There are those who have insisted that truth (or knowledge) is indefinable. Obviously that is literally false. Certainly nothing stops me from defining truth as, for example, the neighbor’s cat. Obviously that is an “incorrect” definition, in the sense that it isn’t conventional, but finding a correct definition is no harder than picking up a dictionary. Clearly then the claim that truth is “indefinable” must mean something else. Indeed the claim stems from the confusion of the philosophers who make this claim between definitions and theories. What they mean to say, I suspect, is that it is impossible to create a complete and justified theory about truth, something quite different from a definition.
But what is so hard about creating a theory about truth? I myself provided one only a few days ago, so it certainly doesn’t seem impossible, although whether the theory is properly justified or complete might be disputed. But I could only construct that theory because of the method I subscribe to and my theory about knowledge. And I would argue that while creating a complete and justified theory about truth isn’t impossible by itself doing the same simultaneously for truth, knowledge, reference, and method is a much harder task.
The problem lies not in simply creating these theories but in grounding them in some way. Certainly it is easy to simply proclaim something to be the case, but no proclamation, no matter how forcefully made, is worth putting any faith in. If we are rational about what we tentatively accept to be the truth then we will want the claims presented to us to be supported in some way, such that they are more likely to be the case than a claim produced by a random process. The problem is that the very things we are trying to theorize about are the same things that ground our decisions concerning which claims are justified. That is how I arrived at my list of truth, knowledge, reference, and method (assuming justification falls under knowledge). Claims concerning any one of these can be used to support claims, and so I assume that in our evaluations of claims we rely on a theory about at least one of them. (Reference may seem like the odd man out, and it is true that not every theory about reference can provide a foundation for other theories, but some theories about reference, such as those that claim we can intellectually “grasp” the subjects of reference when they are abstract objects, can be.)
If this is the case it would seem to make it impossible to rationally evaluate theories on these subjects. If we evaluate them by whatever process we are already unconsciously using to decide whether claims are justified or unjustified then it would seem that we are illegitimately prejudging the issue. Specifically we are assuming that the theory we are already judging claims by is a good one, and this may illegitimately rule out better theories which contradict it in certain ways. Of course on one hand the method by which we judge claims seems fairly reliable, at least when it comes to the claims we encounter most often in everyday life. The problem is that while our methods for judging claims may work well with claims about the everyday world they may fail miserably when confronted by more abstract claims, such as claims about the nature of justification itself. And it may very well be that the method that we use to evaluate claims may be open enough to other alternatives that they will indeed endorse better methods. But there is no way to be sure of that, because we would have to know what the better methods in fact are, and then see whether our existing methods of deciding would endorse them, and obviously that is impossible because it presumes that we have some way of determining infallibly what the better methods in fact are, which obviously we don’t since it would instantly resolve this entire problem.
To avoid that problem we might instead decide to judge these theories based only on their own standards (a kind of coherentist approach). Unfortunately this actually leaves us in a worse situation, because many of these theories will endorse themselves while rejecting their competitors (in fact they all should if they are constructed well). So instead of endorsing a single theory in a relatively questionable way we are now endorsing a large number of mutually exclusive theories, which is not an improvement (unless you like being confused). Alternatively we might hope to proceed from some unquestionable stripped down theory regarding justification, including only what we can be absolutely sure about from our intuitive thinking about justification, and from this theory arrive at a conclusion about which of our candidate theories concerning truth, knowledge, and so on is best. While this is an intriguing possibility there doesn’t seem to be any way to establish what belongs in this minimal theory without introducing the possibility that we are illegitimately presupposing something about the very subjects that we wish to objectively investigate.
This exhausts what I think of as the constructive ways to approach this problem. These approaches are constructive because they all attempt to determine in some way the correctness of these theories by deducing or establishing them from some pre-existing foundation. But obviously when it comes to questions about the foundation itself such constructions are hard to justify. A similar problem plagues epistemology in general, it seems obvious that most beliefs are justified by appeal to other beliefs, but when considering the beliefs on the bottom, the foundation, it seems as though they must be unjustified. (Some have proposed that they are self-justifying, but obviously that is unacceptable because it would mean endorsing a kind of circular logic.) The solution to the epistemological problem, as I see it, is to stop seeing the foundations as certain or unshakable, but rather to understand them as a kind of hypothesis, which may be later overturned. Of course that presupposes that what we take as foundational (evidence in my terminology) is overall more right than it is wrong in certain ways (such that is not perfectly and consistently wrong). And that is an assumption I can make only because of another principle I accept about knowledge, namely that what matters is only how well it explains that very evidence. Thus it doesn’t make sense to suppose that the evidence is consistently “wrong”, if they theories it leads us to construct are in perfect harmony with it then they are correct, at least as far as it matters to us.
Unfortunately this resolution of the epistemological problem cannot be easily carried over to the general problem of justification, because it presupposes something about our claims, namely that all that matters about them is how well they agree with experience. (And it may be hard to extend this account to cover all kinds of claims.) But if we are to solve the problem of justification in general I think it must be through a similar approach, perhaps by defining standards or some kind of framework and proceeding from there. A such an approach entails outlining what is acceptable, and, from there, deducing what is the case by determining what falls within those guidelines. (Of course the guidelines by themselves may not be sufficient to label just one theory as acceptable. However, we may be able to deduce from the standards certain features that any theory that can meet them must include. And since these are theories that define what justification is we can use those features as the kind of minimal set mentioned above in order to pin down more and more details about the correct theory, ideally.) Obviously I am concerned primarily with philosophy and philosophical justification, and when it comes to philosophy I think what we need to decide is what the correct method for philosophy is, and then, using that method, arrive at theories about truth, knowledge, and reference. In terms of the philosophical method this approach would be to decide what we want our philosophical method to yield (in terms that don’t appeal to any of the questionable concepts, such as truth). I’m not in a position to provide a complete list of standards, but I can list two big ones: we want it to produce understandable theories and objective assessments of theories (which is not to say that disagreement will not exist, just that in some ideal sense all disagreement can be resolved if all the parties involved are rational and actually follow the method). From the first requirement it follows that theories can’t contain contradictions (because the human mind seems incapable of understanding a contradiction except by putting it in such a way that it isn’t a contradiction). And from the second requirement it follows that we cannot be allowed to appeal to our intuitions, since the people differ in their intuitions. Naturally not everyone will accept such standards, and obviously we cannot defend them. However, we can point out that if they really disagree with us about these standards that they simply aren’t doing the same thing as us, they are pursing some different goal, even if we both describe ourselves as philosophers. And given such a fundamental disagreement we can safely ignore each other and go our separate ways.