Falsifiability is a stupid criterion

[Angrboda]

It's morning, so you're not going to get anything incredibly coherent from my direction, but FWIW. (And I have someone waiting, so I don't have time to properly proof and edit this; mea culpa.)

I haven't read Kuhn except in the secondary literature, so I can't really comment on that. I do think, however, that there is a conflict between falsifiabilty as a criterion for science and how most people think about truth and knowledge. I conjectured with some support that most people naively operate using what is known as a correspondance theory of truth. A correspondance theory of truth is a theory that what makes a statement or proposition "true" is that there is a fact of the world that "corresponds" to the fact expressed by the statement. (E.g. The statement "all swans are white" is true only if the swans that exist in the world are colored white.) This seems a natural and intuitively correct way to view things. In the first part of the twentieth century, a group of philosophers known as the Vienna circle set the tone of thinking about the philosophy of science and other things. One of their notions, verifibility, grew into its own major philosophical position known as logical positviism (whihch in turn greatly influenced Ayn Rand and Objectivism). Verificationism and logical positivism assert that the truth of a theory is measured in how well confirmed that theory is, how many observations appear tot validate the thory, and, that only propositions that are and can be verified this way express "truths" that have any real meaning; "facts" or theories that can't be verified, are on this view meaningless. Now I won't go into detail on the philosophical problems with this view other than to make the common observation that this view tends to strengthen the problems in human psychology known as confirmation bias, the human tendency to only look at things that agree with our view, and ignore or minimize those that don't.

Well, how is this a problem? Well, when verification was the standard of how one determines the utility of a hypothesis and whether it is true or not, there wasn't much problem. Truth consisted in pairing up statements about the world with facts about the world that match that statement. And verificationism made that the gold standard by asserting that what made a scientific theory meangingful and valid was how well it lent itself to being confirmed by finding observations which appeared to pair what was observed with what the theory suggested we should expect to observe. The theory of what made good science, and the theory of what made statements true (correspondance) fit together hand and glove. Popper and others changed that by introducing notions such as falsifiability as the criterion by which the quality and meaningfulness of scientific theories are judged. However, people did not change their notions about truth, that truth is correspondance between statement or theory, and a fact in the world. So people's ideas about the fruits of science, that it produces statements which correspond with or describe facts of the world, is in some sense mismatched with the current epistemology, philosophy and practice of science.

I can't claim to understand what this means in practical or philosophical terms, nor do I really understand how to characterize the understandings of science in a philosophical way as a theory of truth, a statement about how science and scientific statements are judged in terms of truth content. What I will say, however, is that there appears to be a mismatch between what people in general (and many scientists) believe that scientific statements are saying, what they mean, and the current philosophy about what having a good, falsifiable but not yet falsified theory means in terms of notions about its "truth" and "validity". People still expect science to confirm theories, but it's not at all clear that science is doing this, can do this, or even should do this. (The other side is a view that the "truth" of a scientific theory is embodied by how useful the theory is; that if a theory leads to the production of useful technological goods, that is a measure of how true that theory is. I'm not going to go into that aspect here.)