Discussion on ‘Performativity: will any theory do?’
June 4, 2008
As usual, Peter Levin’s comments are illuminating and thought provoking, and this one is not different. So, I thought that it would be better to put it here as new post (with my reaction), rather than just as a comment
Interesting post, addressing a concern I think about all the time. But might I suggest that there are alternatives as well. You note that if the theory ‘fails’, or if it has no ‘public experiment’ available to it, there is no possibility that it will be performative – at least in the sense you mean.
I was thinking about the cases where the theory remains, despite evidence to the contrary. My art specialists are as likely to speak of a price as ‘not yet achieving’ its market price, or that its market price is not yet in line with its cultural value, or something like that. In other words, out-of-line prices are explained not by a failure of theory (in this case, theory is that mess of valuation done by specialists), but by a failure or anomaly in the marketplace.
A second case recalls to mind that someplace (I can’t seem to recall it now) I remember a study asking whether economists were judged based on their predictive abilities. This is interesting because of economics’ claim to at least aim for prediction. I recall, though, that prediction was less important than providing ‘interesting explanation’ or somesuch. I read this, again, as some kind of bending of the world to match what theorists do, rather than the reverse.
I realize these are not exact counter-factuals, but I’m interested in your thinking about how wide a range of possibilities there might be. The idea that there is a reality out there, and that theories ‘work’ or ‘don’t work’ in explaining-then-shaping those realities, seems a bit constraining.
About situations of ‘price not achieved yet’. Yes, we see such situations frequently in markets, when the public experiment is not delivering the results one expects them to deliver. In these cases the burden of proof is transferred elsewhere. That is, in the case of the art specialists they will need to explain why it is that the price they predict for that piece is not achieved in the market. Another example, where such burden of proof is really pressing, is the example of convergence arbitrage. There, a trader’s position is dependent on asset prices in two markets to converge, but they may diverge. In such cases, it is not uncommon for the trader to ‘defend the position’, either successfully or unsuccessfully, against his/her manager who may want to unwind it There is a very good paper about this by two finance scholars (“The Limits of Arbitrage” by Andrei Schleifer & Robert W. Vishny. The Journal of Finance, 1997). So, you are right, a one-time discrepancy between prediction and market would not fail the theory, but gradually, if such incidents mount and if these incidents cannot be defended successfully, then we should expect the performativity circle to be broken in some way.
About the second case: economists who simply produce ‘interesting explanations’. In the cases where experts do not take part in the prediction-performativity game then no public experiment (at least in the restrictive sense I am suggesting here) exist. But, of course, the claim you mention is interesting with regard to expanding the concept of the public experiment to the (relevant) realms of aesthetics of theories, their appeal to policy and so on.
About ‘reality out there’: the public experiment concept is agnostic about the existence of external reality. We follow the agents here. As long as the agents find the predictions useful in their interventions, the theory will be supported.