Popper, Falsification, and Social Theory 7 June 2006Posted by Todd in Philosophy & Social Theory, Philosophy of Science, Queer Theory.
Theory and Reality is blowing my mind and making me stop and think about things I've taken for granted since high school. I've always been taught that scientific theories can never be proven true, but only be proven false. This is the theory of falsification that comes from Karl Popper (from Chapter 4). Popper's theory of science break down into two parts, both of which have had an incredible impact on the way we think about and teach scientific inquiry over the past 50 years.
1. Falsification (see above). For Popper, there can be no induction (that is, the quest to infer general principles from observation) because confirmation is an impossibility. Only falsification is possible.
2. Demarcation. Popper wants to distinguish between scientific and pseudo-scientific theories. For Popper, this comes down to a rather simple-seeming idea that scientific theories are those wherein scientists took risks by stepping outside of the known observations to try out new “conjectures.”
Only in reading this book have I come to confront the deep problems of scientific confirmation, about how we can actually know when a theory has been confirmed (because science presumes a constant degree of uncertainty, philosophers of science avoid the word “prove”). But I had always assumed the theory of falsification. Godfrey-Smith shows that most scientists use Popperian language and theories of science without understanding what he was actually arguing. Following G-S's argument, here are the major problems with Popper's two points:
1. The problem with falsification is the problem with scientific method in general. It's called the problem of “holism in testing.” Simply put, in order to falsify (or confirm) a theory, we must make assumptions about our ability to observe, quantify and analyze data. If any of those assumptions are faulty, our falsification (or confirmation) is false. Popper got around this by arguing that science was a set of behaviors that included making decisions about how and when to observe and what data counts. The problem with Popper's answer is obvious: If it boils down to “decisions”, then anyone can accept or falsify any theory based on which decisions they make about observation and data. If that is the case, there is not only no confirmation, but there is no falsification either. Hopefully it is obvious why this is an unacceptable theory of science. Popper then argued that you could choose to accept a theory if it was improbable that it would ever be falsified. But again we are left with the question of how improbable does it have to be before it's no longer falsified? Instead of his original claim, that observation, tightly linked to theories, have the power to falsify said theories, Popper ends up saying that falsification can occur without observation, because observation is probabilistic and based on decisions.
The problem really boils down to Popper's rejection of confirmation (G-S calls philosophers of this ilk induction skeptics). G-S illustrates the problem with a problem: If you have to build a bridge that has to carry load X, and you have two theories of how to build such a bridge, Bridge A has been tested for 50 years and never been falsified and another Bridge B is brand new and has never been falsified, Popper gives us no way to choose between the two. If Popper is correct and there is no way to confirm a theory, then two theories that have never been falsified are of equal value. Intuitively, the problem of the two bridges probably seems obvious to most people–you obviously choose the bridge that has been tested for 50 years. But remember, for Popper, Bridges A & B are theoretically equal, because there is no such thing as confirmation. Popper worked most of his later life trying to theorize an idea of corroboration, but never succeeded in solving this problem.
In the end, induction skepticism is as problematic as a naive belief in confirmation and induction proper. And so we are left with the problem of how to conceive of scientific confirmation, a question I'm sure Godfrey-Smith will address more as the book progresses.
2. So the scientific method most of us were taught is actually a kind of combination of Popper's method of demarcation (scientific theories are those which take risks) and a theory of confirmation (which Popper rejects). This boils down to the basic formula:
hypothesis/conjecture + observation/experimentation = theory/confirmation
Godfrey-Smith modifies Popper's demarcation theory to say that risk-taking is not about the production of the theory but about the way the theory is handled. G-S argues that Popper was onto something when he claimed that a scientific theory must be one that is set up for falsification, that it can risk observation. Popper argued that many theories can appear to have lots of possibility of observational falsification, but in actuality don't because they are never exposed to the risk of falsification. Specifically, Popper uses Marxism and Freudianism as examples of unscientific theories that are/were not produced in such a way as to be subject to falsification. G-S argues instead that the demarcation between scientific and non-scientific theory arises in the way those theories are handled. He gives examples of how Freudian theory can be handled as an a priori philosophy not subject to observation and falsification (e.g., in literary theory) or how it can be handled scientifically and open to observation and falsification (e.g., in cognitive psychology).
As I've been working on revisions of my book manuscript, I've been grappling with a comment one of the peer-reviewers made, that I needed to “use” more queer theory. I was educated early in my graduate career in queer theory and read it avidly until about 6 years ago, when I began to loose interest because it didn't seem to correspond to the people I was researching. What I've come away from in reading G-S this morning is a question about how to treat or handle queer theory scientifically and how to frame it for my book. If G-S's critique of Popper and Popper's critique of social scientific theories are correct, then the fundamental question I should be asking is this:
What observations would have to be made to falsify queer theory? Or given the observations already made by scientists and by me in my social-historical research, what aspects of queer theory have already been falsified?
I think in all honesty my research falsifies many parts of queer theory's basic premises of the social role of homosexuality, and I think the biology of homosexuality as it has developed over the past 15 years basically wipes out most strands of queer theory. And yet I'm in a discipline that takes it seriously and so I have to address it somehow (much like literary theory still takes Freud seriously).