Over on OrgTheory, a discussion of the apparent constancy of color perceptions morphed into a(nother) discussion of performativity and, by inappropriate extension, postmodernism and epistemological skepticism. Rather than hijack that post, I’m moving over here to post some thoughts and critique of Teppo Felin and Nicolai Foss’s paper, “Social Reality, the Boundaries of Self-Fulfilling Prophecy, and Economics.”
Felin and Foss (hereafter, FF) present a critique of performativity theory, specifically Ferraro, Pfeffer, and Sutton’s (hereafter, FPS) discussion of the mechanisms of performativity. Unfortunately, FPS claim economics is performative and bad, so FF argue by contrast that it is natural and good. I won’t address the bad/good question, only the performative/natural question.
The crux of FF’s critique is on page 656:
A natural extension of FPS, and related strong forms of the argument is that any (even false) reality can be created through theory, language, and prophecy, which subsequently fulfills itself.
But this extension is anything but “natural.” In fact, the claims in FPS and the literature on which it builds emphasize mechanisms by which theories, including false ones, can affect reality. These mechanisms are key, but virtually entirely ignored in FF in favor of characteristics of the theory itself. Essentially, FPS claim that false theories have effects when they are widely believed to be true and encoded in institutional and technological innovations (“agencements“). The “natural extension,” then, is a straw man, since there is nowhere a claim that any false theory can affect reality. I cannot theorize that the tree outside my window is made of formaggio pecorino tartufo, convince my gullible graduate students of this, and go outside to begin a delicious feast. But the fact that I cannot do this says nothing about whether convincing, institutionally-endorsed false theories of behavior can affect behavior systematically and path-dependently.
FF’s alternative hypothesis is that “Theories…affect reality because they capture and explain underlying objective realities….” In other words, true theories affect reality; false theories don’t. But pause for a moment to consider: what constitutes a true theory? It is one that “capture[s] and explain[s] underlying objective reality.” Since true theories are claimed to be those that change reality, a theory that is true at T1, prior to its promulgation at T2, would no longer be true at T3, after that promulgation, since the reality it “capture[d] and explain[ed]” at T1 was changed by the trueness of the theory!
By contrast, consider a theory that, at T1, is false. At T2, it is widely promulgated, making it widely believed and institutionally/technologically instantiated. At T3 it could quite plausibly be more true than it was at T1 because, through the mechanisms described, the described behavior changed to match the believed and endorsed theory.
This comparison demonstrates that, contrary to FF’s contention, false-but-convincing-and-instantiated theories are more likely to change reality and become true than are true theories, since true theories that affect reality tend to become false!
It’s important to pay close attention to the mechanisms here. FF offer objective reality and human nature as the scope conditions on the performativity of theory; FPS, and the broader performativity literature, actually offer their own scope conditions in the form of mechanisms that must be in place for a theory to become performative. They must be widely believed and instantiated in technological and institutional arrangements. Unlike FF’s contention that truer theories are more effective–for which no evidence is actually presented–FPS stand on the widely-documented case of the Black-Scholes-Merton (BSM) equation: a theory that was relatively untrue at T1, was widely promulgated and convincing and instantiated in technical practices at T2, became truer at T3. FF simply claim, sans evidence, that
the underlying realities that the [BSM] model tapped into better explained a more true value of options
what evidence do we have that the model was “more true” than others? Well, the fact that it convinced enough people, and was widely enough adopted, that it “helped shape and change option prices themselves.” Truth here, in other words, has nothing to do with representing or capturing observable reality; it is an article of faith that the model must have reflected a deeper reality since it turned out to be effective. In other words: at T1, when the BSM model was developed, there was no way to evaluate whether it was true in the way FF understand it to be true.
By now it should be clear that this is a tautology. What theories affect reality? Those that are true. How do we know if they are true? Because they affect reality.
Part of the objection FF raise to FPS has to do with “mindlessness.” FPS (21) claim that:
theories become dominant when their language is widely and mindlessly used and their assumptions become accepted and normatively valued, regardless of their empirical validity.
From this FF argue that FPS think “humans are ‘mindless’ and readily duped into false realities” (658). But there is no claim that humans are mindless in general, or that the duping is done “readily,” just that it is possible, and that when it happens, the conditions are ripe for the theory they are duped into to become dominant. FF essentially commit a macro version of the ad hominem fallacy: drawing a conclusion about human character from a claim about a particular human behavior.
It is certainly the case that there are important and strong scope conditions to performativity. Indeed, I’d go so far as to say that there aren’t that many cases of truly performative theories, meaning theories that make themselves true. I think there are lots of theories that meet the lower standard of reactivity: that is, affecting the reality they conceptualize. I’m convinced that theories of public opinion are performative in a real sense, and my forthcoming article with Kate McFarland in the Annual Review of Sociology makes this case.
Certainly, the project of understanding the appropriate scope conditions of performativity (and reactivity) is an important one. However, I expect the project to be more fruitful when approached through the specification of mechanisms and necessary mediating conditions than when approached through the ontological assumptions of objective reality and human nature.