www.philorum.org
The following are rough notes used as the basis for the speech. Nathan's resolution of the problem, outlined below, is given formal treatment in Sinclair, Nathan (2011) A Distinction between Science and Philosophy, Essays in Philosophy: Vol. 12: Iss. 2, Article 4.
Introduction
A way of distinguishing philosophy from science and other forms of enquiry that has two great features: it is forehead smackingly obvious, both in the sense that it is clear and that it is a clear inheritor of the traditional accounts of analytic philosophy that does not suffer from the defects of those accounts, and secondly no one else has thought of it in the fifty years since the difficulties of those traditional accounts of philosophy were discovered.
Partly because the solution is so obvious once you consider it, I would fail to communicate just why this problem is so important or what is at stake if I just stated the the distinction I am advocating, so lets begin with the problem.
Part 1
Last time I gave a talk here I went over several of the ancient accounts of philosophy, how it was that reasoning alone and without the aid of experience could produce knowledge. Now I am going to concentrate on the major shift in philosophy that took place at the beginning of the twentieth century in england (one of the three major developments in philosophy in all of western history).
This movement was championed by figures such as Russell, Wittgenstein, Moore, Whitehead, and has been the dominant form of philosophy for the last hundred years, though in recent times there has been a counter-revolution.
The central claim of this new movement was that experience was the only source of genuine knowledge. This idea had been around, but the new philosophers took it more seriously and applied it more systematically to philosophical problems.
In large part this shift was negative, all this stuff about the absolute and the world spirit and the unconscious entelechy of universal existence or the collective unconscious or the inevitability of class struggle and the rising of the proletariat, (all the stuff you still find in french philosophy and post modernism and progressive university english courses), all that was to go. Absolute nonsense strictly speaking because there was no experience that could confirm or deny any of these claims.
This claim about nonsense was the chief positive consequence of the new philosophy, the meaning of a sentence was the experiences that confirmed or falsified it. (This is the view that stephen hawkings recently claimed to hold at the same time he decried philosophy as useless to scientists.) Importantly, though it wasn't realised at the time, one could tell what would infirm or confirm a sentence by a sort of imaginative introspection, imagining what experiences would lead you to reject or accept a particular claim.
The problem for this new school of philosophy was to explain both mathematics and philosophy itself, for it seemed that here we had examples of knowledge that was not produced by experience. The solution for this problem was the notion of analyticity: analytic truths were true by definition, or by virtue of meaning, one didn't need experience because these truths would be verified by any experience and infirmed by none and one could tell this merely by contemplating the meanings of these sentences.
So it was that the vienna circle proclaimed in its manifesto that "the task of philosophical work lies in this clarification of problems and assertions, not in the propounding of special philosophical pronouncements".
Mathematical truths too were supposed to come out as mere definitional truths, and so it was that the challenge of either producing definitions which made all the truths of mathematics come out as definitional, or of showing why no such definitions could be produced was the major game in philosophy in the beginning of philosophy; sadly enough that whole project failed but in a way that did not undercut the notion of analyticity, and I shan't talk anymore about it.
Part 2
So what went wrong::
There are two priming factors:
So long as we thought that you could determine the meaning of sentences just by loooking within, by "contemplating concepts", things seemed ok, but this is unacceptable to a truly empiricist outlook - what is the observational evidence for such as Wittgenstine puts it "internal states do not come with labels attahced", to put it in another more empiricist fashion, if meanings are these internal objects what is the evidence upon which translation is based - what confirms or infirms the hypothesis that "chien" means in french what "dog" means in english.
So the notion of analyticity which the early positivists thought you could get for free as a kind of internal bonus, started to be viewed as something that itself stood in need of empirical support. The problem became one of identifying the analytic statements in some foreign language one encountered for the first time. (this foreign tongue is often called "junglese").
Second: almost as soon as the account of meaning was announced it was realised that individual sentneces were not in general confirmed in isolation. Some sentences like "see red" or "it's hot"or "There is a rabbit" are confirmed directly by observation but most claims, like "electrons have negative charge", or "Julia Gillard is the prime minister of australia" or "John Bentley is a bachelor" or that "neutrinos have traveled faster than the speed of light from the cern acellerator in geneva to a detector in italy" are not, they are suppoprted by a web of interconecting sentences which connect them in the end with experience. The analogies that the promoters of this wholistic account of meaning use are of bridges or buildings in which only the bottommost stones are connected with the ground, the rest being supported not only by those foundation stones but also by one another, another analogy is that of a web in which the nodes are sentences the lines between them are inferential connections and the anchor points the connections to experience.
Now the traditional account of anayticity seemed to give rise to a simple account of the empirical evidence which would justify taking a statement in junglese to be analytic, namely that the speakers of junglese would never deny this statement no matter what their experiences. But his has two problems, on the one hand there are other intuitively non-analytic statements which competent english speakers would never deny - consider "there have been black dogs", we all accept this statement and nothing I could do now would induce you to give it up. More importantly in the other direction holism implies that we can maintain any statement, keep any stone in our arch even if the foundation stones are moved provided we are prepared to distort the rest of the arch enough.
I won't develop these problems further but will give a brief summary of the overall flavour of these objections.