Positive theories of knowledge assert that, if they are correct, future guesses are guaranteed to have (at least probabilistically) a marked improvement in their objective verisimilitude, not just in their increased empirical adequacy. If this were true, it would be an immense boon for everyone. Logical negativism rules such a possibility out a priori; in fact, it originates in the supposed failure of all positive theories of knowledge. Therefore, the greatest argument one can muster against this dogma in logical negativism is to demonstrate that some kind of necessary increase in verisimilitude occurs when replacing an old theory with a new one.
This dogma of logical negativism is found in Popper’s treatment of scientific theories, but it is present in, and perhaps central to, Quine’s views on language: The most we can learn from experience is that we are wrong, not that we are right; however, any phenomenon that contradicts the whole will overthrow all theoretical premises. Empirical refutation does not definitively pick out any one statement as responsible, which leaves us with a choice: we may defend our doctrines come what may, or we may choose to ‘lose the argument’ with nature gracefully and revise our doctrines.
Joseph Agassi once said something to the effect of ‘to lose an argument is to win’. He’s right in one respect: ‘winning’ an argument may be the result of skilled oratory at the podium, yet persuasion is not the best strategy towards pursuing the truth. If we choose to search for the truth, we are required to want to lose whenever possible. When we lose, we learn that we are wrong, and from this understanding comes a transformation.
The Socratic dialogue, then, necessitates a particular behavior or methodology, which is dependent on the attitude of wanting to change (this attitude can be instantiated in specific activities or the structure of communities): we must not just give lip service to fallibilism, like so many epistemologies today (read: Bayesianism, externalism, internalism, &c.); we must embrace fallibilism as a central ‘dogma’ to our epistemology.
The problem with negativism, though, is that it does not, and cannot, guarantee that our future guesses will be improvements over our past guesses. All we can know is that we (1) are willing and happy to admit error; (2) we have tentatively adopted a theory that, as far as we can tell, survives criticism better than known alternatives.