A Reply to Mr Sellars

Citation preview

III: A Reply to Mr. Sellars Peter F. Strawson Philosophy and Phenomenological Research, Vol. 17, No. 4. (Jun., 1957), pp. 473-477. Stable URL: http://links.jstor.org/sici?sici=0031-8205%28195706%2917%3A4%3C473%3AIARTMS%3E2.0.CO%3B2-H Philosophy and Phenomenological Research is currently published by International Phenomenological Society.

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/about/terms.html. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/journals/ips.html. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.

JSTOR is an independent not-for-profit organization dedicated to and preserving a digital archive of scholarly journals. For more information regarding JSTOR, please contact [email protected].

http://www.jstor.org Fri May 18 09:05:00 2007

SYMPOSIUM

LOGICAL SUBJECTS AND PHYSICAL OBJECTS

Mr. Sellars' criticism of my paper contains misunderstandings of what

I am about. These I shall try to indicate as briefly as possible. (1) As soon as we understand the fundamental logical distinction of subject and predicate, says Sellars, it becomes obvious that particulars oan never appear as predicates, but only as logical subjects. The asymmetry of subject and predicate, he says, is already present in the singular, unquantified statement. For this logical asymmetry merely reflects another, ontological asymmetry - an asymmetry in the types of the items whioh oan enter into the facts that singular statements state. We see what this asymmetry is like when we see that these types of items can, with respect to the possibility of the items entering into facts, be ordered in a certain serial order, which has perhaps no end, but certainly has a beginning. Thus Socrates (an item belonging to the type of particulars) can be given as an example of wisdom (an item belonging to the type of firstorder attributes); wisdom oan be given as an example of being a virtue; being a virtue, perhaps, as an example of being an attribute of first-order attributes; and so on. But nothing oan be given as an example of Socratea. There is nothing which, in this scale, stands to Socrates as Socrates to wisdom, wisdom to being a virtue, etc. Whence it follows, by a few moves which I shall not repeat, that Socrates can never appear as a predicate but only as a subject. Now Sellars is careful to emphasize that he is not exploiting the actual asymmetry of the expression 'exemplifies' or 'can be given as an example of.' He is ready to follow me to the extent of ignoring grammar and replacing 'exemplifies' by something that doesn't even look asymmetrical, namely a oolon connecting nouns. The asymmetry, then, does not lie, so to speak, in the oolon ; it lies in the fact that the series of types of colon-joinable items (if I may mix the modes in a harmless way) has a definite beginning. I t lies, that is to say, in the fact that items of every type except one oan be colon-joined in two directions in the series. The one exception is the type of particulars. Thus wisdom oan be colon-joined to both Socrates in one direction and to being'a virtue in the other. But Socrates oan be oolon-joined only in one direction, not in two. But now I ask: why not? Cannot I make a series, too, such that Socra-

tes, like wisdom, can be colon-joined in two directions? Suppose I say that this snub-nose or this hand is an example of Socrates, or that remarking that Parmenides was in love with the One is an example of Socrates at his best. Or, changing the terminology in a way that appeals to some logicians, why shouldn't I say that this snub-nose is a sample of a certain head, that the head is a sample of Socrates, that Socrates is a sample of mankind, that mankind is a sample of animality, and so on. Here we have a series, with Socrates in the middle, particulars stretching out indefinitely to the left, universals indefinitely to the right. Or, doing another familiar philosophical violence to a word, why should I not say that this snub-nose is a member of the class Socrates, and Socrates a member of the class mankind? What is there about Sellars' colon which prevents one from forming, in this way, series of colon-joinable items, which extend in both directions from 'Socrates : wisdom?' I t is easy to imagine the protests. First: "But it is a misuse of the word 'example' ('sample', 'class,' etc.) to say that this snub-nose is an example of Socrates." To which I reply : Certainly it is; still, you would know what I meant if I said it. Then the protest might take the form: "But then you would be using the word 'example' in a different sense from the sense in which I use it when I say that Socrates is an example of wisdom, and wisdom of being a virtue; in this sense nothing can be an example of Socrates," And now of course I see: I see that it is a part of the meaning of 'is an example of' that only non-particulars can have examples. I see that when I construct my series of colon-joined items, which contains this snub-nose colon-joined to this head, this head colon-joined to Socrates, Socrates colon-joined to wisdom, wisdom colon-joined to being a virtue, and so on, the colon changes its sense in the middle: it becomes Sellars' colon only when at least one of the colon-joined items is a universal. So of course his series of colonjoinable items has a definite limit at one end, and of course this limit comes where he places it. And then, when the connexions are made as he makes them between the series of colon-joinable items, and the ideas of subject and predicate, of course a particular can never appear as a predicate. But this result is secured by making it part of the meaning of 'predicate' that a thing can be a predicate only if it can be colon-joined in either direction in the series, and then making it part of the meaning of 'colon-joined' that particulars can be colon-joined only in one direction. So Sellars has arranged for it to be part of the meaning of 'predicate' that particulars can never be predicates. He has built the category-distinction 'particular-universal' into the logical distinction 'subject-predicate' at the crucial point. And having built it in, he finds it there. I am not suggesting that there is anything arbitrary about this procedure. I concede that Sellars can perfectly reasonably claim to be 'reconstructing ordinary language' in making these connexions, i.e., to be

constructing a simplified model suggested by many features of ordinary speech, including ordinary speech about speech. Only it is surprising that Sellars should think I have overlooked the fact that the subjectpredicate distinction can be construed in his way. I emphasize as much as he does, for example, that in the ordinary logico-grammatical use of the verb 'to predicate,' it makes no sense to speak of predicating a particular, say Socrates, of anything. My question can be understood as precisely the question, why this use, this concept, exists, the question why we thus inject a category distinction into logical terminology. One way of raising this question is to proceed as I do: that is, first to introduce the terms 'individual' and 'predicate' in the non-committal manner of my first section: then to find a purely formal difference, independent of categorydifferences, in terms of which I can give body to the distinction of individual and predicate, and then to ask why this formal difference meshes in with the category difference of particular and universal, to yield just that mixed concept of the subject-predicate distinction which belongs to the tradition and which Sellars sets before us. Now this purely formal difference I found ( I still think, rightly) in the distinction between entering into discourse directly designated, and entering into discourse under the cloak of quantification, i.e., as one of a set (even if as the one which is uniquely so-and-so). I then gave reasons why particulars should be preeminently the things which enter into discourse in the latter way: and in these reasons found the explanation of the dictum that spatio-temporal particulars can never appear as predicates, only as subjects. Depending on the meaning you give to the terms 'individual' (or 'logical subject') and 'predicate,' this dictum is either an analytic proposition, as it is in Sellars' use, or a falsehood. The matter may be put thus. Suppose Sellars grants me my purely formal distinction between appearing as an individual and appearing as a predicate, and I grant him his mixed categoriallogical distinction between subject and predicate. Then my question can be framed as follows: how is one to explain the existence of the mixed distinction? And my answer to it can be framed as follows: the existence of the mixed distinction (and of our ordinary grammar which reflects it, and of our own ordinary use of verbs like to predicate, to characterize, to exemplify, which also reflect it) can all be explained in terms of the nature of the purely formal distinction, together with the nature of the difference between particulars and universals, our situation in the world and the needs which discourse meets. But my question is simply not one which Sellars faces; and perhaps I am, in part at least, to blame for this, for underestimating the power of the tradition, for overestimating the effect of Ramsey's skepticism. If Sellars wishes to say that my question belongs in the realm of questions about alternatives to our actual logical-ontological framework, he has my permission to do so.

(2) There is a strong suggestion in Sellars' paper that I would have done better if I had stuck to Cook Wilson. This suggestion I want equally strongly to repudiate. Certainly Cook Wilson draws attention to an interesting difference in ways in which items may appear in discourse. I t may be roughly expressed as follows. When we say Glass is elastic we may be talking about glass or we may be talking about elasticity (and we may, in the relevant sense of 'about' be doing neither). We are talking about glass if we are citing elasticity as one of the properties of glass, we are talking about elasticity if we are citing glass as one of the substances which are elastic. Similarly when we say Socrates is wise, we may be citing Socrates as an instance of wisdom or wisdom as one of the properties of Socrates. And of course we may be doing neither but, e.g., just imparting miscellaneous information. Now how, if a t all, could this difference help me with my question? Would it help at all, for example, if it were plausible (which it is not) to say that we were inevitably more interested in determining what properties a given particular had,than in determining what particular had a given property? Wouldn't this at least suggest that particulars were the natural subjects, in the sense of subjects of &erest? Let me answer this question by the reminder that what I have to do is to establish a connexion between some formal linguistic difference and a category difference; and a formal linguistic difference is one which logic can take cognizance of, in abstraction from pragmatic considerations, like the direction of interest. Such a formal ditference exists in the difference between appearing in discourse directly designated and appearing in discourse under the cloak of quantification. But the difference in the use of unquantified statements to which Cook Wilson draws attention is not a formal difference at all. Both glass and elasticity, Socrates and wisdom appear named in such statements, whichever, in Cook Wilson's sense, we are talking about. An appeal to pragmatic considerations is, certainly, an essential part of my own account at a certain point: but this is the point at which such considerations are invoked to explain why a certain formal difference should be particularly closely linked, in common speech, with a certain category difference. The difference of which Cook Wilson speaks is, then, though interesting in itself, irrelevant to my question. Cook Wilson is, and I am not, concerned with what Sellars calls dialectical distinctions. (3) A few remaining comments. (a) Sellars says that I confuse appearing as one item among others in the range of a variable with appearing as an individual, or being a logical subject; and I do this, roughly, because I fail to distinguish satisfying any propositional function from satisfying a certain class of propositional functions which he calls real functions. But of course I do not fail to

make this distinction. For my sense of individual, it just does not exist. For the traditional sense, it does. But I am trying to explain the existence of the traditional sense on the basis of my sense. (b) Sellars' bafflement over my use of the notion of predicate has the same general source in misunderstanding of what I am about. The parallel between wisdom and Socrates as non-primary predicates comes out clearly enough if we consider the forms of question "Which - are wise?" "Which - are of Socrates?" In both cases alike we need a kind of thing (e.g., speeches) specified before the question can be answered. (Let no one be troubled by the presence of 'of' in the second form of question. This is a requirement of ordinary grammar which we agree to neglect when we agree on the colon notation). Now Sellars says: "He (Strawson) sees, of course, that Socrates pure and simple can't be predicated of anything. He puts this, however, by saying that Socrates does not yield a primary principle of counting or grouping."

But of course it is not this which I see or which I put. For this, in my purely formal sense of appearing as a predicate, is false, though in Sellars' sense it is true. In my sense, Socrates appears as a predicate when we speak of certain speeches or virtues as Socrates's without naming the speeches or virtues, just as wisdom appears as a predicate when we speak of certain speeches or men as wise without naming the speeches or men; whereas in the statement that Socrates is wise, in which both particular and universal are named, the question, which appears as individual and which appears as predicate, is a question which does not arise; for they exhibit no formal asymmetry in their manner of appearance. In Sellars' sense of subject and predicate, oi course, all this is nonsense. But once again, I am trying to explain the existence of the traditional mixed sense in terms of my unmixed sense. (c) Finally, Sellars remarks that he is not clear why, according to me, an item appearing in discourse as the so-and-so cannot provide a secondary principle of grouping. But I didn't say it couldn't, and I did explicitly say it could. PETER F. STRAWSON.