Comment on Jay Rosenberg's 'Categorial Ontology'

Jaroslav Peregrin

Feb. 6, 1996

It seems to me that Sellars indeed managed to develop the Carnapian idea of 'Ueberwindung der Metaphysik' in a way which avoids Carnap's philosophical naivete (and in a way which, although not so easily passable as the one offered by the other towering figure of this 'second round of the linguistic turn', Quine, runs terribly deep). However, as much as I am inspired by the Sellarsian picture itself, I am always embarrassed by certain pronouncements Sellars makes when it comes to the questions of the very significance of the picture. Maybe my embarrassment is a matter of missing something in Sellars views; but even in such a case I think this might be a starting point for an interesting discussion.

Let me quote from Sellar's paper Meaning as Functional Classification:

I have often been asked, what does one gain by abandoning such standard platonic entities as triangularity or that 2+2=4 only to countenance such exotic abstract entities as functions, roles, rules and pieces. The answer is, of course, that the above strategy abandons nothing but a picture. Triangularity is not abandoned; rather 'triangularity' is seen for what it is, a metalinguistic distributive singular term. (Synthese 27 (1974), p. 436)

I cannot make sense of this answer. Of course, I, for my part, do not need any answer to this question, for I know that the Sellarsian picture has really tremendously helped me to understand language and the world, and to gain new light on many philosophical problems. However, there are clearly plenty of people who do not share this conviction, and they can legitimately be asking this question. And to tell them: you have been bewitched by a certain picture, and what I now offer you is no picture, but rather things as they really are, seems to me to be hardly of any help. If one needs a reason to appreciate the picture, then it does not help to say to him that the reason is that the picture is correct.

Thus, it seems to me that the question Sellars so quickly sweeps off the table should be considered much more carefully. Let me hint at the possible ways of approaching the question which I can see: One such way might perhaps be to argue that only if we view language in this way, can we manage to reach a naturalistic picture of the world, and then go on to argue that naturalism is helpful in that it, e.g., lets us give an account of the world with the help of "an extraordinarily small store of unexplained explainers". Another way might be to argue that the other view is dangerous in its ability to misguide us into plainly wrong beliefs concerning certain matters of fact: that it is, e.g., likely to lead us to a completely wrong notion of how language (as an empirical phenomenon) works (it may, for instance, lead us to the view that introducing a new word is a kind of christening). I wonder which answer (if any) would be the truly Sellarsian one.

Anyway, it seems to me that it is important to realize that he who believes that meanings are abstract objects and that 'means' is not a copula, but a relation between a linguistic and an extralinguistic entity sees the whole thing, of course, the other way around: for him, his picture is the plausible one (it is in a sense natural to see meanings as objects; we do speak about meanings as about things, we do say such things as 'triangularity and scalene triangularity are two different things'); and the Sellarsian picture is a mere unnatural 'brain teaser'. My impression is that we have to admit that there is no prima facie reason not to take meanings as abstract objects (and I am convinced that this picture may even be--in certain contexts--quite helpful), but perhaps we could show how this way of thinking about meaning may be dangerous in that it is likely to be severely misguiding.

It seems to me that all of this leads us to the important question of whether we should view Sellars as replacing one metaphysical picture by another metaphysical picture (claimed to be simpler, or more perspicuous, or more realistic, or whatever), or as cutting deeper--towards some kind of, to return to the phrase of Carnap's, 'Ueberwindung der Metaphysik'. I would vote (together with, e.g., Rorty) for the latter option, but Sellars' own pronouncements, like the one quoted above, seem to me to sound as if he were after the former--as if what he was after did not amount to replacing one description by another ('better') description, but rather to doing away with descriptions and showing how things really are in themselves.


[Return to Main Page]