When someone cites one of my old papers, I get a notification. And I’m always surprised to see from responses to papers that I wrote a decade or more ago that parts of the scientific realism debate are going on in roughly the same terms that they’ve been going on for 40 years. Realism is taken to be the commitment that we should believe in unobservable posits of our best scientific theories, like electrons— alternately, that our best theories are probably at least approximately true.1
Lots of the debate seems to rely on treating realism as if it’s committed to finalism and overconfidence.
By ‘finalism’ I mean the view that the answers we’ve got now will not need to be revised or abandoned in the future. It is easily rebutted by the historical record of past theories, by any plausible projection of how future science will go, and by noting that the present is not some magically privileged position.
By ‘overconfidence’ I mean the view that belief means certainty. Some anti-realists claim that not believing in electrons would mean taking alternatives more seriously in a way that would be healthy.
I think we should be fallibilists, so I reject finalism and overconfidence. Fallibilism is the the admission that our beliefs might be wrong. It portrays our epistemic and doxastic situation as already and always being underway. We never get to step outside of ongoing inquiry to report final results. This naturally requires attention to well-formulated and potentially fruitful alternative lines of research. Perhaps we’ll find out that our prior beliefs need to be revised; perhaps we’ll just understand their context better.
This is all compatible with believing in electrons. I’m more confident that there are electrons than I am about empirical claims about distant parts of the world that I’ve read about but haven’t thought about too much. I’m more confident that there are electrons than I am about hazily remembered bits of my own biography.
So it seems to me that either (a) a modest realism is right for mundane, common sense reasons, or (b) the usual way of understanding ‘scientific realism’ should be abandoned. On neither horn of the dilemma does contributing more words to the scientific realism debate serve any purpose.2
Two thoughts; the first flip, and the second more serious. First, the fact that everything everyone says about realism is wrong is actually a good reason to expect more articles and books about it to appear: lots of low-hanging refutations are available.
Second, since I find myself (mostly against my better judgment) currently trying to write a book about the pessimistic induction, I wanted to try to articulate where I get off the boat with your line of reasoning, mostly just to get clear in my own head. I think there are two ways of denying ‘finalism,’ one of which is compatible with realism, and one of which is not. You defined ‘Finalism’ as the view that “the answers we’ve got now will not need to be revised or abandoned in the future.” I think the key question is how ‘big’ you allow the revisions to be, or how much we have to abandon. If we have good evidence that the future changes will be relatively small/ mostly consistent with our present theories, then you CAN affirm realism and deny finalism. However, if the future changes are going to be huge, i.e. current and future theories cannot both be approximately true, then that is a good reason against realism. (In your original formulation, ‘revised’ sounds more like the first, small-change option, and ‘abandoned’ sounds more like the second, big-change option.)
Why? Because if you (think you) will get evidence tomorrow that refutes p (even though you don’t have that evidence today), then believing p today is irrational. That is, if you have good evidence for both “All the evidence available to me today supports p, on balance” and “I will acquire new evidence tomorrow that will refute p,” then you should NOT believe p today. For example, imagine you have good reason to believe that the evidence available to you today is in some way misleading or unrepresentative, but that you will acquire less misleading/ more representative evidence tomorrow. Someone in that situation should doubt the hypothesis that is best supported by the evidence available to them today. (See R.A. Briggs’ “Distorted Reflection,” 2009, p.77)
At least, that’s how I’m currently thinking about these questions. The pessimistic induction is second-order evidence (or at least is strongly analogous to 2nd-order evidence).
(P.s.: And who are the people claiming realists can’t be fallibililists? I don’t know anyone who claims ‘Our current theories COULD be wrong’ entails anti-realism. If you want to share, you can email me or whatever if you don’t want to name names here.)
Greg: Thanks for replying!
My fallibilism does not merely allow that my beliefs will need small revisions. I fully expect that some of them will turn out to be pretty tragically wrong. It won’t be all or hopefully even most of them, but I am in no position to identify in advance which are problematic. I believe anyway because the alternative is an all-consuming scepticism.
The schematic version of the Pessimistic Induction that you sketch is coherent. However, as the debate is framed, the problem must arise for beliefs about unobservable entities and phenomena in a way that it doesn’t arise for observables. I doubt that the history of science could underwrite clear enough conclusions for that. And I don’t see how it could be strong enough to make all-consuming scepticism look appealing.
As for the PS: I won’t name anyone, because I suspect my reading is rather uncharitable.