A topic I have been focused on for a bit is this idea of epistemic regress. By the typical definition of knowledge, a piece of knowledge is a belief which is true, that we have a justification for believing is true. The problem of epistemic regress falls out of this definition in the way that we justify beliefs. In short, the problem is that if we have a belief, it is typically justified by appeal to one or more other beliefs. Belief A is justified by belief B, which is presumably justified by belief C, and so on and so forth, causing an infinite regress. Which, if we don’t figure out some way to work out this infinite regress, could mean that we don’t actually know anything.
There are three basic ways we could resolve this regress. The infinitist theory is to say yes it is an infinite regress and that’s just how it is. Which is just terribly unsatisfying. Worse, if A isn’t justified unless B is justified, and B isn’t justified unless C is justified, etc. then we aren’t really justified in any of them because the string of unjustified beliefs goes on forever. We need some root belief that is justified. This brings us to the second way to resolve this problem, which is foundationalism. A foundational belief would be a terminal belief in this chain, that stops the infinite regress. The problem is that any belief that is foundational would need to be justified without needing another belief to justify it, which would be a very peculiar sort of belief indeed. The existence of such a belief is almost as hard to swallow as the idea that the infinite regress goes on forever. Now, the third possible way we could deal with this infinite regress is that there is no stopping point, and it doesn’t continue infinitely, but rather it goes in closed circular loops of indefinite length and structure. This solution is called coherentism. Superficially it seems acceptable, until you realize that you’re basically sanctioning circular reasoning as a means of justifying otherwise unjustified beliefs. That’s like saying God is good because the Bible says so, and we know the Bible is true because God is good (and wouldn’t lie to us). And we know that both those are true because it is circular. This logic is, to say the least, bad. Coherentism seems to suggest that this is acceptable, and that is a hard bullet to bite as well. So we have three bad theories that basically cover all the major solutions to an infinite regress. Either it is in fact an infinite regress, it terminates somewhere, or it just loops on itself. Now what?
Well, we have to test each one. Regarding the infinitist theory, it seems we are forced to conclude that we have little to no reason to think any of our beliefs are justified, and that we have no actual knowledge. This is most likely going to be our choice of last resort, since it basically concedes that we have no knowledge. Either that or infinite regresses are just fine, which is not really logically tenable. As a result this is not so much a serious theory of how to resolve epistemic regress as it is giving up.
Coherentism is a complex theory, and there are any number of ways you could go about attempting to prove it. If in fact our beliefs loop back on themselves, it also doesn’t appear we are justified in believing them, however most coherentists will argue that because the web of beliefs stands up coherently, we have reason to believe the entire group becomes justified to some extent. This seems a little bit plausible, but what about belief sets that are internally consistent, but disagree with the observable world? Does this same justificatory system apply to them? Unless a coherence theory adds in some element of observation then yes, and this would kill the theory. So a coherence theory must have some attachment to the real world in order to stop it from being legitimate to just make up all sorts of internally consistent nonsense and having it be justified as knowledge. Still, the truly knockout blow for coherentism is that a coherentist will generally rely on some belief which it makes little sense for it to be justified by a coherence argument. For example, “I think that coherence theory is true.” It doesn’t really make a lot of sense for a coherence theory to justify assertions about that sort of subject. But a coherence theorist cannot say that we have unproblematic epistemological access to anything, even our own thoughts, because that assertion that we have unproblematic epistemological access to our own minds is itself a belief that requires a justification, and a coherence theory will have to justify it by coherence with other beliefs.
Which leaves us with foundationalism. Which does seem to be the strongest candidate of the three, as well as the most intuitive, and is the one I will support. It is the typical way to resolve other infinite regresses, to find a base case. However its original problem still stands- what on earth could be a self-justifying belief? We need a belief that 1) is justified, and 2) does not depend on other beliefs for its justification. It’s easy to find beliefs that have either condition on their own. The belief that SKLAPPY SCHNAAA does not depend on other beliefs (if you are prepared to accept it is even a belief, which is debatable) but clearly it is nonsensical to say it is justified. Most beliefs we think are true will be justified, but will require other beliefs to justify them, even if those beliefs are as simple as what the subject concerned is. “That object is a chair” is a belief that depends on knowing what a chair is, for example, and possibly even knowing what an object is, or the sense in which one object can be of another class. Finding a truly foundational belief is quite challenging. Unlike a coherentist, a foundationalist can say that one of our foundational beliefs is that we have unproblematic epistemological access to our own minds. That can just be a foundational belief, if indeed it doesn’t require another belief to justify it by inference. And indeed it seems that the act of having a particular thought implies that it has been accessed by your consciousness, and it appears we have located our first foundational belief. Possibly. It could be that that “access” is really an experience of the sort that is suitable for a coherence theory, if you are prepared to accept that introspective experiences and sensory experiences are similar in nature, and can be used to modify a web of beliefs that is justified as a whole when it is internally consistent. Honestly, this is a truly complex topic.
There’s really far too much to this topic to ever hope to discuss all of it in one blog post, much less come to an actual answer. If you’re interesting in the topic do check out the Stanford Encyclopedia of Philosophy on Foundationalism, or on Coherentism.