Have the social sciences become stagnant? Are they even sciences? What can be done to make them more science-like? These questions have preoccupied scholars for quite some time. In fact, these are typical test questions for graduate students taking university research methodology courses. Three articles have just been published that offer some new insight. Let’s take a look at the latest round of this debate.

Irving Louis Horowitz (2006) has written down some emeritus reflections about the decline of the “big five” social sciences (anthropology, sociology, political science, economics, and psychology) and the rise of five little ones (communications, environmental studies, urban affairs, criminology-penology, and policy studies). The social sciences have lost influence and are failing to attract scholars because of declining relevance. Soon after the big disciplines were established, it became difficult to confine analysis to these pigeon holes. This led to fragmentation of the social sciences into narrow, “hyphenated” sub-disciplines (e.g., political anthropology and social psychology). Those who sought to create a truly interdisciplinary and unified science were then treated as fringe players (the equivalent of the much-maligned practitioners of holistic medicine). In contrast, the five little disciplines show promise precisely because they are reversing this development. These are disciplines that are focused on tangible societal needs through the integration of knowledge across traditional disciplinary boundaries. For example, communications studies draw knowledge from across the social sciences, while environmental studies look at subjects in an integrated way from different levels of analysis (from the microscopic to the macroscopic levels). Horowitz overstates this as a “new revolution” but, given my interdisciplinary sympathies, I’m glad he has chimed in for the record.

According to Horowitz, many within the social scientific establishment do not want to admit to this decline of relevance. Liah Greenfeld’s article (2005) echoes Horowitz’s by arguing that many social scientific disciplines have become exhausted and unable to support new careers. They have become unscientific and laden with ideology. She singles out macro-sociology and those branches of history based on economic determinism.

Greenfeld discusses nationalism studies to illustrate her point. A couple of books (Gellner’s Nations and Nationalism and Anderson’s Imagined Communities) inspired many scholars to enter this field of study. The problem is that the main theses of both books are factually wrong, something which should have been obvious to anyone with the slightest grasp of the subject. Ideological presuppositions trumped empirical observation. Within the natural sciences, these errors would have been pointed out and subsequent studies would offer corrections. This self-policing is a necessary part of the scientific method. But in the social sciences these errors were allowed to stand and other scholars build upon these arguments despite overwhelming evidence showing them to be flawed. Well said.

Greenfeld makes the case for non-ideological social sciences. Yet, in doing so, she argues against aping physics, which many consider to be the archetypical scientific discipline. Social scientific subject matter is too historically contingent and variable for the establishment of universal laws. Experimentation and quantification are rarely viable, according to Greenfeld. Thus, Karl Popper’s (1972) model of scientific “conjecture and refutation” is less applicable. Biology is a better model for the social sciences because it developed new methods appropriate to its subjects. Life is difficult to predict. Biology relies more heavily in probabilistic methods instead of universal laws. More effort should be spent conducting systematic comparisons than conducting experiments. The social sciences should do likewise, says Greenfeld.

S. Phineas Upham (2005) offers a very different view. He claims that social scientific attempts to create reductionist models are unfairly pooh-poohed because of a misunderstanding of the role of model-making in science generally. Most of Upham’s argument comes from Nancy Cartwright (1983, 1999), who argues that reality is too unpredictable and complicated to be exactly mirrored by scientific models. Moreover, scientists often experiment by isolating phenomena in order to determine causality. The resulting models are not perfect predictors outside of the laboratory because there are too many intervening forces at play. (For example, there is no such thing as a perfect vacuum or a frictionless plane.) Thus, even scientific models in a discipline like physics are only approximations (simulacra) which do not account for all factors within the real world. So when people attack a social scientific discipline like economics because its models seem unrealistic, they are missing the point of scientific modelling. Accordingly, when economists make a claim and then invoke the ceteris paribus disclaimer (a thing is true only insofar as all other things are held equal), they are not trying to shield themselves from the scientific requirement that a claim be falsifiable (potentially refuted through empirical observation or testing). Instead, they are simply acknowledging the limited applicability of a model within a messy and complicated reality.

These are all interesting contributions to the debate. However, there is much to disagree with here.

I agree with Upham’s (or should I say Cartwright’s) argument that scientific models are an imperfect reflection of reality. Models are a form of simplification whereby important factors are isolated. Applying scientific models in practice means taking into account a variety of other intervening factors. Most scientists and engineers understand this intuitively, especially those who use science to build new technologies given the practical necessity of coping with various exogenous factors. In the case of economics—or rather Upham’s more specific case of rational choice theory—many of the main assumptions of the model would seem to be obviously wrong in most situations (e.g., people always act according to purely rational action, highly selfish utilitarianism, and perfect information). This is not simply a case of inaccurate modelling or simulation, but of using assumptions that can be shown empirically to be false (in most controlled and real-world contexts) and that seem unrealistic on the face of it. Upham admits that some of these “obviously false postulates” would have to be “weakened”, which is a weasely way of putting it. He then goes on to explain that the unrealistic postulates can not be completely abandoned because that would undermine the coherence of the model, which is the way most hack economists would put it. It seems clear to me that the imperfections of non-empirical, rational-choice models are not analogous to imperfections in empirical physics models, such as the law of gravity. Yet this is exactly what Upham claims (repeatedly) because his main argument hinges on it.

The case of rational choice theory supports Greenfeld’s argument that social scientific theories that are based on economic determinism and are obviously wrong often do not get rejected.1 They live on and scholars just add to them. The pile gets taller. And there is certainly an ideological appeal to rational choice theory, for it underpins such things as the Thatcherite philosophy of public administration.

Although I do like Greenfeld’s article, I do have a concern about her misrepresentation of the role of experiments and measurement. She understates biologists use of experiments. What scientific discipline does the terms “lab rat” and “guinea pig” conjure up in your mind? She also understates biologists use of measurement. Doesn’t systematic comparison of species mean measuring gills, wings, beaks, and flippers? In other words, how many biologist attempts to classify species are premised on quantification of the physical attributes of organisms? Quite a few, I would argue. Aren’t cell counts, blood-chemical levels, gene-sequencing algorithms, life spans, reproduction rates (et cetera, et cetera) all forms of quantification? Certainly biologists are more timorous about stipulating concrete scientific laws but, as the fractious debates among evolutionary biologists illustrates, it isn’t like they aren’t trying. It almost seems as though Greenfeld has in mind a stereotype of biologists as 18th Century naturalists, replete with sketch pads, butterfly nets, safari hats, and magnifying glasses. This is interesting because modern-day taxonomists remain on the periphery of the biological sciences precisely because they conform to Greenfeld’s description.

On the other side of the curtain, Greenfeld overstates the difficulties of using experimentation and measurement in the social sciences. Experiments are often the basis for more realistic behavioural postulates in fields like economics. Indeed, the emerging field of behavioural economics is premised on finding out how individuals actually make decisions in the real world. (It’s about time!) Greenfeld simply dismisses this with a backhand and a sidestep. The backhanded slap is this: “The substitution of ‘behavioral’ for ‘social’ in the name of these disciplines, obviously, changes nothing.” (p. 110) Well … obviously … it changes quite a bit if the new term indicates a more diligent use of experimentation and other empirical methods to determine the actual activities of humans. The sidestep is Greenfeld’s insistence that all attempts to explain human behaviour are the exclusive domain of biologists. She claims that social scientists should only concern themselves with the study of culture. Ah, yes: arbitrarily defining yourself out of a debate; I always get suckered by that kung-fu move. Or more pointedly: she seems to advocate creating an iron wall within the nature-nurture debate; scientists get one side, social scientists get the other, and never the twain shall meet. [sigh]

While on the topic of behavioural economics, let me mention that it offers a counter-point to Horowitz. Economics is one of his five big disciplines and I certainly agree that mainstream economics has become stagnant. Indeed, there is an entire genre of books criticizing the lack of relevance and empiricism in many branches of mainstream economics.2 This is not so with behavioural economics, as suggested by the success of several recent authors.3 This field also meets the criteria listed by Horowitz: it is focused on actual problems; it draws extensively from other disciplines (notably biology, psychology, and politics); it is possible to integrate findings across levels of analysis; and so forth. Thus, behavioural economics provides an example of what a more unified approach to social science might look like.

So is there a verdict about what ails the social sciences? Is it a lack of social relevance and scientific unity (Horowitz)? Or is it too much ideology and not enough empiricism using tailored scientific methods (Greenfeld)? Or do people just misunderstand the limits of science (Upham)? Expressed in these general terms, I would have to say all three arguments have merit. I share many of the misgivings of these authors. But these questions will still have to be asked until an entirely satisfactory answer arrives. This debate is to be continued. Don’t seem so shocked by this conclusion.

Review by Peter Stoyko

NOTES

1. Despite my criticisms of this family of theories, I should point out that I do find merit in some of them—those few which have some support of empirical evidence. For an excellent catalogue of one variant of rational choice theory, with empirical commentary added, see: Dennis C. Mueller, Public Choice Theory III (Cambridge: Cambridge University Press, 2006).
2. To name just two examples: David C. Colander, Why Aren’t Economists as Important as Garbagemen: Essays on the State of Economics (Armonk: M.E. Sharpe, 1991); Paul Ormerod, The Death of Economics (Hoboken: Wiley, 1997).
3. To name just two examples: Steven D. Levitt and Stephen J. Dubner, Freakonomics: A Rogue Economist Explores the Hidden Side of Everything (New York: William Morrow, 2005); James Surowiecki, The Wisdom of Crowds: Why the Many Are Smarter Than the Few and How Collective Wisdom Shapes Business, Economies, Societies and Nations (New York: Doubleday, 2004).

Review by Peter Stoyko

REFERENCES