who have been subject to forced intercourse regardless of whether they are pre-menopausal, menopausal, or post-menopausal. Here lies the heart of Gowaty’s critique. She says, “The social danger I see is the possibility of transference of the sociobiological idea to the legal realm so that legally, victims of rape will be said not to have been raped” (Gowaty 1982, p. 630). It is not the mismatch itself that worries her, but rather the possibility that the sociobiological classification will influence the legal conception of rape in ways that will negatively impact victims of rape. Specifically, the negative impact she fears is that victims of forced intercourse will be unable to gain access to the protections and means of redress provided by the legal system because the sociobiologically–influenced legal conception of rape will not classify what happened to them as rape. To put the matter most starkly, if the law changes to incorporate Thornhill’s classification of rape, then the post-menopausal woman of Gowaty’s scenario would not be able to charge her attacker with rape. In response to those who would say that the chances of such a drastic reaction to Thornhill’s work are slim, Gowaty replies, “This is not so unlikely. A similar reversal, in which the victim’s motives and behavior are tried, has led to the further victimization of women in police precincts and courts when they report rapes (Brownmiller 1975)” (Gowaty 1982, p. 630). Given this precedent, we should be concerned that the proliferation of Thornhill’s conception of rape will have a negative impact on the way the legal system treats victims.
Gowaty’s criticism of Thornhill’s conception of rape is, therefore, fundamentally different from that of Mitchell. The latter attacked Thornhill for failing to provide a classification of rape that possessed certain epistemic virtues. Gowaty’s critique, on the other hand, does not question the epistemic merits of Thornhill’s work. Instead, it focuses on the threat his classification poses to society. Gowaty’s criticism, therefore, represents a response to anthropomorphism that focuses on its moral and political implications. In her critique, we find an example of Dupré’s suggested moral evaluation of science.
Having illustrated the use of Dupré’s proposed use of epistemic and moral/political virtues to evaluate scientific classifications, I will now turn to a defense of Dupré’s proposal.
Evaluating science on moral and political grounds—A defense of Dupré’s proposal
Of the two modes of evaluation proposed by Dupré, the use of epistemic considerations to make normative judgments about science is surely less controversial than the use of purely moral or political considerations. While there has been fierce debate surrounding the need for, and status of, epistemic norms in science, there are few philosophers, save perhaps Paul Feyerabend, who reject the use of norms outright. Even those who deny that strict criteria can determine theory choice find some value in less restrictive criteria such as Kuhn’s list of values like simplicity and consistency (e.g. Longino 1990). Therefore, while there might be extensive discussion about the importance and number of the epistemic virtues, Dupré’s general proposal to evaluate science according to some epistemic principles does not stand in dire need of defense.
The same, however, cannot be said for Dupré’s use of moral and political virtues. There is intense resistance in many quarters to the notion that moral and political considerations should play a role in the acceptance or rejection of scientific ideas. To reject a scientific theory because it conflicts with one’s moral or political beliefs is thought to be akin to choosing to live in a dream world. The mere influx of moral and political considerations into science is also thought to undermine one of the merits of scientific practice—the objective, unbiased investigation of nature. Even those who accept the growing evidence that moral, political, social and ideological values play a role in the process and product of science, who accept that science is not completely value-free, maintain that science should continue to strive to eliminate such prejudices as much as possible. As David Hull puts it, “Perhaps characteristics of our society affect the conduct and content of science, but they shouldn’t” (Hull 1994, p. 704). For those of this view, Dupré’s plan to bring moral and political considerations to the fore in evaluating science will seem to take science a further step back into the mire of biased opinions. Hence, it is this part of the Dupré’s method of evaluating science that stands most in need of defense. In what follows, I defend Dupré’s proposed use of moral and political evaluation of science from a number of pressing objections.4
I draw the first two objections from Hull’s review of The Disorder of Things. In this essay Hull responds to Dupré’s discussion of moral and political effects of science as follows: “Dupré details the harmful effects that essentialism, reductionism, and determinism have had on societies through the years….If these three philosophical theses have had any good implications for society, Dupré does not mention them. But as I see it, any idea can be abused, including those that Dupré himself prefers (Hull 1994, pp. 704-5; emphasis added). Hull’s comment that “any idea can be abused” forms the basis for two pressing objections. I will address each in turn.
Objection A: Any idea has the potential for abuse, but that does not mean that it will necessarily have the negative moral consequences imagined by moral critics.
On this reading, Hull is basically charging the moral critics with excessive worrying. He agrees that it is possible for a scientific idea to be abused, but he disagrees that this is the whole story. The very same scientific idea may, in fact, not be abused and may have a positive impact on society. Hull might point out that there is surely a possible world in which I am struck by lightning upon leaving this building, but it would obviously be foolish for me to thereby refuse to move when I have no reason to think a lightning storm is on the horizon. Similarly, for every scientific idea there is a possible world, or maybe several, in which it is abused, but it would be foolish to thereby reject the idea outright.
To make this objection more concrete, let me explain why it I think it has some merit. Critics, like Dupré, who point to the negative social consequences of science can appear to overstate their case. By simply pointing to one negative implication of a scientific view, they can misrepresent the actual situation. Take Gowaty’s criticism of Thornhill’s work, for example. She argues that Thornhill’s reconception of rape poses the social danger of changing the legal definition of rape, which would thereby prevent many victims from getting redress. She is surely right that this is one possible result of the proposal of a sociobiological conception of rape. But there are undoubtedly also other potential outcomes.
Whenever a new scientific classification is proposed, there are a number of conceivable consequences. A list might include some of the following: 1) The scientific classification could sweep away the common usage. This replacement could be, a) detrimental to society because the, now extinct, common usage played some important role (e.g. in the legal sphere) or b) beneficial to society because the common usage was pernicious or not useful (it might have been too vague, prejudicial or incoherent, for example). 2) The scientific classification could be restricted to use in the scientific domain leaving the common usage to survive intact in other domains. 3) The scientific classification might be rejected by all members of the linguistic community, including scientists, leaving the common usage intact in all domains.
The social danger envisioned by Gowaty is that Thornhill’s definition will result in an effect of the kind described in 1a. She worries that Thornhill’s classificatory scheme will replace common usage resulting in a detrimental effect on the legal coverage afforded to rape victims. She is certainly right that this is a possibility, but by failing to mention the other potentialities, she could be accused of making a mountain out of a molehill. Thornhill could argue instead that his classification might follow route 1b. Thornhill thinks that the common ways of thinking about rape have prevented us from eradicating the crime. He could claim that it would be a good thing for his classification to sweep away common usage.
One way to interpret Hull’s criticism is, therefore, that scientific ideas can have both negative and positive consequences and that simply pointing to one possible negative outcome is insufficient grounds on which to reject the idea. These are insufficient grounds because the imagined detrimental impact on society is only one possible outcome. Just as we do not take only one possible outcome of our actions into consideration when making personal decisions, we should not accept or reject science without looking at all of the facts. Moral critics do make themselves liable to this objection when they tell alarmist stories about possible devastating effects of science on society. If they ignore the fact that the science could also do some good, then they are just not taking all the facts into consideration.
That said, I do not think that this version of Hull’s objection deals a death blow to all moral and political criticism. In order to avoid this type of objection, I suggest critics like Gowaty and Dupré follow two commonsense guidelines for making moral criticisms. First, both the potential positive and negative outcomes should be elaborated. This does not mean that moral critics bear the burden of foreseeing in advance all potential results of a scientific view, but they do have to mention enough of the positive and negatives ones to convince the audience that they are not painting an incredibly slanted picture. Second, some reasons or evidence to believe in the likelihood of the negative moral or political outcome should be offered. Gowaty does, in fact, do this in her article. Her moral criticism is strengthened by the fact that she refers to Brownmiller’s work on the treatment of victims in police precincts to back up her claim that Thornhill’s definition will lead to a devaluing of the rape victim’s experience. It could be also be argued that in the context of a society in which women maintain a marginal presence in the legal system, it is more likely that Thornhill’s definition will be used in ways that are insensitive to the needs of women. Recognition of the social context in which a scientific idea will circulate can often reveal whether it is likely that the idea will be abused.
Once moral criticisms of science have these two guidelines in place, they are much less vulnerable to this version of Hull’s objection. It may be true that any idea can be both misused and used wisely, but it does not follow from that that some ideas are not, given the social context, more prone to misuse than others. Moral criticism could still be leveled against those ideas that have a high probability of misuse. Such criticism is not so easily dismissed as groundless worrying. It makes a lot of sense to be hesitant to leave the building during a lightning storm. Likewise, it is reasonable to worry about the misuse of a scientific idea when the social context is ripe for such abuse. That said, there is another interpretation of Hull’s comment that poses a further challenge to moral criticisms. This next objection focuses on the appropriate response to these levelheaded worries about the abuse of science.
Objection B: Since it is the social context which largely determines whether an idea will be used or misused, the target of moral criticism should be the social milieu rather than the scientific claims.
Another way to read Hull’s comment that “any idea can be abused” is as an attempt to show that the blame for the negative consequences of science rests on the heads of those who abuse the idea. On this view, scientific ideas are not (and cannot be) prejudiced. People are prejudiced. Scientific ideas should not be blamed for the bad things that happen as a result of their proliferation in a community. Instead, it is the society that puts a scientific idea to harmful uses that should be taken to task. In the case of rape, one might argue that feminists like Gowaty should aim their attacks at the society that will abuse Thornhill’s definition of rape.
As with objection A, this objection does have some merits. It is important to recognize that scientific classifications can do no harm on their own. Classifications need people to apply them and use them to make judgments about people in order for any pernicious effects to result. An analogy between moral criticisms of scientific classifications and justifiable censorship may be helpful here. In a society that protects free speech, the right to express any idea is protected in all but the most extreme situations, but one’s freedom to act on the basis of one’s ideas is, to a large degree, restricted. For instance, in the comfort of one’s home, one is free to express approval of abortion clinic bombings. However, one would certainly be arrested in attempting to carry out such a bombing. Objection B suggests that we take a similar approach to protecting scientific ideas. Scientists should be free to propose any classification of things, and we should not attempt to forbid the expression of these ideas. Instead, we should restrict attempts to use the classifications to the detriment of society. This suggestion makes a lot of sense in some situations. Scientific freedom is certainly valuable. If, by simply restricting the use of classificatory schemes, we can avoid the moral and political disasters imagined by moral critics without gagging the scientist, then a win-win situation results. For example, a proponent of objection B might respond to Gowaty’s critique by proposing that we avoid her imagined social danger by directing the legal system not to use Thornhill’s definition. In this way the social disaster is averted and Thornhill is still free to use his definition of “rape” in his biological work. In many cases such a response to a moral criticism might well be appropriate. However, there are situations in which our options are not so inviting. It is in these situations, I will argue, that objection B’s response to moral criticism is misguided.
Rejection of scientific ideas on moral grounds is justified when the potential social damage of the idea far outweighs the damage to scientific freedom that results from rejecting it. There are times when a scientific idea can be as dangerous as yelling “fire!” in a crowded theater. Some of the horrors of Nazi science certainly jump to mind here. When the social context is such that promulgation of some classificatory scheme is likely to justify loss of innocent life, I think it is morally irresponsible to advocate that idea in the name of scientific freedom. It is in such situations that I think the advice of objection B is misguided. The analogy to objection B in the free speech situation would have us yell “fire!” and then try to stem the ensuing stampede. To try to restrain the responses to such utterances in explosive situations is simply ineffective. Sometimes the only way to avoid disaster is to curtail speech. Similarly, I would argue that there are cases when rejection of scientific ideas on moral grounds is the only way to avoid disaster.
It will surely be difficult to draw a line between those cases where moral criticism is best aimed at the use of an idea and those times when the only safe course of action is to squelch the idea altogether. My case study of anthropomorphism in sociobiology illustrates this difficulty. The question here becomes, is the social danger envisioned by Gowaty serious enough to justify rejecting Thornhill’s definition of “rape”? One might be initially tempted to think that the situation must not have been that serious since in the years following Thornhill’s 1980 paper we have not seen a dramatic decrease in the legal protection afforded to women. It would seem, therefore, that Gowaty did not provide reasons to justify abandoning Thornhill’s definition. Instead, she may have just given us reason to guard against its misuse in the legal sphere. However, on the other hand, one might also note that Gowaty’s attack on Thornhill was simply supposed to persuade him to use a different term other than “rape” to describe the behavior of non-human animals. It could be argued that asking Thornhill to use “forced copulation” instead of “rape” is not an extreme infringement on scientific freedom. Therefore, if the cost to scientific freedom is negligible, then all that Gowaty would need to do is show a slightly greater cost to the social status of women were Thornhill to keep using “rape.” I am not going to attempt to resolve this particular issue about Gowaty’s criticism in this paper. However, I do hope to have shown that if you believe that yelling “fire” in a crowded theater is morally condemnable, then you should similarly believe that in some situations scientific ideas can be rejected on moral grounds.
To summarize, I have by now defended moral criticism against both versions of Hull’s objection that “all ideas can be abused.” In the course of this defense, I have argued that moral critics should follow several guidelines in order to avoid the challenges posed by Hull’s critique. First, they should outline both the potential positive and negative outcomes of the scientific idea under attack. Second, moral critics should provide some reasons or evidence to believe in the likelihood of the negative moral or political outcome. Third, my response to objection B suggests that moral critics who advocate rejecting a scientific idea outright should make it clear why the benefits of avoiding the moral disaster outweigh the costs to scientific freedom. Moral criticisms that adhere to these guidelines are not vulnerable to objections A and B. However, there is a final objection that I would like to address.
Objection C: Moral criticisms are intellectually dishonest because they ask us to reject the truth about the world in favor of some palatable fiction.
A third objection to moral and political criticism is a common one that finds some expression in the following quote from David Hull:
…science might well change so much that political considerations take the place that theology once played. Politicians, bureaucrats, possibly the People will decide not only what sort of a world they would like to live in but also what this world actually is, the results of scientific investigation of the old fashioned sort to one side. Under such circumstance, science as an historical entity will still be science, but it will cease to instantiate a kind that I find worth studying (Hull 1990, p. 86).
In this passage, Hull expresses aversion to giving political considerations a role in science. Hull seems to think that such a move means that we will no longer be interested in discovering what the world is really like. Instead, politicians and the People will be involved in deciding what the world is. Opponents to moral criticism often charge moral critics with being afraid to face up to the facts. This is a rhetorical strategy which Thornhill and Palmer repeatedly exploit in their book as they make social scientists and feminists out to be cowards who will not face the fact that rape has a biological basis. It is a strategy which is often successful. The specter of the Darwinism vs. Creationism debates still haunts biology and moral criticism is often viewed as an attempt to uphold a dogma in the face of clear scientific evidence to the contrary. But does all moral criticism espouse fiction over fact? I do not think so, and in my response to this final objection I will argue that certain types of moral criticism are immune to this type of attack.
This objection to moral criticism misses its mark because it presupposes that the scientific idea under attack represents a discovery of some immutable fact about the world. I will argue that objection C relies on the assumption that science aims at uncovering, rather than creating, order in nature. Science is, on this view, an attempt to uncover regularities in nature. Kepler’s laws are, for example, descriptions of some regularity found in the motion of planets. When scientific ideas are viewed as descriptions of reality, it makes sense to see the rejection of an epistemically virtuous idea as an attempt to ignore some fact about the world. What is wrong with this picture is that scientific ideas are often not passive descriptions of reality but rather agents in creating order. When does science take on this active role? One area in which this role is most noticeable is when science attempts to classify and explain human behavior. As Dupré points out in his discussion of evolutionary psychology and rational choice theory, “…human behavior is not an immutable set of phenomena awaiting the correct scientific analysis, but is rather subject to constant historical evolution. And theorizing about human behavior is always to some extent an intervention in this evolution” (Dupré 1994, p. 378). Scientific ideas about humans can change the ways that humans act. It is, therefore, not correct to think of moral criticism as always advocating the rejection of some description of an immutable phenomenon. Moral criticism can rather advocate the rejection of one path of the evolution of human behavior in favor of another path.
A distinction introduced by Ian Hacking is useful in clarifying the relevance of this point to the case of classificatory schemes. Hacking distinguishes between two types of kinds, interactive kinds and indifferent kinds. Interactive kinds are classifications that “when known by people or by those around them, and put to work in institutions, change the way in which individuals experience themselves—and may even lead people to evolve their feelings and behavior in part because they are so classified” (Hacking 1999, p. 104). Heterosexuality is one example of such a kind. Individuals often go to extreme measures to control their speech and posture in order to behave in the manner befitting a straight person. Similarly, Gowaty’s criticism clearly points to the interactive nature of rape as a classification. The classification of a woman as a victim of rape in the legal institutions of our society can have significant impacts on the way that the woman behaves (e.g. if she knows that the legal system does not classify her as a victim, then she may not attempt to seek punishment of her attacker). In contrast, indifferent kinds are types of entities that do not evolve simply because they have been classified in a particular manner. To use Hacking’s example, quarks do not change in response to ways in which they are classified: “Our knowledge about quarks affects quarks, but not because they become aware of what we know, and act accordingly” (Hacking 1999, p. 105).
When a science engages in analysis of interactive kinds, its theorizing can have effects on the entities it studies. Sciences like economics that traffic in interactive kinds can, therefore, have a tremendous impact on the way in which humans will behave in the future. This means that it is often inappropriate to view the classificatory schemes of such sciences are merely descriptive of some order in the world. It may instead be the case that they are also actively engaged in creating that type of order.5 In such cases, it is inappropriate to view moral criticisms as attempts to turn a blind eye to unpleasant facts. When scientific classifications of interactive kinds act not as descriptions of human behavior but rather as prescriptions that direct the evolution of behavior, they are no longer simply cold, hard facts that any intellectually honest person must recognize. It is, therefore, appropriate for moral critics to attack such prescriptions on the grounds that they will lead to morally objectionable situations. This means that moral criticisms of disciplines that engage in theorizing about interactive kinds can be defended against objection C.6
In this section, I have attempted to defend moral criticisms against three pressing objections. I have not attempted to provide an exhaustive defense of moral criticism of science in this paper. There may be many more objections to moral criticism that I have not considered. For example, I have not discussed the viability of methodological objections to moral criticism. Future analyses of moral criticism would have to attend to such concerns. At this point, I hope simply to have shown that Dupré’s proposal to use moral criticism is one that is not so easily dismissed out of hand.
In conclusion, John Dupré’s brand of pluralism is not a version of anything-goes relativism. Dupré proposes two types of standards against which science can be measured: epistemic virtues and moral/political virtues. I have presented an example of how Dupré could put these virtues to use in evaluating classificatory schemes. This extended examination of the controversy surrounding anthropomorphic use of language in the study of rape shows that critics do use both epistemic and moral critiques to evaluate anthropomorphic science. I have also defended moral criticisms from three objections. A number of conclusions can be drawn from this defense. First, in order to be effective there are some guidelines that moral criticisms should follow. Second, moral criticisms of disciplines that theorize interactive kinds can be defended from the types of objections discussed here. While much more work needs to be done to justify the use of moral criticism, I do hope to have shown that Dupré’s means for avoiding relativism can be defended from some pressing objections.
Abele, L. and Gilchrist, S. (1977), “Homosexual rape and sexual selection in Acanthocephalan worms”, Science 197: 81-83.
Barash, D. (1977), “Sociobiology of Rape in Mallards (Anas platyrhynchos): Responses of the Mated Male”, Science 197: 788-789.
Bastock, M. (1967), Courtship: An Ethological Study. Chicago: Aldine.
Dupré, J. (1993), The Disorder of Things: Metaphysical Foundations of the Disunity of Science. Cambridge: Harvard University Press.
Dupré, J. (1994), “Against Scientific Imperialism”, PSA 2: 374-381.
Dupré, J. (1998), “Against Reductionist Explanations of Human Behavior”, The Aristotelian Society Supplementary Volume LXXII: 153-171.
Ereshefsky, M. (1992), “Eliminative Pluralism”, Philosophy of Science 59: 671-90.
Estep, D. and Bruce, K. (1981), “The Concept of Rape in Non-humans: A Critique”, Animal Behaviour 29: 1272-3.
Fausto-Sterling, A. (1985), Myths of Gender. New York: Basic Books Inc..
Gowaty, P. (1982), “Sexual terms in sociobiology: Emotionally evocative and paradoxically, jargon”, Animal Behavior 30: 630-631.
Hacking, I. (1999), The Social Construction of What?. Cambridge: Harvard University Press.
Hull, D. (1990), “Conceptual Selection”, Philosophical Studies 60: 77-87.
Hull, D. (1994), “The Disorder of Things: Metaphysical Foundations of the Disunity of Science”, The Philosophical Review 103: 701-705.
Kitcher, P. (1984), “Species”, Philosophy of Science 51: 308-33.
Kitcher, P. (1985), Vaulting Ambition: Sociobiology and the Quest for Human Nature. Cambridge: MIT Press.
Kitcher, P. (1989), “Some Puzzles about Species”, in M. Ruse, (ed.), What The Philosophy of Biology Is. Dordrecht: Kluwer Academic.
Longino, H. (1990), Science as Social Knowledge. Princeton: Princeton University Press.
Mitchell, S. (1996), “The Units of Behavior in Evolutionary Explanations”, in M. Bekoff and D. Jamieson, (eds.), Readings in Animal Cognition. Cambridge: MIT Press.
Thornhill, R. (MS), “Sexual selection and the evolution of courtship”.
Thornhill, R. (1980), “Rape in Panorpa Scorpionflies and a General Rape Hypothesis”, Animal Behavior 28:52-59.
Thornhill, R. and Thornhill, N. (1983), “Human Rape: An Evolutionary Analysis”, Ethology and Sociobiology 4:137-173.
Thornhill, R. and Palmer, C. (2000), A Natural History of Rape: Biological Bases of Sexual Coercion. Cambridge: MIT Press.
1 It is worth noting that Thornhill’s later definitions broaden the definition of rape to include acts of copulation that do not necessarily include insemination. For example, the 1983 paper defines rape as “forced copulation of a female by a male” (Thornhill & Thornhill 1983, p. 140). That said, I will focus on the 1980 paper and responses to it.
2 One might object here that Thornhill would call forced intercourse with a post-menopausal woman rape because he begins the ’80 paper with the following definition “Rape is forced insemination or fertilization” (Thornhill 1980, p. 52). Thus, one might argue that when it comes to determining whether or not a woman has been raped, the major consideration is her psychological state at the time of the attack. On this position, if the post-menopausal woman says she was unwilling and was forced, then the scenario describes a rape. However, this response is not one that Thornhill would likely take, for the following reason. As was noted earlier, Thornhill claims that an act that is reproductively disadvantageous for the female is one that she is “unwilling” to engage in. It is this reasoning that enables him to apply his definition of “rape” to scorpionflies and other non-human animals with which the problem of ascertaining the “will” of an individual is particularly pressing. In other words, the only way for Thornhill to make his definition apply to both humans and non-humans is for him to interpret an act as forced when it is against the reproductive interests of the female. He cannot, therefore, simply take the woman’s account of her intention as the sole criterion on which to judge whether the intercourse was forced. He must also consider her reproductive interests. Hence, in order to save the universality of his definition of rape, Thornhill cannot claim that the scenario describes an act of rape simply because the woman’s will was opposed to the intercourse.
3 Gowaty also uses the example of the question “Is it rape when a virgin is forced to intercourse?” However, this question does not reveal the clear difference between the ordinary and sociobiological classifications that Gowaty needs to make her point. According to Gowaty, the sociobiologist would not call this rape, contrary to common usage, because “the virgin could be impregnated and her genetic fitness thereby increased” (Gowaty 1982, p.630). However, Gowaty overlooks the fact that Thornhill introduces several possible ways in which a female’s fitness can be reduced as the result of a rape. Thornhill could appeal to several of these causes of decreased fitness to claim that the virgin would have been raped. For example, he says, “A human male knowing or suspecting that his mate has been inseminated by another male might also reduce paternal investment and thereby decrease his mate’s fitness” (Thornhill 1980, p. 57). Thornhill might, therefore, argue that the raped virgin would suffer a decrease in fitness due to the reaction of her partner to the attack. I raise this issue not to argue that Thornhill would be justified in counting this scenario as rape under his definition, but rather to show one way in which Gowaty oversimplifies Thornhill’s position. Her example of the post-menopausal woman is a much clearer example of a discrepancy between ordinary and sociobiological classification.
4 While I am framing this defense of moral criticism in terms of a discussion of Dupré’s promiscuous realism, I do not mean to suggest that the promiscuous realist faces any unique challenges in trying to use moral evaluation. All of the objections I consider in this section could easily be applied to any attempt to reject scientific ideas on moral grounds.
5 I am drawing here on an argument Dupré makes in the final chapter. He says, “Where there are background assumptions that affect, but are not determined by, the empirical content of a discipline, we should see that discipline as engaged, at least in part, in constructing a kind of order that is defined by that set of assumptions” (Dupré 1993, p. 260). He identifies economics as one such discipline. In this passage, Dupré is clearly pointing to the same phenomenon in which I am interested, namely the role of science in creating or maintaining regularities. However, he identifies these disciplines that engage in such construction as those that rely on background assumptions that go beyond the empirical content of the field. It is not clear to me that this description adequately captures the type of science we have in mind. As I see it, all sciences must rely, to some extent, on such background assumptions. For example, to believe that Kepler’s laws are true, one must assume that regularities observed in the past will continue in the future. This is a background assumption that, as Hume pointed out and many philosophers have since repeated, outstrips the observed data of any science. I do not think anyone would like to claim that celestial mechanics constructs some kind of order in the heavens. Therefore, I fail to understand why Dupré thinks that background assumptions are the key element of disciplines that construct order. My use of Hacking’s distinction between interactive and indifferent kinds is intended to better distinguish between those sciences that create order and those that simply describe it.
6 It should be noted that I have not taken a position with respect to the legitimacy of moral criticism of disciplines that theorize indifferent kinds. I have only argued that being a discipline that studies interactive kinds is a sufficient condition for being liable to moral criticism. I have not argued for or against the claim that it is a necessary condition.