Kuhn, Scientific Revolutions, and the Social Sciences

by Justin Esarey

The New Atlantis published a recent retrospective on Kuhn's Structure of Scientific Revolutions that's well worth a look. I think the article rightly notes that social scientists have eagerly embraced Kuhn's ideas, in the sense that the average political scientist is quite likely to have at least spent a graduate seminar meeting or two discussing the relationship between Kuhn's thesis and the practice of social science.

Yet I can't help but think that it gets several important points wrong. For instance:

Despite these criticisms, many social scientists embraced — or perhaps appropriated — Kuhn’s thesis. It enabled them to elevate the status of their work. The social sciences could never hope to meet the high standards of empirical experimentation and verifiability that the influential school of thought called positivism demanded of the sciences. But Kuhn proposed a different standard, by which science is actually defined by a shared commitment among scientists to a paradigm wherein they refine and apply their theories. Although Kuhn himself denied the social sciences the status of paradigmatic science because of their lack of consensus on a dominant paradigm, social scientists argued that his thesis could still apply to each of those competing paradigms individually. This allowed social scientists to claim that their work was scientific in much the way Kuhn described physics to be.

I'm willing to be persuaded otherwise, but my subjective impression is that Popper has been far more influential on (quantitative) political science than Kuhn. It's Popperian falsificationism that informs the scientific process that we teach to our undergrads in the standard research methods curriculum. Popperian falsificationism drove the AJPS's previous policy of requiring explicit statements of hypothesis, test, and result in the abstract. Popperian falsificationism informs criticisms of non-experimental empirical social science offered by Green, inter alia.

Some things the author says are kinda true, but I think are taken in a misleading direction:

A scientific way of thinking permeated the writings of Auguste Comte and Karl Marx, and by the end of the century, with the work of Max Weber and Émile Durkheim, the era of social science had begun in earnest. Many of the early social scientists came to view society in terms of contemporary physics; they adopted the Enlightenment belief in science as the source of progress, and considered physics the archetypical science. They understood society as a mechanism that could be engineered and adjusted. These early social scientists began to deem philosophical questions irrelevant or even inappropriate to their work, which instead became about how the mechanism of society operated and how it could be fixed. The preeminence of physics and mechanistic thinking was passed down through generations of social scientists, with qualitative characterization considered far less valuable and less “scientific” than quantitative investigations. Major social scientific theories, from behaviorism to functionalism to constructivism and beyond, tacitly think of man and society as machines and systems.

I'm not an expert on Weber and Durkheim, but I have read some of their methodological writings. Durkheim wrote an entire book about the philosophical underpinnings of his method. Weber's method of verstehen, interpretive understanding, launched a thousand essays about qualitative interpretationalism in political science–most of them decidedly opposed to positivism in the social sciences. That's not to say that many quantitative political scientists don't cling to a somewhat philosophically backward version of epistemological positivism; I suspect that many, maybe even most, do. But methodologists, the teachers and developers of quantitative methods in political scientists, are ceaselessly ragging on people to take a more sophisticated view (just take a look at this naturalistic but non-logical-positivist book just released by Kevin Clarke).

Still, some of the things the author says are pretty interesting. For instance:

A recent paper in the journal Theory in Biosciences perfectly encapsulates the desire for a more biological perspective in the social sciences, arguing for “Taking Evolution Seriously in Political Science.” The paper outlines the deterministic dangers in the view of social systems as Newtonian machines, as well as the problems posed by the reductionist belief that elements of social systems can be catalogued and analyzed. By contrast, the paper argues that approaching social sciences from an evolutionary perspective is more appropriate philosophically, as well as more effective for scientific explanation. This approach allows us to examine the dynamic nature of social changes and to explain more consistently which phenomena last, which disappear, and which are modified, while still confronting persistent questions, such as why particular institutions change.

Reading over a preprint of the article, it makes some interesting (if slightly superficial) observations about the difference between an explanandum that's amenable to repeatable experimentation and one which is data-driven and factual but largely based on uncontrolled observational evidence. The latter is not unscientific, but certainly not idiographic (a new word I learned from this month's Perspectives on Politics). But I'd like to hear a more thorough and more philosophically developed epistemological framework that relates the practice of evolutionary biologists to some foundational perspective on the world in such a way that we would expect one to lead to a better understanding of the other. That would, I think, be necessary before I could get behind using evolutionary biologists as a model for political science.

The Theory in Biosciences article does say:

Many political scientists today are searching for a better understanding of the mechanisms of political change. The problem analytically, is that most political science models are static. For rational choice, this is due to the theoretical argument that any given institutional setting will eventually reach an equilibrium in which “no one has the incentive to change his or her choice” (Levi 1997: 27). Consequently the only source of change is exogenous. As Levi argues, “it is obvious that choices change regularly and constantly. . . To understand these changes requires a set of hypotheses concerning what exogenous shocks or alterations to the independent variables will have what effects on the actions of the individuals under study” (Levi, 1997: 28).39 Given the foundational assumptions and logic of rational choice, “endogenous institutional change appears,” as Hall and Taylor observe, “to be a contradiction in terms.”

Now that's something I completely agree with… and yet, I see very little research being done on truly dynamic theoretical models in political science. Dynamic statistical models, sure. But they can't substitute for institutional political theories that seamlessly integrate change over time into their explanatory framework.