Helen Longino, Studying Human Behavior: How Scientists Investigate Aggression and Sexuality, University of Chicago Press, 2013
Rebecca Kukla
Kennedy Institute of Ethics
Georgetown University
Washington, DC, USA
In Studying Human Behavior: How Scientists Investigate Aggression and Sexuality, Helen Longino meticulously examines a wide variety of research programs devoted to studying human behavior, specifically aggression and sexual orientation. She teases apart the methodologies of the various approaches, examining their epistemic structure, the assumptions they must hold fixed, the research questions they enable and those they necessarily occlude, their underlying ontologies, the interests that motivate them, the ideological investments they help buttress, and the practical and political uses they serve. By the end of the book she has made clear that the so-called ‘nature-nurture’ debates invoke a hopelessly simplistic and distorting dualism when it comes to any interesting question about the etiology of a behavior. This is a readable, informative, rigorous, and elegantly organized new work from a leading scholar who has had a major hand in turning the interplay among science, ethics, and social interests into a topic of rigorous philosophical study.
In the first part of the book, Longino devotes a chapter or substantial section to each major methodological approach to studying aggression and sexual orientation, including (using her labels): quantitative behavioral genetics, social-environmental approaches, molecular behavioral genetics, neurobiological approaches, and several ‘integrative’ approaches (such as developmental systems theory) that focus on the interactions and dynamics between causal factors over time rather than trying to isolate one causal contribution. She also briefly discusses human ecology, which takes the unit of study to be the behavior of a population, studied in its material environment, rather than the behavior of an individual.
Longino shows elegantly how each of these research programs is equipped to reveal certain specific kinds of truths about behavior, while making other important dimensions of the phenomena invisible. Many of her examples are technical and detailed, but I can recap a few relatively simple and powerful examples: (1) Behavioral researchers have no shared understanding of what an ‘environment’ is or which features of it we need to measure. Whatever operationalization of ‘environment’ is used within a study or set of studies, this operationalization is not itself scientifically validated, but merely presupposed (33). Thus how much of a behavior shows up in a study as due to the ‘environment’ depends on just what we measure. On the behavioral genetic approach this will include everything other than genes, but on, say, the neurobiological approach this will be everything other than neural structure. But these are quite different senses of ‘environment’ with different scopes, and there’s no good reason to think they are easily commensurable. (2) ‘GxExN’ researchers use techniques to study the etiology of behavior that were developed expressly for studying disorders (112). This work thus presupposes that we can understand whatever behavior is under study on the disorder model, and will yield no well-formed results to the extent that this model does not fit. But it is at best non-obvious that a disorder model is unproblematic when it comes to aggression or—especially!—sexual orientation. (3) Most of the research programs Longino covers treat the behaving individual as the unit of analysis and study, as opposed to, for example, human ecologists’ focus on whole populations. But as she demonstrates, this affects what kinds of causes we can study, because patterns at the population level may not be detectable using the kinds of measures that can be manipulated at the individual level. Furthermore, some kinds of behaviors may require more than one participant to exist at all; for instance, the very same thing that counts as an aggressive act within one interaction may be a peaceful act in another (175).
Longino does a lovely job of showing what sorts of social interests are served and buttressed by which kinds of research. For example, social-environmental approaches will, in virtue of their methodology, focus on social factors that are manipulable, and this makes them particularly amenable to generating suggestions for social interventions. Such research thus better serves the ends of those who are looking to implement such social interventions. This is so, not in virtue of the science yielding results that are more scientifically legitimate or more biased than its counterparts, but because this program is methodologically set up to display possible social interventions in a way that other research programs are not.
In the second part of the book, Longino develops an argument for a pluralistic approach to behavioral research that neither privileges one methodological research program nor tries to unify all of them. She also examines the concept of behavior itself, arguing that there exists neither an obvious, intuitive definition of behavior, nor any neat unity among the objects of study in these different research programs. She ends by analyzing how the results of the different research programs circulate and disseminate; she argues that there is very little cross-fertilization between programs and that behavioral genetic approaches are especially appealing to a broader lay audience.
One of the philosophically richest and deepest chapters of the book is Chapter 9, “Defining Behavior.” In it, Longino shows just how little clarity or unity there is to our scientific, lay, or philosophical conception of the ontology of behavior. In the first instance, we cannot presuppose that different research programs share an understanding of ‘aggression’ or ‘sexual orientation,’ and in fact they operationalize these notions quite differently and perhaps even incommensurably. All these research programs assume that “the definition and operationalizations of aggressive behavior and sexual orientation pick out a stable class of phenomena into whose etiology or about whose correlates it makes sense to inquire” (49). But it is by no means clear that there is some stable underlying phenomenon there to be studied; aggression and sexual orientation may be no more than strategic posits placed underneath amalgams of particular characteristics and actions.
Even more fundamentally, it’s not clear that we share an understanding of the ontology of ‘behavior’ itself: “By behavior . . . do we mean tendencies in a population, particular episodes in the history of an individual, patterns of behavior, dispositions to respond to situations in one way or another, patterns of interaction?” (113). She argues that in fact, “[t]he question about the proper ontology or the units of behavior does not have a unique, universal answer. The proper ontology depends on the question being asked. The questions are many and reflect a diversity of cognitive and practical interests” (177). Despite this diversity, she argues that most or all of the research programs in practice interpret ‘behavior’ to mean dispositions to behave in various ways (as opposed to, for instance, one-off actions). Even stronger, they in effect posit types of people with stable characters: homosexuals, aggressors, etc.
While I don’t have any definite points of disagreement with Longino’s analysis, I do think that she in various ways understates the significance of her own discussion in this chapter. It seems to me that the ethical and epistemological problems that her conclusions pose for behavioral research are thornier than she acknowledges.
For example, Longino repeatedly bemoans the dearth of cross-fertilization between the different research programs. But if it is true, as she argues, that there is neither a unified understanding of behavior that is common to these different programs, nor a single natural kind phenomenon that all of the programs are trying to capture, however imperfectly, then it seems that cross-fertilization would not only be difficult, but perhaps even undesirable. If we use the results from one study to critique or rethink another study, and if these two studies seem to have the same topic, but in fact share only a homonymous topic, then bad reasoning and bad science will likely ensue. That is, if ‘human aggression,’ say, means different things in the two studies, and if there is no stable phenomena underlying both programs, then an attempted synthesis of the two will confuse metrics and terms. So the lack of cross-fertilization is both less surprising than she makes it out to be, if the rest of what she says is right, and also less disappointing. Indeed, in this light her proposed ‘pluralism’ is a bit puzzling. While it makes sense to argue that each of these research programs has epistemic value and can uncover helpful truths, she sometimes slips into making it sound as though she is taking a pluralistic approach to studying some one thing—aggression or sexual orientation. Really, what she ought to be advocating, in order for her to be consistent, is the value of many different research programs with different but related topics.
I mentioned above that Longino argues that while there are many things that can count as ‘behavior,’ behavioral research focuses on dispositions to act, not on actions. Sometimes, I pointed out, she puts the point more strongly—and I think correctly—and argues that what really interests behavioral researchers are not particular actions, but rather the etiology and development of certain sorts of people: homosexuals, aggressors, criminals, and the like. In a quick footnote, she acknowledges Foucault’s point that research can help produce an association between clusters of behaviors and ‘special categories of persons,’ such as ‘the homosexual.’ Thus we cannot assume that such categories are natural kinds predating their scientific examination; they are instead contingent, culturally specific products. It is not automatic that clusters of particular actions are taken as outgrowths of having a certain kind of identity; rather, such identity types emerge, partly in tandem with research that presupposes them (153, n. 3).
Here too, though, I think she ought to push her own argument further than she does. For to study such kinds or identity types is not just to study ‘behavior,’ under a certain interpretation of the term. Such identities are not behaviors at all, under any interpretation, but rather the theoretical entities that purportedly express themselves in behaviors. When research implicitly or explicitly studies such identities, the behavioral operationalizations serve to allow purported scientific access to the underlying phenomenon; they are not the primary objects of study. If Longino is right that this is the scientific situation—and I find her case compelling—then this is a more radical conclusion than she acknowledges. Behavioral research does not, ultimately, take behavior as its object at all!
Furthermore, acknowledging Foucault’s point that these types are not natural kinds, but rather socially situated categories, has crucial epistemic and ethical consequences that Longino does not acknowledge. It is central to Foucault’s point that such human kinds are, as Ian Hacking puts it, “looping” kinds. Kinds like ‘homosexual’ and ‘aggressive personality disorder’ don’t just happen to be socially contingent. Rather, there is a dialectical relationship between the identification of the kind, especially in expert discourses such as behavioral science that seek to study it, and the solidification of the phenomenon under study into something unified and stable. Unlike electrons or pea plants, humans respond to discourse about them and the institutional practices that are entangled with such discourse. When scientists, policy makers, and other experts label a segment of the population as ‘aggressive’ or as ‘homosexual’ (or, for that matter, as ‘heterosexual’), they do not merely reflect how the world is, whether perfectly or imperfectly; they change the phenomena under study, by affecting the self-understanding and the behavior of the people to whom those labels are applied (among others). Asked to recognize themselves in an ever-tighter network of expectations, narratives, images, and symptoms, people thus categorized may come to form a unified and stable type that would not have otherwise existed.
But this means that behavioral research of this sort has concrete social and ethical implications beyond its direct effect on policy. Expert discourses have the power to solidify new kinds of people in the process of studying them. These are epistemic practices that help constitute their object of study; they have both word-world and world-word fit. Thus they come along with a certain ethical risk: creating and shaping a category of people such as, say, people who are aggressive (rather than just people who engage in specific acts of aggression) matters a great deal to how the people who fall under or struggle to avoid that categorization will in fact behave, experience themselves, and fit into institutional practices and mechanisms. Furthermore, when we take a kind of person as our object of study, we incur not just ethical but epistemological risk. We study only patterns of behavior that can be conceptualized as stemming from some sort of stable set of dispositions rooted in a personality type. This can occlude other equally revealing patterns that require other sorts of explanatory frameworks. For instance—to slightly repurpose one of Longino’s own most powerful examples—this framework keeps our focus on individual behavior rather than interactions requiring more than one participant or population-level behavior. This focus is not epistemologically neutral, but rather shapes the field of possible research questions and conclusions.
I will close with some minor worries. The book ends with Longino’s description and analysis of her study of dissemination patterns in behavioral research. She examined the citation paths of papers representing the various research programs she examined, to see how much they were cited in general, how much they were cited by other papers in the same research program and by papers from other programs (indicating cross-fertilization), as well as how often they were taken up by policy journals and in the mass media. Her main conclusions are that there is very little cross-fertilization, and that work in behavioral genetics has a disproportionate mass appeal and gets cited more in general.
It is a bit hard to know what to make of this study. I’ve already discussed a good reason why cross-fertilization might be rare. Furthermore, researchers are likely under pressure to cite key texts in their own subfields in a way that they are not under pressure to cite other sorts of texts. Indeed, a wide array of pressures, including the editorial practices of journals, researchers’ desires to impress certain people or fit certain grant opportunity descriptions, and much more, surely influences these citation patterns. Longino would undoubtedly acknowledge this point; she emphasizes that her study is merely descriptive. She is tracing paths of influence, not explaining or hypothesizing the reasons for these results. But in the face of such vast complexity, I wasn’t sure what I was supposed to conclude from the patterns. In particular, I was not sure I could conclude that they pointed to a problem with the dearth of cross-fertilization.
If these patterns of influence were clearly indicative of the social impact of the different research programs, then regardless of the reasons for the patterns this would be important information. Sometimes this seems to be Longino’s suggestion. Certainly she promises throughout that she will be examining the policy uptake of the different research programs. But this part of the book, unlike most of the rest, is disappointingly thin. I’m just not convinced that citation in policy journals is an interesting proxy for actual policy impact, given the massively complex and messy reasons why journal articles cite what they cite, and given the gap between scholarship and practice.
My goal in the last few paragraphs has merely been to explore what we can conclude from Longino’s analysis. I’ve claimed that in places she stops short of drawing some strong conclusions that are available to her, and in other places I wished she could have concluded more. These are minor critical reflections, and my comments are not intended to detract substantially from the value of her analysis. Helen Longino’s Studying Human Behavior is, as I hope I have indicated throughout, a thoroughly excellent book. It is important and enjoyable reading for anyone working in bioethics (especially research ethics), science studies, social epistemology, philosophy of the social or life sciences, or the ethics of science.
Rebecca Kukla
Kennedy Institute of Ethics
Georgetown University
Washington, DC, USA