I will be discussing Quine's paper "Designation and Existence," which appeared for the first time in The Journal of Philosophy in 1939. For our purposes I will be using the pagination of its occurrence in Readings in Philosophical Analysis edited by Herbert Feigl and Wilfred Sellars published by Appleton-Century and Crofts 1949.

For Quine "singular existence statements" take the form "There is such a thing as so and so." Further, the word or expression following "as" *purports* to designate a particular. Notice that the question of whether an expression purports to refer even outiside the context of a sentence or the context following 'as' in such constructions does not arise. Quine is explicit, however, in taking the 'is' as tenseless. In this sense, a sentence such as 'There is such a thing as W. V. Quine' is true, even though in four dimensional space-time "the temporally forward" end of Quine "lies behind" February 2001. The same cannot be said of terms that purport to refer to entities which are mythical, such as Pegasus.

In addition to "singular existence statements" we have "general existence statements," such as

There are unicorns.

Here in lieu of

There is such a thing as ...

we have

There is such a thing as a...

And then our example is formulated:

There is such a thing as a unicorn

The most obvious formal difference between singular existence statements and general existence statements is that in their formalization the general statements make use of variables. Thus (adjusting for keyboard limitations on quantifiers):

(Ex)(x is a unicorn)

General existentials are characterized by the absence of names and the presence of variables; these are logically analogous in meaning to "something which." In the case of our example, in words alone:

There is something which is a unicorn

'Unicorn' is a general term; 'Pagasus' is not; but 'unicorn *can* turn up in singular existential statements, however. This happens when we want to assert the existence of a certain *property*, such as being a unicorn. In this case we use the singular existential:

There is such a thing as unicorn.

Quine observes that I may assert this sentence while denying

There is such a thing as a unicorn.

Quine now inches forward, introducing Russell's treatment of the denials of singular existential statements like

There is no such thing as Pegasus.

Such sentences are especially puzzling to those who fail "to observe that a noun can be meaningful in the absence of a designatum." What these philosophers may may be puzzled by is that if 'Pegasus' designates something, then any denial of the existence of Pegasus is going to be false, but if it designates nothing the sentence is even worse off; it is meaningless. So distinguishing meaning and designation is important. The *meaning* of 'Pegasus can be expressed (translated) as "the winged horse captured by Bellerophon."

It's designation is a matter best left to zoology; we don't come to know of animals by only understanding the names for them. Not all nouns are names. Furthermore, nouns that don't name are lumped in with prepositions in this important respect, and so Quine appears to be rejecting the designative function of relational terms, a move we are inclined to link with nominalism. This characterization, however, contains the danger of causing us to miss some important issues. Let's touch on a couple of points.

Quine notes that factual considerations can dispel belief in the actual existence of some abstract entities, such as the mythical disease "hyperendemic fever" in the same way as factual considerations can cause us to relinguish any belief we may have had in Pegasus, but the existence of abstract entities corresponding to prepositions is vastly different. 'Up' has meaning, although Quine alleges that its nonexistence is not a matter of fact in the same way that the nonexistence of Pegasus is a matter of fact. Is he denying the sentence:

There is such a thing as up?

I think he is, but he is saying also that the word 'up' has meaning. In my opinion, when later Quine denies there are meanings (qua entities) he is denying them in the same non fact related sense as he is denying up exists. Notice that the correctness of a translation is not a matter of fact, and that "analytical hypotheses" differ in this regard from the theories of physics, a point of contention between Quine and Chomsky where Putnam sides with Chomsky. But let's stick with this text for now. The point is that the issue of analyticity is grounded in Quine's incipient nominalism, historically at least. We must be very careful to observe the distinction *at this stage* between Quine's nominalism and what realists have called "dead end nominalism." Here Quine is rejecting *not* abstract entities but rather *some* abstract entities, those in particular whose existence is not open to factual examination, such as is the reality of appendicitis. The important thing here is that there are abstract entities the existence of which can be rejected on the basis of empirical examination. This is something Quine will cling to. In fact, he reminds readers in 1943 of what he had said in 1939:

But the idea, according to which the observation of nature is relevant only to determining the existence of spatio-temporal particulars and never the being of universals, is readily refuted by counterinstances such as that of "hyperendemic fever" in my paper "Designation and Existence".... ("Notes on Existence and Necessity" JP. March 4, 1943. p. 116).

Before going any further I want to make one observation that may interests those of you who are linguists. When Quine says in "Notes On Existence and Necessity" (hereafter NEN) that we cannot infer

(Ex)(there is no such thing as x)

from

There is no such thing as Pegasus

the problem may not be that "'Pegasus...never occurs designatively..." The problem may be viewed differently as violating what linguists call the principle of Full Interpretation which requires that all all expressions which are terms of a sentence be interpretable (my formulation sb). If the inference were allowed we would get something like

(Ex)(There is no such thing as Pegasus)

In this case, either 'Ex' or 'There is' would not have an interpretation, violating the principle. Howeva! This violation w.r.t quantifiers, as I recall, is violated in at least one valid proof of Lowenheim Skolem (ascending?). Anyway, I can't go into all that. Having seen that factual considerations can lead us to reject abstract entities - keep in mind that meanings are abstract entities for Carnap - what is to stop the nominalist, Quine asks, from denying there is such a thing as appendicitis?! A nominalist may choose to

...maintain that the word is not a *name* of any *entity*...and that it is a noun at all only because of a regrettable strain of realism which pervades our own particular language

. What is being maintained here is that these non naming terms are merely "syncategorematic," and this word is important. In fact this is among the very first occurances of the term in Quine, a term which Quine will use in _Word and Object_ a term which does not name, occurring opaquely. Here there is anticipation of Quine's later discussion of the de re/de dicto distinction in the quantificational domain, but this comes only AFTER the concept of 'pure designation' is available as a a major tool in Quine's philosophy. The important question remains: why can't *all* terms be syncategorematic? Quine seems to be casual in remarking that "now the whole question of existence is beginning to appear gratuitous, but it will become very important at a latter point that the very much the sameremark will come to apply to 'analyticity'. thing will happen to 'analyticity'. Quine's point is this: So what if all these terms fail to designate an entity? What is the nominalist *denying*? And if nothing, why should we care what he does?

What is left but a bandying of empty honorifics and perjoratives - "existent" and "non-existent," "real" and "unreal"?

The philosopher who "outdoes" the nominalist by treating all words syncategorematically, and not just a few words is not a nominalist at all, it seems to me. For one sort of nominalist there is only reference to concrete particulars. For this guy there is no reference to *anything*. Is this is an idealist gone linguistic? But if the nominalist (attenuated idealist) is to convince us that he has anything of content to say, he

must find some relationship of logical dependence between the singular existence statement and the rest of discourse. (Sellars p. 47)

>From 'There is such a thing as appendicitis' no change is forced on the assertion 'Appendicitis is dreaded'. From 'There is such a thing as appedicitis' no change in truth value is forced on 'Appendicitis has been known to kill'. Nor is it even the case that the denial of the existential affects the truth value of 'Appendicitis has been known to kill', since if it is true there is surely no change and if it is false so too is 'Appendicitis has been known to kill'. So we need something that the existential makes a difference to, and that something is application of the logical rule which most of us know, perhaps, as "existential generalization."

If we affirm the singular existence statement, we must regard any general existence statement "(Ex)(...x...)" as following from the corresponding statement "...appendicitis..." which contains "appendicitis" in place of "x."

And of course if we deny the existential ("there is such a thing as appendicitis") there is no such consequence. Quine goes so far as to say that a word *designates* if and only if (iff) we can perform this logical operation on it (that is, existential generalization). But what might this have to do, if anything, in the least with ANALYTICITY? This is the question we want to answer. Quine while skimming along says

Perhaps we can reach no absolute decision as to which words have designata and which not, but we can say whether or not a given pattern of linguistic behavior *construes* a word W as having a designatum.

Now here is the rub: if we describe an analytic truth as a truth in which there are no designators and so depends only on something like "form," then if there is no deciding as to which sentences contain words that designate then there will be no way of deciding which sentences are ANALYTIC. I think there is this sort of connection between designation and necessity, even in 1939! Two observations in concluding our discussion of "Designation and Existence": first, close to the end of the paper (Sellars p.51) he says something that is probably what made him a famous philosopher: "The universe of entitites is the range of values of variables. To be is to be the value of a variable." Second, he will describe constructing an abbreviation that will allow the nominalist to dispense with unwanted entities. As long as he wants, the nominalist can use 'proposition', for example, but when he wants he can dispense with it by substituting the terms of the abbreviation. (There is something "Mickey Mouse" about this, I suspect) He will restrict his quantification to names of individuals and presumably live happily ever after doing set theory. We have seen, then, that there IS a sense in which Quine moves away from analyticity without discussing meanings, as he did in "Two Dogmas." To be sure the two can be related, but they do differ. However, as early as 1943 Quine says a number of things that can be construed as a proto attack on analyticity. The relevant passages pertain mainly to "analytical necessity." Quine says:

The notion of synonymity is presupposed also in the notion, so current in philosophical circles since Kant, of *analytic* statements. It is usually to describe an analytic statement as a statement that is true by virtue of the *meanings* of the words; or as a statemetnt that follows logically from the meanings of the words. Given the notion of synonymity, given also the general notion of truth, and given finally the notion of logical form...we can define an analytic statement as any statement which, by putting synonyms for synonyms, is convertible into an instance of a logical form all of whose instances are true. (NEN p. 120)

A bit later he notes that "no intensional mode of statement composition is needed in mathematics." Now because analyticity implies such intensional statement composition, for example in the failure of existential generalization in sentences expressing analytic necessity (cf. NEN p.123), the whole notion of analyticity becomes suspect. Also, if meanings are attributes (Carap's intensions) then since the context "the meaning of ..." is not an extensional form of sentence composition, it too, and with it analyticity, becomes suspect. But here is something I find a little puzzling and don't quite know how to deal with it. Quine says:

Conversely, also, given the relation of synonymity it would be easy to derive the notion of meaning in the following way: the meaning of an expression is the classs of all the expressions synonymous with it. No doubt this second direction of construction is the more promising one. (NEN p. 120).

Isn't it possible to have a language with no synonomy? Isn't it possible that this were true of all languages (meta etc). The metalinguistic notion of synonomy doesn't appear essential to characterize semantics. Is this right?

Home