What Is A Priori Justification, And How, If At All, Is It Possible?

A priori justification is essentially justification for a belief that is independent of experience i.e. based on reason alone. I will be mostly looking at Bonjour's account of how a priori justification is possible, and suggesting some problems that arise, and give a sense of why a priori justification is such a widely disputed topic. I wish to demonstrate why my conclusion on the matter is ultimately a tentative acceptance that a priori justification may be possible, although there is by no means certainty around the issue.

Firstly, Bonjour has given a negative account of a priori justification, stating a belief is justified independent of experience if the justification is by some other source than the 5 senses, introspection, memory or testimony, however it should be noted that it need only be independent of any experience beyond what is needed to acquire the necessary concepts to understand the proposition in question. This is sufficient if we accept we can know necessary truths (2+2=4), analytic truths (a bachelor is an unmarried man) and arguably some a priori contingent truths (If Jones fixed the reference of 'one meter' by 'the length of S at t0', then S is one meter long at t0) independent of experience. However, these examples still rely on the acceptance of the existence of a priori truths, which some sceptics claim is not possible, as they argue that everything is known empirically. It is arguable that only a priori truths can be a priori justified, but I will return to this later in the essay.

If we assume a priori justification is possible, there remains the question of how. In order to be based on reason alone that implies there must be some kind of unmediated grasp of the truth of the proposition at hand. Bonjour argues for a non-inferential, immediate act of rational insight/intuition called 'appearing/seeing', and this is the foundation of the justification. Anything that relies on intuition (or similar) immediately faces problems. For example, a sceptic may argue that a basic 'seeing' is merely 'gut-instinct' that would need to be verified empirically anyway if it were to have much worth. A defence may be that at least in the case of a priori truths they do not need empirical verification because of their very nature. But this too is unconvincing as a sceptic could simply question how we know we are not mistaken in believing this a priori proposition (an extreme example being we don't know there isn't an evil demon making us believe them despite them being false). It would appear at least that anything relating to the a priori cannot easily be justified without (very simply put) resorting to an 'it just is the case that' approach.

Instead, we could consider acts of induction, and how we come to the conditional claim that a repeated premise is highly likely to have a certain explanation, rather than simply being by chance or coincidence. In other words, not a spontaneous moment of enlightenment when we come to know something, but rather the act of considered reasoning that is justified a priori. Note also that 'highly likely' truths could be believed rather than necessary a priori truths in general as a response to sceptics, meaning we merely claim it is highly likely they are to be necessary rather than that it is certain. They could still be justified a priori, even if we did not regard them as a priori truths. Bonjour argued that many errors involving apparent rational insight can be corrected internally by further reflection, and it is this further reflection that is a priori justified. Most would agree that we all do this is one form or another on a day-to-day basis, and there does not appear to be any better justifier. For example, we think we know something, but after considering the matter more deeply (and have not conducted further research) we then conclude something different. However, this argument is self-defeating. Since there is no way to verify self-reflection, we could just as easily later conclude that the second reasoned conclusion is also incorrect, and so on. Any internal reasoning cannot be checked against anything to determine whether it is accurate, and therefore it is arguably not very convincing.

Putnam argued that there seem to be no truths where it would never be rational to give up in some set of circumstances under which it would be rational to accept its denial. A key motivation for this is that we don't know that future experiences wouldn't contradict a seemingly a priori justification. For example, some physicists have claimed that some subatomic events occur at random, so the claim 'every event has a cause' which is commonly held to be true could potentially be false. However, the idea of disregarding certain truths (or at least propositions commonly held to be true) on the basis that something, at some point may show them to be false seems dangerously close to radical scepticism. For example, if we entirely mistrust everything that can be doubted (in true Cartesian fashion), then we are arguably left with very little knowledge on anything except perhaps ourselves. Though some people are happy to accept this, it seems more likely that there can be some kind of middle ground that at least allows for truths to be held on a day-to-day basis.

The process of induction (if we accept this to provide an a priori justification for a proposition) could be an a priori justification for an a posteriori proposition. If we found that without exception copper was observed to melt at 1083°C, in a variety of circumstances including different times, places etc. then it seems we have what Bonjour would call a standard inductive premise: "all observed As (copper objects) have been observed to be Bs (to melt at 1083 degrees centigrade)" . But this still faces the problem that something could at some point prove it to be false.

In conclusion, it does seem that a priori justification is possible in some sense, but in what sense is unclear. The problem closely relates with the existence of a priori truths, which requires an individual be open to non-empirical forms of justification. For the time being, this seems possible, however as more scientific discoveries are made which may go against commonly held beliefs, it seems we should at least be tentative towards anything relating to the a priori, even if we do not want to be entirely sceptical of all knowledge in general.