Statements and Propositions

Citation preview

Statements and Propositions Bruce Aune Noûs, Vol. 1, No. 3. (Aug., 1967), pp. 215-229. Stable URL: http://links.jstor.org/sici?sici=0029-4624%28196708%291%3A3%3C215%3ASAP%3E2.0.CO%3B2-H Noûs is currently published by Blackwell Publishing.

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/black.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.

The JSTOR Archive is a trusted digital repository providing for long-term preservation and access to leading academic journals and scholarly literature from around the world. The Archive is supported by libraries, scholarly societies, publishers, and foundations. It is an initiative of JSTOR, a not-for-profit organization with a mission to help the scholarly community take advantage of advances in technology. For more information regarding JSTOR, please contact [email protected].

http://www.jstor.org Fri Jun 22 07:21:55 2007

Statemats and Propositions BRUCEAUNE UNIVERSITY OF MASSACHUSE'ITS

A common assumption in much recent philosophy is that the predicate "is true" applies fundamentally to statements rather than to propositions or sentences. Propositions are generally regarded as excessively mysterious entities, whose relation to observable acts of statement-making is every bit as puzzling as the relation between Plato's Doghood and the familiar creatures that chase cats and mailmen. Sentences are not thought to be mysterious in this way, but it is felt that they can be true or false only in a derivative sense. Fundamentally, it is what a man says that is true or false; yet if a certain sentence is commonly used, in certain circumstances, to make a particular statement, then that sentence might conveniently be considered true or false as well. Although I am prepared to grant that some clarification has been achieved by insisting that statements are the fundamental bearers of truth, I am nevertheless convinced that this clarification is far less extensive than is commonly believed. For one thing, a clear, unambiguous account of what, exactly, a statement is and how statements are related to speech acts has not been provided. The usual explanation, that statements are "what is said" when certain speech acts are performed, obviously tells us very little; and when it is pointed out (as by Strawsonl) that what is said on a certain occasion may be the same as what is believed or doubted on that or some other occasion, statements begin to appear just as mysterious as old-fashioned propositions. The first part of this paper will be concerned with the general question of what a statement is. My aim will be to outline some= See P. F. Strawson, "A Problem About Tmth-A reply to Mr. Warnock," Truth, ed. George Pitcher (Englewood Cliffs, N. J., 1964): 88-84.

215

thing it would be reasonable to mean by the word "statement." Since it will turn out that this word cannot felicitously do all of the work philosophers expect of it, I shall try to make sense of the notion of a proposition as well. In attempting this latter task, I shall endeavor to avoid some of the worst puzzles that seem inevitable with Platonistic conceptions-specifically, the puzzles that concern the relation between propositions and the actual thoughts, beliefs, and statements of living beings.

Although philosophers obviously use the word "statement" in a technical sense, it will nevertheless be instructive to begin by considering its standard use in English. As even a small dictionary indicates, this word is not strictly synonymous with related words that philosophers often treat, for their purposes, as interchangeable with it-I mean such words as "assertion," "affirmation," "claim," "remark," or "declaration." These nouns are not only different in meaning from one another, but they are each related to (actually built up from) special verbs, which are not synonymous either. These verbs are "to state," "to assert," "to affirm" "to claim," "to remark" and "to declare." In general one might say that in the strict sense an actual (as opposed to a potential) statement is what is or was stated, while an assertion is what is or was asserted, an affirmation is what is or was affirmed, and so on. It should be noted that anyone who makes a statement, an assertion, or an affirmation, is also saying something. According to J. L. Austin, stating, asserting, and remarking are illocutionary acts, and in performing such acts one is also performing a locutionary Thus, whether one states, remarks, or deact of saying-~omething.~ clares that it will rain, one is nevertheless saying, in all three cases, that it will rain. Since different illocutionary acts may, in this way, have a common locutionary or "saying" aspect, we commit ourselves to less in claiming that a man said something than we do in claiming that he made a certain remark or declaration. The verb "to say" is accordingly a fairly noncommital one. Since it is convenient to have a noncommittal noun corresponding to this verb in the way that "assertion" corresponds to "assert," the dictionary a See J. L,. Austin, How To Do Things With Words, ed. J. 0.Urmson ( Oxford, 1962) : 98-107.

STATEMENTS AND PROPOSITIONS

217

allows a secondary use of "statement." In this secondary sense a statement may be regarded simply as what is said when a man makes a remark, affirmation, and the like. This secondary use of the word "statement" seems most closely in line with philosophical usage, and in what follows I shall be concerned with the word in this minimal sense. In identifying an actual statement as "what is or was said (by a certain person, on a particular occasion)," one seems to' be treating statements as peculiar objects that are related in some way to a speaker. This appearance has been the source of familiar philosophical difficulties, notably those connected with Platonism. Yet this appearance is only appearance, for a man's statements are objects only in the sense in which a girl's smiles are objects, or in the sense in which a sense-impression is an object. I shall call such objects "nominal objects" and I shall say that they exist only in name. To gain some perspective on nominal objects, consider the following locutions :

( 1 ) Mary's smile was enchanting.

( 2 ) Tom's serve was forceful. ( 3 ) Dick's handshake was firm and manly.

The possessive singulars involved in these locutions (e.g. "Mary's .smilev) are what grammarians call "verb nominalizations." Though they may seem to refer to peculiar objects privately possessed (only Mary can smile her smile), their reference is actually nominal; it is a reference only in name. This may be illustrated by the following locutions, which unpack the force of ( I ) , (2), and ( 3 ) :

( 4 ) Mary smiled enchantingly.

(5) Tom served forcefully. (6) Dick shook hands firmly and manfully. Here the only things to which we are strictly referring are the persons Mary, Tom, and Dick. We are characterizing these people as doing something, and we are qualifying their doings by adverbs of manner. In analogy with these cases I want to say that if we regard "Tom's statement is true" as containing a referring expression other than "Tom," the reference it possesses is purely nominal. As

I see it, the force of the statement, so understood, would be made fully explicit by

( 7 ) Tom speaks (or spoke) truly, where only the man Tom is referred to and where his activity of speaking is characterized by what grammarians call "an adverb of affirmation."

I regard the remarks just made as bearing on the ontological status of the sort of statement (there are others) Austin evidently had in mind in his early paper "Truth and later attempted t o identify in connection with his doctrine of locutionary acts3 Strawson, in attacking what he took to be Austin's early conception, argued that the predicates of statements, such as "is true," cannot be transferred to speech acts or anything like speech acts.4 In this Strawson was clearly correct: we can no more apply the predicates of statements to acts of speaking than we can apply the predicates of smiles to acts of smiling. To grant this point is not to deny, however, that we can transform the predicates of statements into related predicates that can apply to speech acts. As just indicated, the predicate applicable to Tom's statement, "is true," can in fact be transformed into "truly," which characterizes his locutionary act of saying-something. In general I would say that the only reason for insisting that the predicates of statements (of the kind in question here) cannot apply to speech acts is the grammatical one that while the word "statement" is a noun to be modified by adjectives, the word "said is a verb to be modified by adverbs. In order to transfer the predicates of statements to the corresponding locutionary acts of saying, we must accordingly convert those predicates into adverbs. It must be admitted, of course, that the task of converting these predicates into adverbs is not always easy to accomplish; in fact it may sometimes seem to be impossible unless we introduce some rather contrived terminology. This difficulty might be regarded as an understandable consequence of the fact, emphasized by Strawson, that substantives such as "statement" are used chiefly to simplify our discourse. But while it is undeniably correct to say, as Strawson does, that it is "an ancient, but no longer respectable, error" to suppose that "whenever we use a singular substantive,

' Ibid. See Strawson, "Truth," Proceedings of the Aristotelian Society, Supp. Vol. XXIV (1950); see reprint in Pitcher, p. 35.

STATEMENTS AND PROPOSITIONS

219

we are, or ought to be, using it to refer to something,"6 we obviously need more assurance than this that our use of particular substantives is in fact ontologically noncommittal. I say this because if our use of certain substantives were merely a matter of convenience, it must in principle be possible to dispense with them, however awkward this might be. As an example of the difficulty we immediately encounter in attempting to find an adverbial translation for remarks about a man's statement, consider "Tom's statement that snow is white is true." Although we have no trouble finding an adverbial translation for the predicate "is true," if we want a similar translation for "'that snow is white" we shall no doubt have to invent it. Contrary to what one might initially think, however, our difficulty here is really not very serious. If we take the above remark as equivalent to "Tom said that snow is white and, in doing so, he spoke truly," there is actually no need to find an adverbial translation for the ingredient noun clause. I say this because there is no reason to assume that the noun clause is, in this context, a referring expression, whose use commits us to something other than Tom's speech act. On the contrary, when we say "Tom said that snow is white" what we mean, I believe, is that Tom performed a certain speech actthe sort of speech act one would normally perform by using the words exhibited (but not mentioned) in the oratw obliquu clause. And as I see it, we mean soonething similar when we report a man's remark in direct discourse, that is, when we say "Tom said, 'Snow is white'." Accordingly, I want to insist that the quotations and oratio obliqua clauses that may fill the blank in "Tom said . . ." are not used to refer to (or mention) something, but rather serve to characterize a man's locutionary act as to specific type by exhibiting the words he actually used, or, as in simple forms of indirect discourse, by exhibiting words that could be used to, perform an act of that type.6 Ibid, p. 34. Further qualifications must be made in order to account for the peculiarities of indexical clauses. In such contexts as "John said that he would come" the words following "that" are obviously not the ones John used or even a good translation of the ones he used; they rather constitute a differently indexed counterpart to his actual words, which might have been "I will come." Limitations of space make it necessary to ignore such contexts in the body of the paper, but they do not affect the general point I am concerned to make, namely, that the relevant that-clauses serve to characterize a man's act by exhibiting certain words-either the words that he did use or words that are related to his actual words in a special way.

Our inability to find a familiar adverbial translation for the noun clause in such expressions as "said that snow is white" is thus not very serious. It does, however, call attention to the fact that our standard means of characterizing locutionary acts is signscantly misleading. In using such sentences as "Tom said that snow is white" we characterize a man's act by focusing attention on certain words-either the words that he did use or words that might have been used to perform a locutionary act of that type. In doing this, we inevitably give the impression that statements are simply uses of words. Yet to accept this idea is, among other things, to take an important step on the unfortunate road to Platonism. For once it is granted that statements, taken as actual uses of words, are genuine rather than merely nominal items, there is no longer a plausible basis for denying that the "objects" of other conceptual acts, or attitudes, are genuine as well-and some of these "objects" will turn out to be highly Platonistic. Thus, although one may insist that statements are mere uses of words, one can hardly deny that when a man uses certain words he means something by them. Since one may predicate "is true" of what Tom meant in just the way that one may predicate "is true" of what he said, one's willingness to grant more than merely nominal existence to statements will naturally result in a commitment to grant more than merely nominal existence to meanings or "intensionsu-to that which a man meant in using certain words. But while a tough-minded interpretation of statements may seem to be attainable along the above lines, no similar tough-minded interpretation of meanings is even remotely in sight. For reasons of this kind, I have been anxious to regard statements as objects only in name, and I have thus stoutly resisted all temptation to allow that acts of stating may have genuine objects. By an informal commentary I have tried to show that while the expression "Tom's statement that snow is white7' has the grammaticalstatus of a singular term, it does not have to be regarded as a genuine referring expression. Although my informal commentary has been, I believe, adequate to its purpose, it may prove clarifying to supplement it with a very brief description of how we might characterize locutionary acts in a direct, purely verbal way if we were prepared to employ some slightly contrived terminology. Consider, to begin with, the French verb "tutoyer." In such sentences as "I1 Za tutoie" this verb provides a means of characterizing a man's locutionary act in a wholly nonsubstantive way.

STATEMENTS AND PROPOSITIONS

221

With a little commentary, "He thou's her" could be taken as a contrived English counterpart to the French sentence. Using adverbs already available in English, we could form more complex sentences such as "He thou's her fondly." In analogy with this, we could construct a new verb to characterize the locutionary act of saying that snow is white. The verb would be "to snow-is-white." The claim "John's statement that snow is white is true" could then be expressed in a purely verbal manner by "John snow-is-whites truly." Using this contrived verb would also allow us to make explicit various important dimensions of a complex speech act. Since "saying," as I have been using it, picks out the locutionary aspect common to certain acts of asserting, remarking, declaring, and the like, we could append what I call "adverbs of mode" to the contrived verb. Thus we could express "John asserted that snow is white" as, first, "John snow-is-whited in the asserting mode" and then as "John snow-is-whited assertingly." As I shall subsequently indicate, these contrived verbs of stating may be improved in certain ways so as to clarify such locutions as "Tom thought (it occurred to him) that 5 is a prime number."

The type of statement I have been considering, namely the actual statements of particular persons, is obviously not the only type of basic philosophical interest. Another important type may be identified by reference to such locutions as "the statement that snow is white." As I see it, the word "statement" in this sense means roughly "sayable," and the key to understanding it has been located, I believe, by Wilfrid SellarsS7According to him, the definite article in the expression "the statement that snow is white is true" is used in the "institutional" sense involved in "The whale is warm blooded." As is clear from the latter sentence, the singular term formed by prefixing this use of "the" is not a referring term, least of all a term referring to an abstract individual. While whales are plainly warm blooded, no abstraction is; abstractions possess no blood at aU. This means that if "the whale" referred to an abstract individual, the above statement would evidently be false, which it surely is not. SeUars has introduced a convenient designa-

-

'See Wilfrid Sellars, "Abstract Entities," Review of Metaphysics, XVI (1963) : 627-671. I am greatly indebted to Professor Sellars in this paper, and 1 want to thank him for his help.

tion for the singular terms formed by this use of the definite article; he calls them "distributive singular terms" (or DSTs). The sense of these terms may be indicated by the following schema: (8) The K is f = df All Ks are f Q , where ( i ) "f" differs from "f*" in a manner required by the usual grammatical transformations, (ii) being an f is at least a determinate feature falling under a determinable that is either essential to Ks or inductively implied by some feature that is essential to them, and (iii) the "all" on the right-hand side of the identity is adequate to warrant subjunctive conditionals. If "the statement that snow is white" is a DST like "the whale," a number of important consequences immediately follow. First, and most obviously, the use of such expressions will not commit us to abstract objects that are related in some puzzling way to human acts of stating and inferring. Since they are not referring expressions, there is thus nothing that they denote. To speak of the statement that snow is white is rather a shorthand way of speaking of human acts: to say that this statement is true will be to say, roughly, that anyone saying that snow is white will speak truly. This approach plainly brings statements down to earth, and shows us just how the basic reality concerning statements can be the locutionary acts discussed by Austin. In the second place, this approach allows us to appreciate the extent to which the predicates "is true" and "is false" are evaluative, and to see that their basic function is to assess (to authorize or discourage) assertions. If "the statement that snow is white" were a referring expression picking out some abstract object, it would be difficult to understand how the truth of this "object" could bear on the question of what one is entitled to do, that is, to assert. Yet even Tarski, who applied the predicate "is true" to abstract classes (classes of sentence tokens), constantly spoke as if the question of what is true immediately and unproblematically involves the question of what we may justifiably asserts8On my view of the matter, the fundamental application of "is true" is (indirectly) to locutionary acts or their mental analoguesg, and its basic function is to assess those acts as worthy of being done. See Alfred Tarski, "The Semantic Conception of Truth," Readings in Philosophical Analysis, eds. H . Feigl and W. Sellars (New York, 1949), p. 55. Strictly, the direct application of "is true" is to the nominal objects associated with these mental or verbal acts.

STATEMENTS AND PROPOSITIONS

223

Something must now be said about propositions. Since the topic of propositions is enormously complex, requiring for its adequate treatment a carefully worked-out philosophy of mind, only a few basic remarks can be made here. At the beginning of this paper I noted that there appears to be a sense in which what one .says, thinks, believes, conjectures or doubts may be the same. In traditional language it would be said that all these activities or .states may have the same intentional object or propositional content. This common object or content is what a proposition was traditionally taken to be. The range of so-called propositional attitudes potentially involving the same content or intentional object is very large, and ,some of them are very difficult to analyze.1° For simplicity, I shall here assume that the range of these attitudes is restricted to speaking and thinking, where thinking is regarded as an analogue, in some sense, of speaking. Given this simplification, the notion of a proposition may be treated in the same general way as I have just treated unowned statements. That is, "the proposition that snow is white" may be regarded as a DST (not as a referring expression), which is contextually eliminable according to the following general schema: ( 9 ) The proposition that snow is white is f = df All acts of saying or thinking that snow is white are fD,where . (as in schema (8) above).

..

Assuming that I am right about the expression "the proposition that snow is white," it then follows that in an important sense there are no propositions. A less paradoxical way of putting this, perhaps, is to say that nothing is denoted by such singular terms as "the proposition that snow is white." This claim is true, if I am right, because these singular terms are not referring expressions. In order fully to appreciate the schema ( 9 ) it will be necessary to grasp some further points regarding the general character of a statement or thought that p. To make these points maximally clear, I shall begin by commenting on the analogy Wittgenstein drew between using language and playing chess. Although there are well-known limitations to this analogy, I have no wish to dispute the use that Wittgenstein made of it; my purpose is rather to l o I have attempted a fairly detailed analysis of the most important of these attitudes, such as belief, in my book, Knowledge, Mind, and Nature (New York, 1987): Ch. 8.

explore an element that he did not explicitly consider. The element in question concerns the fact that the game may be played with a wide variety of other than standard pieces. Lacking a standard set of knights, pawns, and bishops, men might proceed to play the game with buttons, dice, numbered cards, lumps of candy, or even automobiles, given that they have a suitable board. The possibility of playing the game with these atypical pieces suggests that the general question whether a certain object is to count as a knight or pawn is basically independent of the specific empirical features the object possesses. Its identity as the piece that it is consists, rather, in the role that it plays in an actual game. Being a pawn is, one might say, a formal rather than an empirical (or material) characteristic of a concrete object. Note that "the pawn," like "the whale," is a DST. Accordingly, such statements as "The pawn captures diagonally" are reducible to such statements as "All pawns capture diagonally." Since, as just indicated, the question whether something is a pawn is answerable by reference to its use or potential use in an actual game, the statement "All pawns capture diagonally" may presumably be rewritten as "All concrete objects used as pawns in games of chess capture diagonally."ll If we take seriously the analogy between using language and playing chess, we might say that empirically different linguistic acts may be formally the same in just the way that empirically different chess moves are formally the same. We may affirm a conditional by saying "If p then q," "p only if q," or "q in case p." Here the empirical differences between our actual statements is entirely compatible with their essential formal sameness: they are the same in the sense that they play essentially the same "premissory" role. When a Frenchman says "La neige est blanche," we would express his statement to an Englishman by using indirect discourse: "He said that snow is white." The function of this particular construction is to locate the Frenchman's statement by exhibiting words that the Englishman might use to make a statement of a similar type. The similarity here is, of course, formal rather than empirical. The words used are empirically different, but the assert-

'' This is, of course, an oversimplification, since "diagonally" would require radical reinterpretation depending on the structure of a given board, for instance a curved one. In discussing the example of chess, I am ignoring the possibility of playing the game without a board, e.g., wholly in oral terms.

STATEMENTS AND PROPOSITIONS

225

ing job done is essentially the same. It will prove useful to have a special general term to mark linguistic similarities of this formal type. I shall construct such a general term by employing a special form of quotation.12 In analogy with the familiar practice of using ordinary quotation marks to form a general term applicable to words (as in "There are several 'the's on this page"), I shall use dot quotes to form a general term applicable to formally analogous statements. Thus, I shall say that both the Frenchman's "La neige est blanche" and the Englishman's "Snow is white" are 'Snow is white's. As I shall use it, the expression (in dot quotes) '"Snow is white"' is an illustrating common noun. The noun is illustrating because, like the "that snow is white" mentioned above, it characterizes a man's statement by exhibiting (rather than mentioning) a form of words that would normally be used, by certain speakers, in making a statement of that specific formal type. Since both the Englishman's and the Frenchman's statements mentioned above are of basically the same type in this formal sense, they may both be regarded as 'Snow is white's. I mentioned earlier that thinking may be regarded as an analogue in some sense of speaking. Although the concept of thinking is much too complicated to be analyzed here, I can give some substance to my remark by pointing out that a man's occurrent thoughts are normally characterized in the way that the Frenchman's statement was characterized, that is, by an oratio obliqua construction: "He thought (it occurred to him) that snow is white.'" In view of this familiar means of characterizing a man's thoughts, I shall extend my special use of illustrating common nouns to characterize thoughts as well as statements. I shall then say that while thinking is empirically very different from speaking, it is nevertheless formally analogous to it, being subject to the same basic logical and semantical norms.13 Thus, the thought that p may be inconsistent with the thought that q in just the way that the statement that p may be inconsistent with the statement that q, and so on. With these remarks in hand, I can now give a more explicit account of propositions. I have already indicated that "the proposition that snow is white" is a DST, not a referring expression. Having said that "The proposition that S is W is f' (where "S is W* la

cit.

Here again I follow Wilfrid Sellars; see his "Abstract Entities," loc.

l3 On this point see my Knowledge, Mind, and Nature, Ch. 8, and also Wilfrid Sellars, Science, Perception, and Reality (London, 1963): Ch. 5.

abbreviates "snow is white") may be regarded as equivalent to "'All thoughts and statements that S is W are f*," my new terminology now allows me to reexpress this last statement as "All 'S is W's, whether thoughts or statements, are f*." In order to have dot quotes on both sides of my equivalence, I shall introduce a new DST, "the (proposition) 'S is W'," as definitionally equivalent to "the proposition that S is W." I then have the following reduction schema for propositions: The 'S is W'is f = df All 'S is W's, whether thoughts or (as in (8) above). statements, are f*, where

...

In view of the complexity of the preceding remarks, a number of clarifying comments are in order. First, the suggestion made at the beginning of this paper that a statement (=what is said) may be the same as a certain thought or belief (=what is thought or believed) turns out to be true, though misleadingly expressed. The suggestion is true in the sense that many statements and thoughts may be characterized as, say, 'Snow is white's. With respect to a formal mode of classii?cation, such statements and thoughts may be regarded as the same kind of nominal object. What I say may be the same as what you say in the sense in which what I eat may be the same as what you eat if we both eat beans. As implied above, the expression "what he said" is not happily regarded as a referring expression. In the context "What he said is true" the expression is best understood as a placeholder for an answer to the question "What did he say?" Since "that snow is white" may be a satisfactory answer for both "What did he say?" and "What did he think?", and since the function of such an answer is to classify a conceptual act as to specific type, we see that the sameness of a given thought and statement will be a (formal) sameness of kind, not an identity of object. In saying that thoughts and statements that are 'Snow is white's may be regarded as the same kind of nominal object, I am of course reaffirming my earlier claim that such statements and thoughts are objects only in name, like smiles and handshakes. The illustrating common noun '"Snow is white"' is thus slightly misleading in the way that quotations and matio obliqua clauses were noted as being. It may, therefore, be useful to note that such nouns

STATEMENTS AND PROPOSITIONS

227

are in principle eliminable in favor of a contrived verb. Since the example given earlier, "to snow-is-white," was explained wholly in reference to speech acts, it will obviously not possess the required generality for replacing ".Snow is white'," which may apply to thoughts as well as statements. It is, however, very easy to improve the contrived verb so as to gain this added generality. We might simply place it in dot quotes-thus, "to 'snow-is-whitee"-and then say that the activities to which this verb applies are either speech acts or mental acts of the formal type indicated by the words exhibited in the contrived verb. My next comment concerns one of the reservations appended to the reduction schema for the institutional use of the definite article. The reservation was that "The K is f' is elirninable according to the schema only if f is (roughly) an essential feature of Ks. This reservation is needed because if the whale is thought of by John, it does not follow that all whales are thought of by him. I shall not discuss the appropriate treatment of nonessential predications of the whale here. But a brief remark must be made about nonessential predications of the proposition that snow is white. Suppose it is said that John entertains the proposition that snow is white. Exactly what is this supposed to mean? As I see it, the claim is a vague one at best, something to the effect that John wonders, considers, or perhaps weighs the question, whether snow is white. Whatever exactly John's specific attitude might be, he is in any case performing some mental act, which is an analogue of what Austin called an "illocutionary" act. If we h o w just what his act is, we would be in a position to classify it by a contrived verbal construction such as "He 'snow-is-white's wonderingly." If we prefer not to employ such artscial verbs, we could nevertheless insist that the so-called object of his act ( H e wondered, "Is snow white?") is purely nominal, like the so-called object of an act of stating. In either case we would not have to agree that he stands in an "entertaining" relation to some Platonic object. In general I would maintain, though I cannot argue the point here, that all nonessential predications of propositions are either psychological or else metalinguistic, as in " 'A' is used to express the proposition that snow is white." In the psychologicaI cases the expression "the proposition that snow is white" is to be eliminated in accordance with the strategy indicated in the previous paragraph, that is, by providing a classification for a mentaI

or verbal act.14 In the metahguistic cases the elimination of the propositional singular term is to be effected in some such manner as this: "the expression 'A' is normally used in making statements that are 'Snow is white's."15 The essential features of propositions are, as I see it, either semantical or syntactical. An example of a semantical predicate is "is true," and an example of a syntactical predicate is "implies that A is B." The reduction schema given above will allow the elimination of "the proposition that . . ." from statements involving these essential predications. I have already shown how this is to be done when the predicate is "is true," and I might add that when the predicate is "implies that A is B," the reduction can be carried out as follows: ( a ) The proposition that S is M implies the proposition that A is B. ( b ) The 'S is M' implies the 'A is B'. ( c ) All 'S is M's, whether mental or verbal, imply the 'A is B', where . . . (as in ( 8 ) above). ( d ) The 'A is B' is implied by all 'S is M's. ( e ) All 'A is B's are implied by all 'S is M's. ( f ) All 'S is M's imply all 'A is B's. The schema I have suggested for eliminating "the proposition that S is M in contexts of essential predication does not, of course, show us how to treat token-reflexivity. The treatment of this matter is, however, straightforward-and rather instructive. Since a tokenreflexive statement may allude to persons, places, and times, the general strategy for treating such a statement may be indicated by the equivalence: They might also, of course, require the classification of a belief, conjecture, or opinion. Limitations of space make it impossible to consider such cases here. l6 In connection with the remark made on p. 220 above, it should be noted that such statements as "In uttering 'La neige est blanche' Jacques means that snow is white" can be interpreted as "In uttering 'La neige est blanche' Jacques is making a statement of the .Snow is white. kind." This interpretation highlights the important fact that in order to mean something specific by certain words one must use them in accordance with certain logical and semantical norms. This important fact is commonly ignored by those who regard meaning something by S as a conceptually unanalyzable state of mind.

229

STATEMENTS AND PROPOSITIONS

The 'S is M' (if said of Tom, here, now) is f = df All 'S is (as in M's (if said of Tom, here, now) are f * , where ( 8 ) above).

...

On this accounting, one who says, speaking of Tom, "He is ill" does not make exactly the same statement that Tom does when he says "I am ill"; for while "He is ill" is a 'He is ill', "I am ill" is not. The difference of statement here, which will not surprise anyone familiar with the logical peculiarities of so-called intentional discourse, is traceable to the difference of words used or, in Austin's terminology, to the difference of rhetic act performed.16 Nevertheless, in view of the following equivalences based on the definition of "said of (by)" and "now," we can easily justify the less exacting sense of statement identity that is enshrined in recent philosophical usage: (10) "I am ill" (if said by Tom, now) is true= Tom is (now) ill. (11) "He is ill" (if said of Tom, now) is true = Tom is (now) ill. What these equivalences allow us to say is that while the two statements concerning Tom are strictly different, we may consider them the same in the weaker sense that their truth-conditions are by definition identical. Summary: I have been concerned to distinguish two basic categories of statements. One includes statements in the abstract, such as the statement that snow is white, and the other includes what might be called "personal nominal objects," such as Tom's statement that snow is white. I have argued that statements of both categories are strictly pseudo-objects existing only in name, the former being reducible to the latter and the latter, in turn, being reducible to locutionary acts of intelligent agents. Propositions, I argued, are also mere nominal objects, similar to statements of the "abstract" sort. Although discourse about propositions is entirely legitimate, it is merely a fagon de parler, equivalent to discourse about acts of stating or thinking and states of believing and doubting. The basic aim of my discussion has thus been the dual one of clarifying the philosopher's notions of statement and proposition and of relating these notions to the basic semantic reality of the speech and thought of intelligent agents. See How To Do Things With Words, especially p. 97.