An Epistemological Note

In this course, we are investigating metaphysical issues: What is the nature of the relationship between a thing and its parts? What is the nature of the relationship between a person and its parts?, etc. An entirely different question, which we are not investigating, is epistemological: How do we know about things and their natures? Though the latter question is entirely different from the former, it is not completely unrelated.

I get the impression from some that the only good answer to the epistemological question is that we know that p just in case p is proven, where this requires something very strict, like deducing p from purely logical truths. This raises many delicate issues in epistemology that are outside of the focus of this course (but well within the focus of Professor Feldman's class, 'Theory of Knowedge', for those who are interested). In this post, I want to give some reason for thinking that requiring logical proof in order to accept p is too strict. I'll conclude with a bit about how I see the role of arguments and proofs in our metaphysical investigation. This is by no means a compulsory way to view the role of arguments, but it is one that seems compatible with a less strict account of what it takes to know that p.

So, to be clear, here is the Strict Epistemological Thesis (SET):

(SET) S knows that p only if S deduces p from purely logical truths.

Here's an objection to (SET):

1. If SET is true, then if a car is bearing down on S at high speed, S does not know that there is a car there and that S is in danger. (This proposition, I assure you, is not derivable from solely logical principles. It requires something like seeing that a car is there, which is no truth of logic.)
2. If S does not know that there is a car there and that S is in danger, then S's dodging the car is irrational. (The idea is that there is a connection between what we know and what we do, and what we don't know and don't do.)
3. S's dodging the car is not irrational. (We would think someone who saw the car but just stood there is crazy or otherwise mentally unfit.)
4. So, SET is false.

I think a proponent of SET should deny (2); S may have a justified belief that there's a car and S is in danger. And having a justified belief is a sufficient basis for rational action.

This objection seems fine to me. But note what it concedes. If this reply is correct, we don't need deductive proof from logical premises to have a justified belief that p, and having a justified belief is sufficient basis for rational action. So, for instance, I could be justified in believing that tables exist, and this allows me to rationally do stuff like put things on tables and say things like 'Tables exist'. In other words, if some proposition is justified for us, then it is rational to accept it. And being justified does not require strict proof.

The moral seems (to me) to be that, even if SET is true, it is rational to accept what we have justification for, and justification requires something less than proof. Consider, for instance, how a proponent of a related view, (SET'), would reply to (1-4):

(SET') S has justification for p only if S deduces p from purely logical premises.

It's puzzling why a proponent of (SET'), if she really accepted her view, would dodge the oncoming car.

So I think that all of this shows is that, regardless of the status of (SET), (SET') is false. But all we are after in a metaphysical investigation is to be justified in accepting some one metaphysical account versus the others. This, in turn, only requires that (SET') is false.

Sophisticated sympathizers of (SET') may have noticed that the above argument against their view does not meet their standards for proof. After all, though the proof was deductively valid, its premises were far from logical truths. Thus, a proponent of (SET') might maintain that this argument does not compel her to abandon her position.

I supect that, if a proponent of (SET') is sufficiently committed and resourceful, there are no arguments that an opponent of (SET') could supply that would force her to abandon her position. But there are two different aims we might have when thinking about (SET'). First, we might aim to provide arguments that a proponent of (SET') would accept which would show her view is false. Second, we might aim to argue that those who are "shopping" for a view of justification or knowledge that (SET') is not the way to go. Though I think the first aim may be impossible, the second may be attainable.

G. E. Moore, an analytic philosopher who wrote early in the last century, argued as follows:

1. Here (holding up one hand) is one hand.
2. Here (holding up his other hand) is another hand.
3. If (1 and 2), then there are at least two hands.
4. So, there are at least two hands.
5. If (4), then there are at least two things external to our minds.
6. So, there are at least two things external to our minds.

(6) is a conclusion that proponents of (SET') cannot accept, since (1) and (2) are clearly not truths of logic. A proponent of (SET') might argue like this:

1. If S is justified in believing p, then S deduces p from logical truths. (statement of (SET'))
2. Alleged justified belief in hands, cars, etc. is not derived from logical truths.
3. If (1 and 2), then no one is ever justified in believing that they have hands, that there is a car, etc.
4. So, no one is ever justified in believing that they have hands, that there is a car, etc.

What Moore wants to know is, if he is not justified in accepting his premises, what justifies the proponent of (SET') in accepting hers? Suppose we are shopping for a view and we are undecided on whether Moore's way or (SET') is the way to go. Moore asks, how can it be more reasonable for you to believe an abstract philosophical thesis like (SET') than it is for you to believe his premises? Common sense compels us very strongly to accept that there are hands, cars, etc. This is not the case with (SET'). So if we are shopping for a view, we should prefer Moore's to the more skeptical alternatives.

The upshot seems to be that, if Moore is correct, (SET') is unacceptable. But if that's so, there's room for us to have justified beliefs in things that are not established solely on the basis of principles of logic.

I'll end by explaining one way in which a Moorean might view the role of epistemology in our metaphysical investigation. We start, at the outset, with our common sense and the dictates of science. (It's worth noting, very briefly, that science, all by itself, does not imply that there are no tables, etc. One route to this view is Ontological Fundamentalism, which is explained and evaluated here.) We are, by and large, justified in accepting what common sense and science tells us. In many cases, our reasons are very strong, though they do not constitute a proof.

We want to find out the answer to some metaphysical question, such as 'What is the nature of the relationship between an object and its parts?' Insofar as possible, we should try to accept a view that comports with our pretheoretic conceptions. To help us focus our investigation, we consider some problems about the relationship between an object and its parts. We want the answer to our metaphysical question to adequately address these problems while preserving what's right about our pretheoretic conceptions. So we consider several views and arguments for and against them. The idea is that the view that fares best against objections and otherwise meets our criteria is the one we should accept.

This turns out to be enormously difficult, since it involves discovering the very best arguments for and against the views. But even if we do not decisively resolve the issue, we gain the benefit of a better understanding of the relationship between objects and their parts. (This point generalizes far beyond metaphysical investigation; what is offered here is one picture of how we might view the project of understanding the way the world is which is the common goal of all disciplines, not just philosophy. On this picture, the sciences, for instance, primarily differ from philosophy only in offering empirically based arguments.) No deductive proof of any view is required or expected. No deductive proof from purely logical principles that an objection is sound is required or expected. Rather, the arguments for and against give us an idea of what we should believe "on balance". So if a view implies (for example) that there are no hands, that is a cost of the view, even if it does not show (in the sense of (SET')) that the view is false. What we need to figure out is which costs are worth paying and why.

This approach is broadly accepted by analytic philosophers. That does not mean that you should accept it, of course. To re-emphasize, the point of this post is to give one picture of the role of arguments in metaphysical investigation and how this relates to the epistmological question of what we should believe. As with any view, there are alternatives. If you are attracted to an alternative view, it may be helpful in developing your view to see where, exactly, the Moorean view goes wrong and why. And if you are attracted to the Moorean view, it is also fruitful to try to see how one who rejected it would argue against it. A fundamental presupposition of this course is that, by carefully considering our views and the views of others, we arrive at more well-considered views and a better understanding of the subject of those views.