Empiricism as the Default Position

It seems to me like all you’ve said above is “Empiricism is the default position” at greater length. When you say, “Anything that can be accounted for by rationalism can be accounted for by the empiricist,” what you mean is that empiricists are cconfident that anything that can be accounted for by rationalism can be accounted for by the empiricist — rationalists would deny it, and empiricists have certainly not addressed every rationalist argument on this question. I don’t even consider myself a rationalist, and I think it would be absurd to say that empiricists have shown that they can solve even the challenges that were put forward against empiricism in the early modern period (the few attempts that have even been made, at least as far as I have seen, have the same flaws Malebranche pointed out in the seventeenth century when he ripped up those versions of empiricism). Rationalists appeal to entities and like that are mysterious on empiricist principles; rationalists would simply point out that however different they may be from things we sense, they really aren’t so mysterious at all: we know lots and lots about them. (To give a version of an old rationalist joke, only empiricists think triangles and the set of natural numbers are completely mysterious, and perhaps then so much the worse for empiricists.) I don’t even know what you can mean when you say, “No one has ever even attempted to seriously give an account of what we know and how on a rationalist conception,” since surely you can’t mean that rationalists don’t give accounts of things we know and how we know them. And rationalists have often denied that empiricism can be formulated in purely empiricist terms, and, apparently, just as often been ignored. All that I think you’ve done is reiterate why, if you already accept empiricism, you’d take empiricism to be the default position. What you would need to show is that rationalists themselves are committed to taking empiricism as the default position.

I then clam that empiricism is simpler and better understood than rationalism and so should be preffered.

No doubt; this is also precisely what the rationalist disputes. He will say that empiricism is not simpler: its epicycles in order to explain things like the coherence of talk about infinites in mathematics get so complicated as to be absurd. He will claim that it is not better understood than rationalism: the empiricist has plenty of mysteries, which is why we get, to name just one example, all these interminable disputes about the qualia of sense-consciousness. He will go farther and claim that the empiricist helps himself illicitly to rationalist moves when it suits him, e.g., to avoid the specter of Humean skepticism about objective causation, or to allow in mathematics as consistent with empiricism. And rationalists have given arguments for such points; to reiterate just one example that I’ve already mentioned, Nagel’s _The Last Word_ is precisely a statement of traditional rationalist views on the inadequacies of empiricism.

I have no idea what you mean by a “predictive theory” of mathematical objects. What are you demanding? That the rationalist predict that all triangles on a Euclidean plane will have angles adding to the sum of two right angles, and that imperfect triangles on relatively flat surfaces will approximate this to the degree allowed by the imperfection of the plane and the lines? Done. What other prediction is required, that does not assume that the rationalist must pass some test that already presupposes that the empiricist is the default position?

My claim is not that the rationalists are right in any of this. My claim is that your argument that empiricism is the default position just appears to assume that empiricists are right. As I said, I’m inclined to think that the state of the argument is both sides giving themselves an A+ by assuming that they’re right, and giving the other side an F for exactly the same reason.

To add if it wasn’t clear, I’m not disputing that there appears no starting point to say one or the other is the ‘default’ position. It just seems that the empiricist can appeal to theories about objects earlier taken as immaterial (i.e. platonic objects) and that their answers appear to explain math as we develop it quite well. Arguably better than I think rationalists or platonists can. I’m not saying they are right. But, to return to mathematics, there’s something appealing about a philosophy that also deals with how we construct proofs.

Clark,

Constructivism et al. don’t actually provide an adequate response to rationalism, because the empiricist needs to show not merely that empiricist approaches can often or even for the most part approximate rationalist approaches (which is something that has usually been held by rationalists too) but that the rationalist approach is incoherent (which it must be if empiricism is right). So, for instance, it is not enough to show that you can do finitistic mathematics; the empiricist must show that you can’t do infinitistic mathematics (or show that there is a purely empirical way to admit genuine infinities) , otherwise he is leaving open a ground only the rationalist can cover.

Richard,

You say that by simplicity you mean “that the epiricist theory of how we know is itself simpler, not that the completed empiricist account of any given phenomenon is simpler.” But then it clearly cannot do the function that you intend for it to do. The reason for the importance of simplicity, remember, was that “when two theories are capable of explaining the same phenomena then the simplest, and best understood theory is to be preferred”; thus the simplicity has to attach to the explanation, not to some point upstream. That is, to be preferable, the value of simplicity can only be that it gives simple explanations, not merely that it has some simplicity somewhere in the theory (which is useless for telling us whether it is better).

But for the rest you’re clearly just using the empiricist approach to decide whether empiricism or rationalism is the default position. For instance, you say,

The empiricists says that the belief that p is true in virtue of the worldly fact that p or in virtue of conceptual connections between beliefs which are true in virtue of worldly facts.

So indeed does the rationalist; after all, if it’s not a worldly fact that 1 + 1 = 2, what is? Even the empiricist will usually admit it. And thus the empiricist cannot deny that the rationalist is talking about worldly facts unless he arbitrarily confines ‘wordly fact’ to ’empiricist interpretation’ of what virtually everyone agrees to be wordly facts. And so with the rest, as far as I can see.

It is in no way a problem for the rationalist that it is a priori conceivable that rationalism is false; obviously it would be so, because in a world without rational creatures, which is a priori conceivable, rationalism would apply to no cognitive agents at all. The rationalist doesn’t need rationalism to be necessary, but to be true.

I think it is actually; it is one that empiricists have traditionally had extreme difficulties with. I think one can argue that the early modern empiricists never found a way to solve it. To handle it Berkeley flirts with the idea that the Pythagorean theorem, while useful, is false (in his notebooks — he doesn’t seem to have done more than flirt with it), and Hume develops an extremely elaborate account of geometry, sprawling over Book 2 of the Treatise, and concluding in part that geometry is an inexact science (he also seems to be committed to the claim that the Pythagorean theorem is false). Malebranche considers a number of possible empiricist responses and shows that the empiricist can not show that any of them actually work unless they implicitly appeal to an idea of the infinite not derived empirically. I don’t think it has been very much on the radar of a lot of empiricists in the past century or so; I’ve never come across an answer that wasn’t just a version of one of the ones Malebranche noted. And certainly there are rationalists still proposing it as a problem — I mentioned Nagel before.

I wouldn’t want to rule out there being some adequate empiricist response, somewhere, but I don’t think most empiricists have the resources, and am inclined to think that the rationalists are right about at least one thing: empiricists may be thorough empiricists in principle, but in practice they’ll slip into rationalist modes of thought and speech whenever it’s easier to do so.

You are right that they can say that it is a worldly fact that 2+2=4, but they mean by ‘worldly fact’ something which is mysterious.

As you might guess, I don’t think the rationalist has to concede this; all they have to concede is that what they mean by ‘worldly fact’ is something that is mysterious to someone trying to fit it into an empiricist interpretation. I’m a little unclear about your response, though; while here you put the emphasis on the rationalist interpretation of ‘worldly fact’ (i.e., what the rationalist takes the things we know to be); but then you go on to talk not about this but about causal interaction, which is one way in which empiricists have tried to characterize the act of knowing (we are acted on by the environment), and talk as if this is where the mysteriousness of rationalism is. Is it the (interpretation of) the worldly fact, i.e., the thing known, or the manner in which it is known that is supposed to be mysterious?

(I say that causal interaction with the environment is ‘one way’ in which empiricists have characterized the act of knowing because, as you note somewhere above, some empiricists are constructivists. But constructivists are distinguished by the fact that they hold that at least some things that we know are not known in this way at all, but by constructing them. The point is actually an interesting one, because there is nothing about constructivism as such that requires empiricism; you could just as easily be a constructivist rationalist, and most rationalists have held that some knowledge is constructed.)

Rationalists tend to hold that causation is not a primitive: i.e., they tend to think it has an explanation in terms of something like structural constitution. E.g., my fist is the cause of the imprint in the dough because the collection of properties in my hand are united to the collection of properties in the dough, both being at least in principle capable of rigorous mathematical description, and the fist’s causing the dough to take on an imprint is in principle describable as an instance of a mathematical function. Thus causation is simply what you get when you have a system structurally constituted in a certain way, and the only reason for thinking causation relevant to knowledge is that we can at least rough out how the information is distributed across the unified system. (Thus, as a side note, most rationalists hold that causal systems are only well-understood if understood in rationalist terms.) If we take this route, though, then it isn’t so clear why only this particular type of structurally constituted system can allow for knowledge; why, for instance, couldn’t it be the case that a different interaction of systems in the brain is so structured that it yields a necessary and certain answer, even though that answer is not wholly derived by information coming in from the environment.

On the conceivability issue, I think it is simply open to the rationalist to deny the claim that it’s a priori conceivable that any given candidate for a necessary fact is conceivably not necessary; and if even one remains, the rationalist has identified a truth only the rationalist can prove (at least as the rationalist is inclined to think).

Also, where you say:

As for the causal stuff. How is the mathematical constituitional story you allude to above a distinctively rationalist account? Everything that you said is compatible with empiricism. Unless you are assuming something about mathematics, but then you need an argument for that.

But rationalists do give arguments for this; it’s a rationalist commonplace, in fact. So even if the form of explanation did not overlap, it would seem that your claim, which seems key to your argument, that empiricists can explain knowing, depends on the assumption that rationalists are wrong about the limitations of empiricism. If that’s so, it would prove my point that your argument for empiricism as the default presupposes that empiricism is right and rationalism is wrong.

The generally accepted picture emerging for cognitive science that talks about neurons firing and encoding information about the environment as an account of sensation is surely correct in broad strokes.

Yes, but what you keep forgetting is that the rationalist will say that this is a rationalist account of the brain — you don’t sense information, which is an intelligible structure found in communication channels. Remember, the rationalist complaint is that the empiricist is only an empiricist in principle; when he tries to show that empiricism is the way of doing science, etc., he always goes beyond what sensation can give. Now, this may be wrong. But it cannot be merely assumed to be wrong in an argument that empiricism is the default.

Rationalists aren’t committed to non-naturalism, simpliciter; they are committed only to non-naturalism of the sort that the empiricist would recognize. There are plenty of naturalist rationalists; Thomas Nagel, for instance. There are indeed rationalists who hold that mathematical tokens exist in space and time; after all, if mathematics describes the leaf on the tree, one can take that as meaning that the leaf on the tree is a particular token of a type of a mathematical object. What such rationalists will deny is that the the empiricist can recognize the leaf on the tree as a mathematical object at all, without assuming some principle or set of principles that can only be known if the rationalist and not the empiricist is right. More commonly, though, and Nagel would be an example, such rationalists are committed not to non-natural objects but to non-natural (where ‘natural’ here means simply ‘what can consistently be known on empiricist principles, i.e., by finite operations upon finite sensations’) features of natural objects.

Rationalists, of course, deny, and have always denied, that the empiricist can account for the infinite; and if you tell that to a rationalist, I guarantee that some of them would call you a liar to your face. Rationalists of various sorts have been arguing for the past three hundred years that empiricists can’t have a theory of the infinite without supposing something that can’t be known on empiricist principles, and they are occasionally touchy about it, because empiricists repeatedly have claimed to be able to handle the infinite without ever showing that the rationalist arguments on the subject are wrong.