IV: THE CHISHOLM-SELLARS CORRESPONDENCE

The purpose of the present chapters is to discuss, and to offer some evaluation of, the dispute between Chisholm and Sellars concerning intentionality as it appears in their published correspondence. [Chisholm and Sellars 521-539] Because one of the most puzzling features of this correspondence is the difficulty one encounters in attempting to become clear on the precise nature of the disagreement between Chisholm and Sellars, one of our chief concerns in this chapter will be to present a plausible account of this disagreement. Since the foregoing chapter was devoted to an examination of Sellars' position on intentionality, it will be especially useful against that background to become clear on those respects in which Chisholm seems to be in disagreement with the views advanced by Sellars, both in his letters to Chisholm and elsewhere. For this reason, the first section of the current chapter will consist in an attempt to present two claims, in the making of which Chisholm appears to take exception to Sellars' views, and the second section will be an attempt to make the first of these claims as explicit as possible. The third and fourth sections, then, will be devoted to an evaluation of the disagreement between Chisholm and Sellars which seems to result from Chisholm's making these claims. A fifth and closing section, finally, will be devoted to some points made in summary and the conclusions reached earlier in the chapter.

I. As we have seen, one of the chief features of Sellars' position concerning intentionality is his insistence that we are able to imagine a community of language-users whose language contains the linguistic resources required to talk about linguistic items in semantical terms, but does not contain the linguistic resources required to talk about psychological items. For unless the language-users who, in Sellars' myth, have not yet been exposed to Jones' theoretical reasoning, speak a language devoid of discourse which is about psychological items, Sellars' myth will not succeed in presenting an answer to the question of how a Rylean language might have become enriched in such a way as to allow the introduction of discourse about psychological items.

On Sellars' view, as we have seen, there are two ways in which what he calls a Rylean language must be enriched if we are to be able to see how discourse about psychological items might have been introduced. The first is semantical discourse, and the second is theoretical discourse. For if, as Sellars believes, psychological discourse is simply discourse about certain sorts of theoretical entities which are conceived of as exhibiting properties like the semantical properties exhibited by linguistic items, then in order to construct psychological discourse we shall need to have both the ability to talk about linguistic items in terms of their semantical properties and the ability to talk about theoretical entities and to establish their existence on the basis of appropriate sorts of theoretical reasoning. Since it seems clearly possible that there could be a language-using community whose language includes semantical discourse but no even rudimentary theoretical discourse, it seems clearly possible if Sellars is correct that there should be a language-using community whose language does include semantical discourse but does not include psychological discourse.

It is this feature of Sellars' position which intuitively seems to generate much of Chisholm's disagreement with Sellars. Thus in his first letter to Sellars, Chisholm suggests that the fundamental issue between them is with regard to the answers which each would be inclined to give to the following questions:

(1) Can we explicate the intentional character of believing and of other psychological attitudes by reference to certain features of language; or (2) must we explicate the intentional characteristics of language by reference to believing and to other psychological attitudes? [Chisholm and Sellars 521]

Chisholm's initial account of the issue separating him from Sellars consists in his suggestion that whereas he would answer affirmatively to the second question and negatively to the first, Sellars would be inclined to answer affirmatively to the first and negatively to the second. [Chisholm and Sellars 521]

The considerations which might lead one to advance such a characterization of Sellars' views would seem to stem from that aspect of his views which we have considered above, namely, that it is possible that there might be a community of language-users able to talk about linguistic items in terms of their semantical properties but unable to talk about psychological items at all. Similarly Chisholm's insistence on an affirmative answer to the second question above seems to result from a feature of his position which we have examined above. For while Chisholm concedes that "the sentences we use to describe the meanings and uses of words" ordinarily exhibit the sort of logical characteristics which Chisholm takes to be definitive of sentences which describe psychological phenomena [Chisholm (17) 516], Chisholm argues that these sentences are elliptical for sentences which are explicitly psychological. [Chisholm (17) 516-517] Thus it seems that Chisholm's position commits him to holding that when we talk about linguistic items in such a way as to characterize them by reference to their semantical properties, what we say makes implicit reference to psychological phenomena such as believing. For this reason, Chisholm seems committed to an affirmative answer to the second of his two questions.

The difficulty with this sort of characterization of the issues separating Chisholm and Sellars is that it is not clear how the claims made by each are strictly incompatible with one another. Thus Sellars' position involves the possibility that there might have been a language-using community with a language which had certain linguistic resources but not others, whereas Chisholm's claims seem to be about a language-using community (possibly ours) which has both sorts of linguistic resources in question. Thus in his last letter to Sellars, Chisholm writes that he concedes "that it is conceivable that people might make semantical statements about one another's verbal behavior before arriving at the conception that there are such things as thoughts." [Chisholm and Sellars 537] What seems unclear here is whether in making this "concession" Chisholm is granting a point which is not relevant to what he conceives to be the issue which separates him from Sellars, or is in fact granting something which results in diminishing the difference between the two positions.

A second difficulty with this initial characterization seems to result from the fact that there is a certain terminological disparity between Chisholm and Sellars at the outset of their correspondence, and although this appears to be eliminated by the closing letters of the correspondence, no specific mention of the matter appears to have been made at any point. In framing the possible objection involving semantical statements to his position in "Sentences about Believing," Chisholm writes that

if we think of a word as a class of sounds or of designs, we may be tempted to say, at first consideration, that intentional sentences about meanings and uses of words are examples which run counter to our general thesis about intentional sentences. For here we have sentences which seem to be concerned, not with anyone's thoughts, beliefs, or desires, but rather with the properties of certain patterns of marks and noises. [Chisholm (17) 516]

Thus Chisholm seems to wish to consider words, as typical cases of linguistic items, as "certain patterns of marks and noises," and to argue that we cannot describe a person's "use of language, or his dispositions to use language in those ways, unless we refer to what he believes, or knows, or perceives." [Chisholm (17) 517] Thus on this view, sentences about linguistic expressions per se would not be intentional in Chisholm's sense; they would be sentences simply about certain physical objects and episodes. By contrast, sentences about people's use or dispositions to use these sorts of physical objects and episodes would be intentional in Chisholm's sense; but on Chisholm's view they would be elliptical for sentences which make explicit reference to certain psychological phenomena.

It seems to be this feature of the terminology which Chisholm adopts in both "Sentences about Believing" and his initial letters to Sellars which leads him to characterize Sellars' position as involving the claim that "marks and noises have a funny kind of characteristic [namely, intentionality] that living things and other physical things don't have." [Chisholm and Sellars 524] For Sellars characterizes linguistic expressions as items which are intentional, in the sense of being about something, in a sense which is the same as or analogous to the sense in which psychological phenomena are intentional or are about something. Thus Sellars writes that "marks and noises are, in a primary sense, linguistic expressions only as 'non-parrotingly' produced by a language-using animal," [Chisholm and Sellars 524] thereby indicating that when he speaks of linguistic items, he intends to be taken as speaking only of those marks and noises which have the requisite history to be counted as linguistic expressions. Because Chisholm seems to view linguistic items as no more than physical events and objects which exhibit certain physical properties, such as phonetic or geometrical properties, he is led to characterize Sellars as assigning such physical items intentional properties such as being about something solely in virtue of their physical properties as marks and noises.{1}

Somewhat later in the correspondence, however, Chisholm appears to adopt a different terminological convention. For whereas he had earlier spoken of linguistic expressions as simply "certain patterns of marks and noises" which exhibit certain specifiable physical properties [Chisholm (17) 516], he writes in his penultimate letter that "linguistic entities (sentences, etc.) are . . . intentional," [Chisholm and Sellars 533] thereby seeming to indicate his acceptance of a terminology which involves characterizing anything which is, properly speaking, a linguistic entity, as about something, or making reference to something, and so forth. In order to have available a convenient terminology with which to deal with the issues which seem to separate Chisholm and Sellars, therefore, we shall use the expression 'semantical item' to apply to all and only items which are used as parts of language and which may appropriately be characterized in semantical, or metalinguistic, terms. Similarly, the expression 'semantical episode' will be used, as it was in the foregoing chapter, to apply to episodes which satisfy those conditions. For the adoption of this terminology seems to reflect not only the distinctions implicit in Sellars' discussion, but those which Chisholm comes to accept in his later letters.

With this possible source of confusion disposed of, let us turn to an attempt to provide some acceptable formulation of those respects in which Chisholm wishes to take exception to Sellars' position. The initial characterization of one of these respects rested on the idea that on Chisholm's view, we must "explicate the intentional characteristics of language by reference to believing and to other psychological attitudes." [Chisholm and Sellars 521] In his later letters, however, Chisholm seems to advance several claims which may express more accurately what, in this initial characterization, he had in mind. Thus Chisholm claims that if one is to explicate what is involved in semantical discourse, in discourse, that is, about semantical items, one must "introduce some term which, if it means anything at all, will refer to what . . . [Sellars calls] thoughts." [Chisholm and Sellars 523] Similarly, Chisholm claims that sentences conforming to what Sellars' calls the 'means'-rubric--sentences, that is, of the form " '. . .' means ---."-- are "to be analyzed as" sentences about certain psychological phenomena, and not conversely. [Chisholm and Sellars 524] And finally, at the close of his last letter, Chisholm writes:

If the people of your myth were to give just a little bit of thought to the semantical statements they make, wouldn't they then see that these semantical statements entail statements about the thoughts of the people whose language is being discussed? [Chisholm and Sellars 537]

In each of these claims, Chisholm appears to wish to advance the idea that there is a certain relationship between sentences which characterize semantical items in terms of their semantical properties, and sentences which are about psychological phenomena, and that the relationship is not symmetrical. Letting a semantical sentence be any sentence which is about some semantical item and which (at least partially) characterizes that item in semantical, or metalinguistic terms, we might formulate the claim which Chisholm has in mind as

(1) Every semantical statement entails some psychological statement, and it is not the case that every psychological statement entails some semantical statement.

It seems clear that the second conjunct of sentence (1) does not represent a point of difference between Chisholm and Sellars, given that semantical statements are about linguistic expressions and characterize such expressions (at least partially) in semantical, or metalinguistic terms. For it in no way seems to be a part of Sellars' position that every psychological statement entails some semantical one. (Similarly, it seems clear that both Chisholm and Sellars would agree that some psychological statements do entail semantical ones. Thus the psychological statement

George expressed his thought that it is raining by saying that it is raining.

clearly entails, in a non-controversial sense of 'entail', the semantical statement

George said that it is raining.

where the first of these statements is to be counted as psychological because it is (at least in part) about some psychological phenomenon.)

The point of difference between Chisholm and Sellars which sentence (1) might capture would derive, therefore, from Chisholm's assertion of the first conjunct of (1):

(2) Every semantical statement entails some psychological statement.

In order to exhibit the way in which Chisholm's assertion of sentence (2) represents a point of dispute between him and Sellars, it will be important to become somewhat clearer on precisely what it seems Chisholm might have in mind in making this claim. In particular, it will be important to try to specify exactly what sense of 'entail' is required in reading sentence (2) if that sentence is to serve as a focus of one area of disagreement between Chisholm and Sellars. Thus in section II, we shall attempt to make more rigorous the claim which we are envisaging Chisholm as making. Before turning to this task, however, it will be useful to ask whether there is any other claim which Chisholm makes in his correspondence with Sellars which, like that expressed by sentence (2), may be used in order to try to refine the points of disagreement between the two.

Because of possible difficulties resulting from special uses of such terms as 'entail', and 'is to be analyzed as', Chisholm suggests in one letter that it might be a useful policy to try to state what is at issue between him and Sellars without making use of such terms. In his penultimate letter, Chisholm attempts to do this, with the apparent result that a different point of disagreement is generated. As we have seen, both Chisholm and Sellars agree that both psychological items and semantical items are intentional. It appears, however, that on Chisholm's view there is a feature of the way in which psychological items are intentional which is not shared by the way in which semantical items are intentional. Thus Chisholm writes that "thoughts are peculiar in that they have an important characteristic which nothing else in the world has," namely, that they "are a 'source of intentionality.'" [Chisholm and Sellars 533] Chisholm goes on to specify that what he has in mind in characterizing thoughts as a source of intentionality is that even though "thoughts would be intentional even if there were no linguistic entities, . . . nothing would be intentional were it not for the fact that thoughts are intentional." [Chisholm and Sellars 533]

That this claim is independent of Chisholm's claim as expressed by sentence (2), and is intended to be independent of it, is suggested by the following considerations. Chisholm remarks that his claim that "nothing would be intentional were it not for the fact that thoughts are intentional" is intended as an explication of his claim, made in an earlier letter, that "language is meaningful because it is the expression of thoughts--of thoughts which are about something." [Chisholm and Sellars 529, emphasis original] In particular, he remarks that the former claim may be regarded as providing the sense in which the term 'because', as it occurs in the latter claim, is to be understood, and he explicitly contrasts this sense with that of "the 'because' of analysis." [Chisholm and Sellars 533] Thus Chisholm seems to be making explicit his intention that the claim under consideration not be construed as a matter of analysis of expressions or of entailment relations between statements. For this reason, it seems plausible that we should accept the claim under consideration as independent of that expressed by sentence (2).

The claim, then, that thoughts are a source of intentionality, in the sense of that term provided by Chisholm, may be regarded as a second claim of Chisholm's which may serve as a focus in our attempt to formulate and evaluate the disagreement between Chisholm and Sellars. As we shall see, it will be useful to formulate this claim more rigorously before using it in an evaluation of this disagreement, but in order to have at least a preliminary formulation of the claim we may express it as follows:

(3) If there were no thoughts, then nothing whatever would be intentional, and even if there were no semantical items, thoughts would nonetheless be intentional.

II. In the foregoing section, it was suggested that there are two claims which Chisholm seems to make, against the background of which we might come to be in a position accurately to describe the points of disagreement between Chisholm and Sellars. These two claims were provisionally formulated by sentences (2) and (3), respectively. It is the purpose of the present section to discuss various respects in which it will be important to become clear on precisely what is involved in the first of these claims, before moving on to an evaluation of these points of apparent disagreement in sections III and IV.

The most crucial point with respect to which sentence (2) requires clarification concerns the precise sense of 'entail' which we are to understand as involved in making such a claim. In order to present a satisfactory construal of sentence (2), it must be understood in such a way as to meet two conditions. First, we must understand (2) in such a way as to render it plausible that Chisholm does wish to assert it, and secondly, we must understand it in such a way as to render it plausible that Sellars would wish to dissent from it. For unless we understand the claim which Chisholm is making in such a way that Sellars would dissent from it, the claim cannot be used in our attempt to capture a point of dispute between the two. In particular, then, we are faced with the following problem. We must understand (2) as involving a sense of 'entail' such that the resulting claim is one which Chisholm wishes to assert, and from which Sellars would wish to dissent. For this reason, it will be useful to begin by asking whether there is some sense of 'entail' such that if sentence (2) is understood as involving that sense, then the resulting claim is one which Sellars would concede. For no reading of sentence (2) which satisfies that condition will be acceptable for our purposes.

That there is a sense of the term 'entail' such that the claim expressed by sentence (2) is acceptable to Sellars seems clear from the fact that in attempting to get clear on the issues separating him from Chisholm, Sellars writes that he "would not for one moment wish to deny that as we use these sentences there is a legitimate sense in which 'x makes meaningful assertions' logically implies 'x has thoughts.'" [Chisholm and Sellars 535] What is important, then, for our present purposes, is to become clear on just what sense of the relevant terms is required for Sellars to assent to (2). For whatever the sense of 'entail' which results in a version of the claim under consideration which is acceptable to Sellars, it must be compatible with Sellars' insistence that it is possible that there should be a language-using community whose language while adequate to characterize semantical items in semantical terms, is nonetheless not sufficiently rich to allow talk about psychological phenomena.

In attempting to clarify the sense he has in mind, Sellars invokes what he seems to believe is a parallel situation. Conceding that "as people once used the word 'water', 'x is a piece of water' clearly did not imply 'x consists of molecules of H(2)O'," Sellars suggests that as a chemist uses the word today against the background of present scientific knowledge, the first sentence does imply the second--"in a sense of 'imply' which it is quite legitimate to call 'logical'." [Chisholm and Sellars 535-536] Given that this is so, it is the sense in which sentences about water do "logically" entail sentences about molecules which Sellars seems to have in mind in claiming that there is a sense in which sentences about non-parrotingly produced semantical items entail sentences about psychological items.

It is of course crucial for any attempt to become clear on just what sense of 'entail'{2} Sellars has in mind here to recognize that there are two cases in which the question of whether the entailment relations under consideration do or do not hold. The first case is that in which all of the expressions between which such entailment relations may hold belong to the language currently in use, and the second case is that in which some of the relevant expressions do not. Thus in connection with the example which Sellars believes to be parallel to the case involving semantical and psychological statements, there was a time when sentences containing 'water' were entirely acceptable parts of the (English) language, but the term 'molecule' simply did not exist as an English term. Similarly, if Sellars is correct, there could have been a time at which expressions characterizing semantical items in semantical terms were entirely acceptable parts of a language, but that language contained no expressions which could be used to talk about psychological phenomena. Thus it seems clear that as things stand, sentences about semantical items might all of them entail psychological sentences, even though there had been a time at which, with respect to a precursor of our current language, no semantical sentences entailed any psychological ones, for there were no psychological sentences in that language to be entailed.

These considerations might lead us to adopt the following account of the sense in which Sellars would insist on the truth of sentence (2), as well as the sense in which he would deny it. We might assert that he agrees with Chisholm in claiming that as things stand in our language, semantical statements do, in fact, entail psychological ones, but that there could have been a time at which semantical statements failed to entail psychological statements because there were none of the latter in the language, as it then stood. In order to become clear, however, on exactly what Sellars is claiming, it will be useful to get clear on (i) precisely what sense of 'entail' is involved here, (ii) why Sellars believes that as things stand in our language semantical statements do entail psychological ones, and (iii) what sorts of psychological statements Sellars believes are entailed by semantical ones. Against this background, then, we will be in a better position to formulate a precise claim which Chisholm would wish to advance.

(i) Since Sellars talks about the claim he has in mind by using the expression 'logically imply', we may offer the following as a plausible account of the sense of 'entail' which he would offer in making precise his claim. A sentence (or statement) entails another if, and only if, we would be prepared to deny that the conjunction of the first with the denial of the second could possibly be true. It is of course important to add that we would not be inclined to deny that this conjunction could possibly be true on the basis of a conviction that it is in some relevant sense meaningless, and that we would be inclined to deny that it could be true (roughly) no matter what sort of empirical discoveries which we could envisage being made. It is not intended, in advancing this account of the term 'entail', that we deny that there could be empirical discoveries which would lead us to change our attitude towards the conjunction in question, but only that we are unable (as things stand) to envisage any such empirical discoveries. Thus the account of 'entail' presented here is not intended to beg questions concerning the viability of a hard and fast distinction between sentences which are analytic or self-contradictory, on the one hand, and sentences which are synthetic, on the other. Rather the fact that we take a certain sentence (or statement) to entail another is intended to be a matter which is (at least in part) relative to a certain state of empirical knowledge. It does not seem clear that such an account will fail to do justice to sorts of claims which both Sellars and Chisholm appear to have in mind.

(ii) In attempting to become clear, given this account of entailment, why Sellars should think that given the presence in our language of both semantical and psychological statements the former entail the latter, it will be useful to return to the parallel case concerning 'water' and 'molecule' with which Sellars presented us. For in addition to the fact that, on Sellars' view, the two cases are parallel because in each case we can envisage a precursor of the language which did not contain one of the two sorts of expressions under consideration, there is another parallel between the two cases. For on Sellars' view, each case involves one of the two sets of expressions under consideration as having been introduced as theoretical expressions--as expressions under which (at least initially) were applied on the basis of theoretical considerations. Thus on Sellars' view, discourse about psychological items was introduced, as was discourse about molecules, as theoretical discourse whose application depended upon certain sorts of theoretical reasoning--the sorts of reasoning involved, respectively, in Jones' theory and in modern chemical theory. In order to become clear, therefore, on why Sellars seems to believe that as things stand semantical statements do entail psychological ones, it will be useful to see why he suggests that as things stand statements about water entail statements about molecules.

Since discourse about molecules was introduced (so we may suppose) in a theoretical context, the following sort of reasoning about such discourse seems warranted. Given that molecules are to be thought of as theoretical entities--entities, that is, the existence of which is to be thought of as having been established on theoretical grounds--it is not at all clear that we should be able to make sense of the assertion that something is water, but does not consist of molecules of one sort or other. For to talk about molecules seems to be to talk about just those sorts of things which, on the basis of certain sorts of theoretical considerations, we take it that all physical things are made up of. For this reason, it seems that the only way in which we should be able sensibly to deny that something which is water consists of molecules of some sort or other would involve denying that talk of molecules is warranted in the first place. Thus given that we assume that discourse about molecules is a warranted part of our language, it does not seem possible that we should be able to envisage empirical discoveries which, though compatible with that assumption, would lead us to accept that something could be water but nonetheless fail to consist of molecules.

There are several matters concerning this point which are important to notice. First, we are assuming that discourse about molecules consists of discourse about certain items out of which all physical objects are made up. Thus it might be argued that with the advent of recent scientific techniques, it now makes sense to suppose that talk about molecules is not talk about such building materials of all physical objects, but rather is simply talk about certain items which can be detected by the use of, say, an electron microscope. If we adopted this understanding of discourse about molecules, then it seems that we should be able to envisage empirical discoveries which would entitle us to say that, for example, no water consists of molecules. On the present line of argument, however, we would be entitled to say this only if what is involved in talk about molecules were to change, that is, only if the way in which we characteristically apply the term 'molecule' were to change.

Secondly, it is clear that these considerations do not, by themselves, allow us to conclude that sentences about water entail sentences about molecules. For if scientific discoveries were made which led us to reject talk about molecules as a "warranted" part of our language, such discourse might still retain some place in our language, and we would nonetheless be inclined to deny of something which was water that it consisted of molecules. Thus terms like 'phlogiston', 'witch' and so forth, which no longer make up a warranted part of our language in the sense that we no longer regard ourselves as entitled to apply them to anything whatever, still do occur in our language.

Given these qualifications, however, we do seem entitled to draw a conclusion which is similar to the claim that, in the sense of 'entail' presented above, sentences about water entail sentences about molecules. For given that (a) discourse about molecules is a part of our language, (b) the basis on which we apply discourse about molecules is as we have supposed, and (c) we rule out from consideration any empirical results which would lead us to reject such discourse as it is currently applied, it does seem that we would be unwilling to accept that a certain thing is water and at the same time deny that it consists in molecules.

It seems reasonable that on Sellars' view we should be able to draw a parallel conclusion concerning statements about semantical items and psychological statements. Thus given that (a) discourse about both sorts of items is present in our language, (b) the way in which we are able to apply such discourse is as Sellars envisages it to be, and (c) we rule out any considerations which would lead us simply to reject discourse about thoughts as unwarranted, it seems that we would be unwilling to accept that a certain person non-parrotingly produces a semantical item and that that person nonetheless has no thoughts. For discourse about thoughts was, as Sellars envisages it, introduced in such a way as to provide that thoughts are just the sort of thing which are causally connected to occurrences of items of intelligent behavior, and in particular, to non-parrotingly produced linguistic items. In order to have an abbreviated way of talking about this claim, we may introduce the following terminology. A sentence (or statement) will be said to entail (with qualification) another if, and only if, we would be unprepared to accept as true the conjunction of the first with the denial of the second subject to the three sorts of qualification we have considered above.{3} On this basis, we may formulate the claim which Sellars would wish to advance as

(4) Every semantical statement entails (with qualification) some psychological statement.

(iii) The third question which we mentioned above concerned just what sorts of psychological statements Sellars might wish to assert are entailed by semantical ones. For our argument above was sufficient to suggest that on Sellars' view every semantical statement would entail (with qualification) a psychological statement to the effect that someone has some thought, just as every sentence about water would entail (with qualification) some sentence about molecules, tout court. The example which Sellars uses concerning 'water' is somewhat more specific than this, however, for he speaks of sentences like 'x is a piece of water.' logically implying not merely sentences like 'x consists of molecules.' but rather sentences like 'x consists of H(2)O molecules.' as well. This suggests that there might be a sense in which Sellars believes that semantical statements entail psychological ones which are more specific than simply a psychological statement to the effect that someone has some thought. What sorts of considerations might lead one to accept the view that such specific "entailments," in some sense of 'entailment', do in fact hold?

It will be useful in considering this question to begin, as we did above, by a consideration of the case involving water and H(2)O molecules. It seems clear that we will not be warranted in claiming that sentences about water entail (with qualification) sentences about H(2)O molecules as well as sentences simply about molecules of some sort or other. For even granting that discourse about molecules is a warranted part of our language, and that to talk about molecules is to talk about items which are constituents of all physical objects, it might be discovered, say, that there are some sorts of substance which while exhibiting all the relevant properties which we would require in order to count something as water, nonetheless does not consist of H(2)O molecules but of molecules of some other sort. For it seems clear that such a discovery, while incompatible with certain generally accepted scientific beliefs, is nonetheless compatible with the idea that the general outline of molecular theory is accurate. Thus it seems that we should need stronger assumptions if we are to justify an insistence on denying the possibility that we should count something as water which, although consisting of molecules of some sort, does not consist of H(2)O molecules. One assumption which would seem adequate to justify such an insistence would be the assumption that the well established results of current chemical theory are correct. Thus we might describe the situation as follows. Given the correctness of the well established results of current chemical theory, it is ruled out that we should ever count something as water which fails to consist of molecules of H(2)O.

A parallel situation seems to obtain with regard to the case of semantical and psychological statements. For a case of a person candidly and non-parrotingly uttering a .It is raining.. (to use the dot quote convention proposed by Sellars) would rule out the possibility that that person did not have the thought that it is raining, given the truth of Jones' theory. For it is a part of Jones' theory that the thought which is causally related to occurrences of candid and non-parrotingly utterances of a given linguistic expression should be truly characterizable as exhibiting an intentional property which is analogous to that exhibited by the linguistic expression. Thus given that we accept Jones' theory as true, we would be prepared to deny that in any case where a person candidly and non-parrotingly produces a certain semantical item, that item fails to be accompanied by an occurrence of a thought episode which exhibits analogous intentional properties.

If, on the other hand, we are to characterize discourse about psychological phenomena as a sort of theoretical discourse, specifying as Sellars does, that the theoretical entities in question are causally related to items of overt behavior (including overt linguistic behavior), then it would seem that it must be possible, in principle, to discover, say, that there are occasions on which candid, non-parrotingly utterances of a .It is raining.. are not accompanied by thought episodes playing the analogous intentional role. For unless we were in principle able to discover such occurrences, the claim that semantical episodes are accompanied by such thought episodes would appear not to be a statement of a causal relation, but rather a statement concerning what we mean by characterizing a thought as playing a certain intentional role. Thus it would seem possible that a certain person should, say, be given a certain drug which results in his non-parrotingly and candidly uttering linguistic items when he has the thought corresponding to the denial of what he utters. Given our ability to make non-inferential reports on our own thoughts, we should thus be able to identify a certain thought as exhibiting certain intentional properties independently of the overt verbal utterances with which it is, at some given time, causally connected.

If the foregoing arguments are correct, then we would seem to be warranted in assuming that something like Jones' theory would need to be assumed if we were to regard it as impossible that a certain person should candidly and non-parrotingly make a certain utterance, and that this event should be unaccompanied by the corresponding thought episode. In order to have a convenient way of expressing matters of this sort, we may adopt the following terminology. A sentence (or statement) will be said to entail (relative to a certain theory, T), another if, and only if, granted the truth of T, we would be unwilling to accept as true the conjunction of the first with the denial of the second. Using this terminology, we may formulate a claim which, in addition to that expressed by (4), Sellars would appear to accept:

(5) Every semantical statement entails (relative to Jones' theory) a psychological statement about a psychological item which exhibits intentional properties analogous to those exhibited by the semantical statement in question.

Thus it would seem that, granted a different sense of 'entail', Sellars would wish to claim not only that every semantical statement entails some psychological statement, but that what psychological statement is entailed can be specified in terms of the intentional properties of the psychological item in question.

Sentences (4) and (5), then, may be accepted as a formulation of the senses in which Sellars would be prepared to assert that semantical statements "entail" psychological statements, for some appropriate sense of 'entail'. Our concern to become clear on what sense there is in which Sellars could plausibly be seen as asserting that every semantical statement entails some psychological statement stemmed from our concern to see whether the fact that Chisholm seemed to wish to advance such a claim represented a point of dispute between him and Sellars. Thus it will be important, in settling this matter, to see whether Chisholm's claim, which we provisionally formulated by sentence (2), amounts to anything beyond what is expressed by sentences (4) and (5). For unless it does, it is difficult to see how such a claim, on Chisholm's part, could result in any issue separating him from Sellars. Before turning to an examination of this question. however, it will be useful first to become clear concerning the question of whether the claim of Chisholm's which we provisionally expressed by sentence (3) results in a point of dispute between the two. For we shall see that becoming clear on this latter point will be helpful in determining what it is which Chisholm may have had in mind in claiming that every semantical statement entails a psychological one.

III. The second of the two claims of Chisholm's which, it was argued in section I, might profitably be used in attempting to come to terms with the issues separating Chisholm and Sellars, was the claim that thoughts, unlike semantical items, are what Chisholm calls a source of intentionality. This second claim was provisionally expressed by sentence (3), as follows:

(3) If there were no thoughts, then nothing whatever would be intentional, and even if there were no semantical items, thoughts would nonetheless be intentional.

There are several respects in which this formulation can be made more rigorous, however, before we turn to the question of whether this claim of Chisholm's results in a point of disagreement between him and Sellars.

In the letter in which Chisholm formulates his claim that thoughts are a source of intentionality, Chisholm also remarks that while both psychological and semantical items are intentional, nothing else is. [Chisholm and Sellars 533]{4} As a consequence, when Chisholm goes on to claim that "nothing would be intentional were it not for the fact that thoughts are intentional," he seems to be asserting that if there were no thoughts, then the physical objects which, as things stand, are semantical items, would not be intentional. Thus when Chisholm writes, making his point more specific, that "if there were no thoughts, linguistic entities would not be intentional," we cannot formulate that claim by using our terminology to say that if there were no thoughts, then semantical items would not be intentional. For as we are using the expression 'semantical item', that expression can apply to all and only those items which, as things stand, are semantical items, with the qualification that whereas to say of something that it is a semantical item involves saying of it that it exhibits intentional properties, to say of it that it is a sign design is, on this usage, neutral with regard to the question of whether it exhibits any such properties. Thus this part of Chisholm's claim may be formulated as follows:

(6) If there were no thoughts, then sign designs would not be intentional.

A second point of clarification concerning the claim provisionally expressed by sentence (3) concerns the assertion which Chisholm wishes to make in writing that "thoughts would be intentional even if there were no linguistic entities." As we have seen, it is a part of Sellars' view that "intentionality is a necessary feature of thoughts, it being absurd to say of anything that it is a thought but lacks intentionality." [Chisholm and Sellars, emphasis original] For since, on Sellars' view, thoughts are to be conceived of as modeled on semantical items in respect of the intentional properties of the latter, if anything is to count as a thought, it must exhibit some intentional properties. Since this feature of Sellars' position seems accurately to reflect our intuitive views concerning the nature of psychological phenomena, it seems reasonable to believe that Chisholm would accept this claim as well. Given that we accept that Chisholm would agree that it is necessarily the case that thoughts are intentional, we seem to be in the following position with regard to the claim under consideration. For in making that claim, Chisholm would seem to be urging, in effect, that something which is necessarily the case (namely, that thoughts are intentional) would be the case even if something which is contingently the case (namely, that there exist semantical items) were not the case. It does not seem unreasonable, however, to suppose that Chisholm may wish to be making a stronger claim than this. Thus, in particular, it seems reasonable to suppose that Chisholm might be claiming that even if there were no semantical items, there might nonetheless be thoughts. Invoking the terminology adopted in the foregoing paragraph, we might, finally, so revise this claim as to make it parallel to that expressed by sentence (6):

(7) Even if sign designs were not intentional,{5} there might nonetheless exist thoughts.

If the foregoing arguments are correct, then we may use the conjunction of sentences (6) and (7) in order to express the claim which Chisholm has in mind by writing that thoughts are a source of intentionality instead of our initial, provisional formulation provided by sentence (3). For convenience in dealing with this claim, then, we may label the entire claim as follows:

(8) If there were no thoughts, then sign designs would not be intentional, and even if sign designs were not intentional, there might nonetheless exist thoughts.

Our question, then, is whether Chisholm's claim, as expressed by sentence (8), is incompatible with anything which Sellars would wish to assert.

Let us begin by considering the first conjunct of (8), namely, sentence (6). Given Sellars' views, then if we are able to talk about thoughts, we are able to talk about a kind of item which, having been introduced as a theoretical entity, is to be thought of as having been introduced on the model of semantical items in virtue of the intentionality of the latter. Thus if there were no semantical items, that is, if there were sign designs but none were intentional or if there were no sign designs, then we could not have introduced discourse about thoughts in the way in which Sellars supposes that we could have. From this, it would seem to follow that if there were no semantical items, then we could not have begun to talk about thoughts in the way Sellars envisages we could have. It does not follow from this, however, that if there were no thoughts, then sign designs would either not be intentional or not exist at all. Thus it does not seem that without some additional assumptions, we should be able to defend the truth of (6) given the nature of Sellars' views on psychological discourse.

Given the nature of Jones' theory, however, the matter is somewhat different. From the fact that George non-parrotingly utters (or writes) an expression which is a .It is raining.. , together with the truth of Jones' theory, we can infer that George had some thoughts. If, for example, George is imagined to speak (or write) candidly and sincerely, then on the basis of Jones' theory we can infer that he had a thought which (to use Sellars' diamond quote convention) is a <>It is raining.<>. For a part of what Jones' theory provides is that a non-parrotingly produced semantical item is the causal culmination of a process which begins with some appropriate thought episode.

Let us suppose, then, that there are no thoughts. From this supposition, together with the supposition of the truth of Jones' theory, we can infer that if there are any sign designs, then they are not meaningful (in the sense of being non-parrotingly produced). For since Jones' theory provides that all non-parrotingly produced sign designs are the causal culmination of processes initiating with thoughts, if there are no thoughts, then there are no non-parrotingly produced sign designs. It is reasonable, moreover, to suppose that if an utterance or an inscription is not meaningful, in the sense of not being non-parrotingly produced, then it is not an intentional item. For such a sign design would not be produced as tokening a properly formed expression in a language, nor would it be properly understood as tokening such an expression (although it might, of course, mistakenly be so understood). For this reason, such a sign design would not properly be described as being about anything, and thus it would not be properly described as being intentional. From these considerations it seems to follow that from the truth of Jones' theory together with the supposition that there are no thoughts, we may infer that whatever sign designs there might be, none would be intentional.

If these considerations are correct, then we would seem to be warranted, using the terminology introduced in section II, in expressing the matter by asserting that

(9) There are no thought episodes.

entails (relative to Jones' theory)

(10) If there are any sign designs, then none are intentional.

Given, moreover, that acceptance of Jones' theory is a part of Sellars' position, it would seem that Sellars would accept the truth of sentence (6), namely, the claim that if there were no thoughts, then if there were any sign designs they would fail to be intentional. It is not clear, then, that Chisholm's insistence on the truth of the first conjunct of (8) represents a substantive point of dispute between him and Sellars.

Let us turn, then, to a consideration of the second conjunct of (8), that is, of sentence (7). It does not seem that considerations parallel to the ones we have used to defend the truth of (6) on the basis of Sellars' position fare any differently here. For it does not seem that we should be warranted, given the supposition that sign designs are not intentional and the supposition that Jones' theory is true, in inferring that there are no thoughts. For to accept that we should to be warranted in so concluding would be tantamount to asserting that the truth of Jones' theory rules out the possibility that there could be living things which had thoughts but nonetheless never expressed their thoughts.

There is, however, a consideration which seems to present a difficulty here. If Sellars is correct concerning the way in which discourse about thoughts is to be thought of as having been introduced, then if sign designs were not intentional, it is difficult to see how discourse about thoughts could ever have begun. For on Sellars' view, the introduction of discourse about thoughts depended on the idea that Jones should have been able to model his theoretical items on other items which were intentional. Thus given Chisholm's conviction that nothing except psychological and semantical items are, in fact, intentional, if it were not for the fact that there existed intentional sign designs (that is, semantical items), it is difficult to see how discourse about thoughts could ever have begun, given that it did begin in the way in which Sellars supposes.

These considerations do not, however, appear upon closer examination to present reason for Sellars to disagree with the truth of Chisholm's claim as expressed by sentence (7). For as we have seen, it is a part of Sellars' view that pre-Jonesean people did "think, but they don't know that they think." [Chisholm and Sellars 526, emphasis original] Thus what is ruled out by the considerations presented in the foregoing paragraph is not the possibility that there should be thoughts but no intentional sign designs, but rather that we should have any linguistic means for talking about thoughts if there were no intentional sign designs. Thus it does not seem to be the case that Sellars has grounds for ruling out the possibility that there should be beings who could be determined, say, by us, to have thoughts, in spite of the fact that they do not express these thoughts in language. Thus it seems possible, given Sellars' position, that there should be beings who, in virtue of extremely complicated non-verbal behavior we should wish to say have thoughts, in spite of the fact that they use no spoken or written language.

If the foregoing arguments are correct, then it seems that Sellars would not wish to deny the truth of the claim, advanced by Chisholm, that thoughts are, in a certain sense, a source of intentionality. The only error which, from the basis of Sellars' position, we might wish to impute to Chisholm in this context is that in making his claim that thoughts are a source of intentionality, he does not make explicit the sorts of assumptions which would be required to make the claim defensible. Thus a less misleading way of putting the point, as it has been defended above, would be to claim that thoughts are a causal source of intentionality, thereby making explicit the need to be prepared to accept some account of the causal relations between thoughts and semantical items if one is to be warranted in asserting sentence (8). Correspondingly, the strength of Sellars' position in this context would seem to lie in the fact that he attempts to make explicit just what sorts of factual truths one would wish to appeal to in defending the truth of sentence (8). For it is not clear how, aside from considerations of the sort invoked above, one might go about trying to defend sentence (8) as true.

The only difficulty which this account of the possible point of dispute between Chisholm and Sellars represented by Chisholm's assertion of (8) would seem to face, would be the possibility that Chisholm had in mind to claim something stronger than what we have seen Sellars would assent to. But it is not clear what stronger claim Chisholm might wish to advance. If one regarded the truth of Jones' theory as open to serious question, then one might wish to claim that some defense of the truth of (8) could be given even though Jones' theory might be false. Thus one account of why Chisholm seems to feel that his assertion that thoughts are a source of intentionality conflicts with Sellars' views might be offered on the basis of Chisholm's conviction that (8) should be regarded as defensible whether or not that aspect of Sellars' account which involves Jones' theory is accepted. In one respect it seems clear that Sellars would agree with such a conviction. For it seems clear that Sellars does not present his "myth" involving Jones' theory as the only kind of adequate account which could be given of the nature of psychological discourse, but rather as one kind of account which is, in fact, adequate. What is important is that Sellars would wish to assert that if claims like that made by sentence (8) are to be defensible, then it must be on the basis of factual and causal considerations of the sort involved in his account. But it does not seem clear that there would be any reason to believe that Chisholm would dissent from this claim.

IV. Accepting the foregoing as an adequate evaluation of the positions which Chisholm and Sellars would take with regard to the claim expressed by sentence (8), we may turn in this fourth section to examine the question of whether there is any dispute between the two resulting from Chisholm's assertion of sentence (2). As we have seen in section II, Sellars is prepared to concede the truth of sentences (4) and (5). For this reason, our problem in the present section is to determine whether, given that Sellars is prepared to make such a concession, there is any point of dispute between the two resulting from Chisholm's insistence on the truth of (2). In particular, it will be important to determine whether in asserting (2), Chisholm has in mind to assert anything stronger than (or different from) what is asserted by sentences (4) and (5). For if he does not, then it does not appear that there is a substantive dispute involved in his assertion of (2).

Before turning to this question, however, there is a preliminary consideration which it will be useful to raise. In discussing the question of whether there is a sense in which Sellars would accept that every semantical statement entails some psychological statement, we have seen that there are two ways in which one might be prepared to advance such an assertion. The first way would involve simply that, for some sense of 'entail', every semantical statement entails some psychological statement of some sort or other; while the second way would involve there being a sense of 'entail' such that every semantical statement entails some particular sort of psychological statement, As we have seen, Sellars accepts such an assertion made in both ways, although the sense of 'entail' which it was necessary for us to presuppose differed between the two cases. Thus our formulation of what Sellars would be prepared to accept involved both sentences (4) and (5).

If we simply view Chisholm as claiming that (in some sense of 'entail') every semantical statement entails some psychological one, then this claim is neutral as between the first sort of claim, and the second. It appears, however, that there is reason to believe that Chisholm would wish to defend the stronger of the two sorts of claims. Thus Chisholm writes that "we cannot describe . . . [a person's] use of language, or his dispositions to use language in those ways, unless we refer to what he believes, or knows, or perceives." [Chisholm (17) 517] And again: "if the people of your [Sellars'] were to give just a little bit of thought to the semantical statements they make, wouldn't they then see that these semantical statements entail statements about the thoughts of the people whose language is being discussed?" [Chisholm and Sellars 537, emphasis supplied] In these places and in others, Chisholm seems to wish to assert that not only do semantical statements entail (in some sense of 'entail') statements to the effect that certain people have thoughts, but also statements about the particular thoughts which they have. Thus it seems that it is only the stronger of the two sorts of claims which will satisfy what Chisholm has in mind.

With these considerations in mind, we may return to our consideration of whether the claims expressed by sentences (4) and (5) accurately reflect the sort of claim to which Chisholm seems committed. There is one respect in which it seems clear that Chisholm might not find these two claims acceptable. For sentence (4) rests upon a particular feature of Sellars' account concerning the nature of discourse about thoughts, namely, the idea that given that we have such discourse as Sellars envisages it, the items which such discourse is about are items which accompany any occurrences of non-parrotingly produced linguistic items. Similarly, sentence (5) rests upon the notion that a certain sort of theoretical reasoning, namely, that involved in Jones' theory, be accepted as accurate in connection with our beliefs about thoughts. If Chisholm were to deny the adequacy of these features of Sellars' account of psychological discourse, then he would not be also in a position to wish to insist upon the truth of (4) and (5), as they stand.

As we have seen, Chisholm wishes not only to assert that (for some sense of 'entail') every semantical statement entails some psychological statement, but, more strongly, that every semantical statement entails some psychological statement about a psychological item exhibiting a particular set of intentional properties. Since the former sort of claim will surely be true if the latter sort is, we can conveniently restrict our discussion at this stage to a consideration of the sense in which Chisholm would want to urge the truth of a statement of the latter sort. For this purpose, it seems that we can easily revise sentence (5) in such a way that the revised version does not obviously conflict with anything which Chisholm would wish to maintain:

(5') Every semantical statement entails (relative to some appropriate theory, T) a psychological statement about some particular psychological items characterizing them in terms of their intentional properties.

It seems clearly possible for one to assent to the truth of our revised sentence (5') without thereby being committed to any position with regard to the particular theory which Sellars defends under the label 'Jones' theory'. Since, moreover, it seems clear that if Sellars accepts (5), he must also accept (5'), our question is whether Chisholm would accept (5'), and whether Chisholm would wish to assert anything stronger.

The only way in which it seems obviously suggested that we might wish to attribute to Chisholm a claim stronger than that expressed by (5') is as follows. It might be thought that Chisholm wishes to assert simply that

(11) Every semantical statement entail (tout court, not relative to a theory and not with qualification) some psychological statement about some particular psychological items characterizing them in terms of their intentional properties.

That is, it might be thought that Chisholm believes that the conjunction of a sentence like

(12) George (non-parrotingly and candidly) uttered a .It is raining...

with the denial of a correspondence sentence like

(13) George has a thought which is a <>It is raining.<>.

must be rejected as false, no matter what, and independently of our acceptance or rejection of some causal theory, or of other matters of fact. If Chisholm were to be regarded as urging the truth of this sort of claim, then, it seems clear, he would not only be urging something stronger than the claim expressed by (5'), but also something which Sellars would find unacceptable. For independently of the sorts of considerations which are involved in Sellars' account of psychological discourse, it does not seem that there is any reason to insist that the conjunction of (12) with the denial of (13) must be false.

The only support which, it would seem, would warrant our attributing to Chisholm so strong a claim would be the fact that he does not, in advancing his views, spell out any considerations which would be invoked in defending the claim he has in mind. Thus it might seem tempting, in the absence of any such considerations, to attribute to Chisholm a claim which does not need to be defended, as Sellars' claim does, by recourse to some account of the nature of psychological discourse such as that embodied in Jones' theory. In particular, since Chisholm does not make any reference to considerations of a factual or causal sort which could be used in defending his claim, it may be tempting to suppose that he neither believes that any such considerations would be needed nor that they would be helpful in defending such a claim. Thus one would picture Chisholm as asserting something which is entirely independent of such considerations. |

It does not seem, however, that these sorts of considerations are particularly compelling. For if the account of Chisholm's views concerning the truth of sentence (8) which was presented in the foregoing section are correct, then it does not seem unreasonable to suppose that Chisholm should have in mind a claim which, although he nowhere makes it explicit, must be defended by recourse to empirical considerations such as those embodied in Jones' theory. For in that section we found it reasonable to suppose that Chisholm's assertion of the truth of sentence (8) could plausibly be regarded as a claim which, although it did not make this explicit, required to be defended by reference to some such considerations. In the absence, then, of any more explicit account, by Chisholm, of the nature of the claim which he wishes to make, we may reasonably regard it as being adequately expressed by sentence (5'). In any event, it does not seem that there are any adequate grounds for supposing that Chisholm had in mind a claim which is either stronger than, or different from, that expressed by sentence (5').

If the foregoing arguments are correct, then it would seem that an evaluation of the matter parallel to the evaluation, presented at the close of section III, of Chisholm's claim that thoughts are a source of intentionality, would be warranted. For just as we concluded that the difficulty, from Sellars' point of view, with Chisholm's claim that thoughts are a source of intentionality was that it was expressed in such a way as to fail to make explicit the sorts of considerations which would be required to defend its truth, so the same criticism can be leveled against Chisholm's claim that (for some sense of 'entail') every semantical statement entails some psychological statement. Moreover, just as it was suggested that Chisholm's feeling that his claim that thoughts are a source of intentionality might represent a point of dispute between him and Sellars could be accounted for on the basis of his conviction that such a claim should be defensible independently of the truth of Jones' theory, so here a parallel remark may be made. For it seems plausible to suppose that Chisholm's feeling that his claim that (in some sense of 'entail') every semantical statement entails some psychological one represents a point of dispute with Sellars might be accounted for on the basis of a conviction that this claim also should be defensible independently of the truth of Jones' theory. Thus the revision of sentence (5) to (5') was suggested, in order to remove this difficulty. So long as the claim which Chisholm wishes to advance is a claim to which considerations of a factual and causal nature along the lines of Jones' theory are relevant, however, it does not seem that Chisholm's claim represents a substantive disagreement with anything which Sellars would wish to insist upon.

Before closing it may be useful to raise one final sort of consideration in connection with Chisholm's claim that every semantical statement "entails" some psychological one. The example of a semantical statement which we used in our argument above represents a case of a particular speech act, namely, the act of George saying or writing that it is raining. There is another sort of semantical statement which is not only of interest to Sellars, but to Chisholm as well, namely, what we have seen Sellars call the 'means'--rubric. An example of such a statement would be expressed by

(14) 'It is raining.' means that it is raining.{6}

Since sentences like (14) are clearly semantical sentences, it may be reasonably asked what psychological sentence Chisholm would believe that (14) entails, for a suitable sense of 'entail'.

The psychological sentence which, it seems reasonable to suppose, would be "entailed" by sentence (14) might plausibly be expressed as

(15) Whenever 'It is raining.' is non-parrotingly and candidly uttered or written, it expresses a thought that it is raining.

It is worth noticing, however, that as Sellars points out, sentences like (15) are subject to a certain sort of confusion. Thus there seem to be two assertions which we might wish to make in saying that a certain sign design expresses a thought. Thus Sellars writes that "in one sense, to say of an utterance that it expresses a thought is to say, roughly, that a thought episode causes the utterance." [Sellars (8) 655] In this sense, sentence (15) would be understood as asserting that a thought episode which is a <>It is raining.<> is causally linked to non-parrotingly and candidly produced occurrences of the semantical item 'It is raining.'. The other sense which might be given to sentences like (15) would result in its being an assertion that non-parrotingly and candidly produced occurrences of 'It is raining.' express the proposition that it is raining, which is in effect to say no more than that they mean that it is raining. [Sellars (8) 655-656] Sellars draws this distinction by stipulating that the first sense of 'express' be called the "causal" sense of 'express', and the second the "logical" sense, and he further distinguishes between the two senses of 'thought' as thinkings and propositions, respectively. [Sellars (8) 656]{7} (Continuing the usage adopted in the foregoing chapter, we shall continue to use 'thought', 'thought episode' and 'thinking' to apply to what Sellars here calls thinkings, and 'proposition' in the way suggested by Sellars.)

Given this distinction, it seems clear that (15), as it stands, is ambiguous as between

(16) Whenever 'It is raining.' is non-parrotingly and candidly uttered or written, it causally expresses a thought episode which is a <>It is raining.<>.

and

(17) Whenever 'It is raining.' is non-parrotingly and candidly uttered or written, it logically expresses the proposition that it is raining.

Now it seems clear that sentence (14) entails (tout court, without qualification and not relative to any theory) sentence (17). For by hypothesis, sentence (17) is simply a different way of expressing what is expressed by (14). If one were to understand, then, Chisholm's claim that every semantical statement entails (in some sense of 'entail') some psychological statement as providing that sentences like (14) entail sentences like (17), then it seems clear that the sense of 'entail' would not have to be that of 'entail (relative to a theory)' which was invoked in formulating sentence (5') above. But it also seems clear that to regard Chisholm as making such a claim would be mistaken if we are to take seriously the idea that Chisholm is urging some sort of "entailment" as holding between semantical statements and genuine psychological ones. For sentence (17) is clearly not about any psychological phenomenon, and thus cannot be counted as a psychological sentence.

It seems clear, therefore, that Chisholm's claim must involve the assertion that sentences like (14) entail sentences like (16). But if this is so, then it seems that the conclusions which were reached concerning the sort of "entailment" which Chisholm can reasonably be viewed as believing to hold between sentences like (12) and those like (13) will apply to the present case as well. For it does not seem that there is any reason to believe that Chisholm would wish to urge that sentences like (14) entail sentences like (16) in any sense stronger than that provided for in sentence (5').

V. In our attempt to evaluate the nature of the dispute presented in the correspondence between Chisholm and Sellars, we have examined two claims urged by Chisholm which seem, in the context of the correspondence, to be intended to represent points of issue between the two. In examining what is plausibly regarded as being involved in each claim, however, we have argued that it does not seem clear that what Chisholm wishes to assert conflicts with anything in Sellars position.{8} In view of this conclusion, however, it is reasonable to expect that there should be some explanation of why, if Chisholm makes no important claim with which Sellars would disagree, the correspondence between the two appears to have reached so inconclusive a result. There seem to be three principal reasons for this. The first, which has been implicit throughout our discussion of the nature of the claims which Chisholm wishes to make, is that Chisholm does not seem to be explicit concerning the nature of his claims. Thus while it seems that he is at great pains to formulate his assertions as unambiguously as possible, he nonetheless does not, as Sellars does, make clear what sorts of considerations would be involved in the attempt to defend such claims. As we have seen, much of Chisholm's convictions concerning the nature of psychological discourse seem to stem from his convictions concerning the possibility of formulating a definition of intentionality, which, in conjunction with a suitable thesis of intentionality, will succeed in presenting a certain sort of characterization of psychological discourse. It does not seem, however, that Chisholm has succeeded in formulating such a definition and such a thesis. Moreover, it is not clear that even had Chisholm succeeded in that attempt, the resulting definition and thesis would be relevant to a defense of the claims considered above. For even if Chisholm could establish some logical characteristic which was exhibited by all and only psychological sentences, such a result would seem neutral as regards the truth of his claims examined in the present chapter and the denials of these claims. Since, moreover, Sellars does not dispute the possibility of presenting such a logical characteristic, it does not seem that invoking such a characteristic would be relevant to defending claims which are intended to be at odds with one or another aspect of Sellars' position.

The second reason for the apparent inconclusiveness of the correspondence is somewhat more complicated. Thus if what has been argued in sections III and IV is correct, then it seems that one possible source of Chisholm's resistance to accepting Sellars' position would be his conviction that the claims he advances should be defensible independently of the sort of account of psychological discourse which Sellars presents, involving, as it does, Jones' theory. It is not clear that even if the attribution of this conviction to Chisholm is correct, that Chisholm would therefore wish to dispute the truth of Sellars' account of psychological discourse; what does seem clear is that he is unhappy about resting a defense of the claims he advances on such an account. Thus it seems clear that he would like to advance these claims as acceptable independently of the question of whether we are to find such an account as Sellars offers also acceptable.

As we have seen, however, it is not clear just what sort of defense Chisholm would offer for his claims, if it is not at least along the general lines of the account of psychological discourse sketched by Sellars. In particular, it seems, on the basis of the arguments presented above, that if any satisfactory defense of such claims is to be presented, it will have to be along the lines of that offered by Sellars. That is, a satisfactory defense of the sort of claims which Chisholm urges would seem to involve a sketch of psychological discourse as about items which are, like semantical items, intentional, and which are causally related in a suitable way both to occurrences of semantical items and to other sorts of items of overt behavior. But it is just these sorts of considerations which typify the account of psychological discourse which Sellars wishes to defend, and on the basis of which Sellars would be prepared to accept claims of the sort which Chisholm advances.

At different points in his letters to Sellars, Chisholm seems to adopt different positions concerning what might be involved in defending such claims. Thus in writing that if pre-Jonesians "were to give just a little bit of thought to the semantical statements they make, wouldn't they then see that these semantical statements entail statements about the thoughts of the people whose language is being discussed?", [Chisholm and Sellars 537, emphasis supplied] Chisholm seems to be suggesting that it would not be necessary, in defending the claims he wishes to advance, to engage in a discussion of considerations of the sort involved in Sellars' account of psychological discourse. For in this passage, Chisholm seems to be suggesting that independently of the theoretical reasoning introduced by Jones, not only could we introduce discourse about thoughts, but we could simply "see," upon reflection, that semantical statements "entail" psychological ones. Thus in this passage, Chisholm seems to insist that his claims should be defensible independently of any considerations whatever of the sort invoked by Sellars.

On the other hand, in discussing his claim that thoughts are a source of intentionality, Chisholm writes in such a way as to suggest that perhaps causal and factual (and perhaps even theoretical) considerations of the sort invoked by Sellars might be relevant to, and required for, a defense of his claim. Thus in discussing a figure proposed by Hospers, Chisholm likens thoughts as sources of intentionality to the sun as a source of light, and semantical items as, presumably, "recipients" or "reflectors" of intentionality to the moon as a recipient and a reflector of light [Chisholm and Sellars 524, 533],and metaphorically speaks of "the astronomical truth" that thoughts are a source of intentionality. [Chisholm and Sellars 533] Thus Chisholm seems to suggest that perhaps factual matters of a certain sort would need to be invoked if we are to defend the claim in question, and thereby allows for the possibility that the factual matters should be of the sort which Sellars has in mind.

The third reason for the inconclusiveness of the correspondence seems to be that Chisholm, while seeming to insist upon the idea that knowledge about thoughts increases our knowledge about the nature of semantical items, does not appear to recognize that Sellars would clearly concede this point. For the fact that thoughts and semantical items are, on Sellars' view, not only causally connected among themselves but causally connected to non- linguistic items of overt behavior, certainly would be seen by Sellars as increasing our understanding of the nature and behavior of semantical items. In particular, since on Sellars' view semantical items along with items of intelligent non-linguistic behavior are culminations of causal processes initiating with thoughts, knowing that semantical items are related in these ways to thoughts certainly increases our understanding of the role which semantical items play in our general pattern of behavior.{9} But it would seem, on Sellars' view, that it is just the fact that semantical items do play such roles that allows us to characterize them as intentional. Thus understanding the intentional nature of semantical items must surely be facilitated by knowledge about thoughts.

Chisholm concedes the possibility that there should be a community of language-users whose language contains discourse suitable to in some way characterize semantical items but does not contain psychological discourse, but claims that this has no bearing on the "astronomical truth" that thoughts are a source of intentionality, and that "a little bit of thought" would suffice for speakers of such a language to see that their semantical sentences entail psychological ones. What he seems chiefly to be concerned to defend, here, is the idea that one cannot have an adequate understanding of the nature and behavior of semantical items unless one has knowledge about thoughts, and about the relationship between thoughts and occurrences of semantical items. But he does not seem to recognize that Sellars would concede, at least, that we cannot have as full an understanding of the nature and behavior of semantical items as we, post-Joneseans, do, unless such knowledge concerning thoughts is had.

If to have an "adequate" understanding of the nature and behavior of semantical items is to have (roughly) the degree and sophistication of understanding of such items that we have, Sellars would certainly concede Chisholm's point. But if Chisholm has in mind some other account of what it would be to have an "adequate" understanding, he has failed to make that clear. For the one obvious possibility for what Chisholm might claim, namely, that unless one has knowledge about thoughts and their connection with semantical items one cannot even characterize the latter in terms of their intentional properties, seems to be ruled out by Chisholm's concession that it is possible that a community of language-users should have a language which although including semantical locutions does not have any locutions by means of which its speakers might talk about thoughts. Thus Chisholm's lack of clarity concerning the sense in which he might wish to claim that without knowledge of thoughts one cannot have a "full" understanding of the nature and behavior of semantical items, combined with the resulting lack of recognition that there is a sense in which Sellars would agree with this claim, seems also to contribute to the inconclusiveness of his correspondence with Sellars.

It is worthwhile remarking, in conclusion, that there is at least one sort of view which might be suggested by Chisholm's insistence that if the people of Sellars' myth were to give "just a little bit of thought to the semantical statement they make" they would then "see that these semantical statements about the thoughts of the people whose language is being discussed," which would be antipathetical to Sellars' position. This view is that the fact that thoughts and semantical items exhibit analogous intentional properties is a fact which can be seen simply by examining items of both sorts. Thus Sellars writes that one might "interpret the analogy between thoughts and speech as a discovered analogy, as one might discover analogies between colors, sounds and tastes" [Sellars (6) 103, emphasis original], and in what follows, he makes clear his rejection of this view. [Sellars (6) 104-105] For it is a part of Sellars' position that discourse about thought episodes is (at least in part) discourse which is intended and is able to provide a certain sort of explanation of both linguistic and intelligent non- linguistic behavior, and the intentional properties of thoughts are attributed to them in connection with their role in such explanations.

For these reasons, it seems clear that Sellars would want to insist that the sense in which one might appropriately speak of discovering that thoughts exhibit intentional properties analogous to those exhibited by semantical items is the same as one of the senses in which one might be said to discover that items invoked by a theory exhibit properties analogous to items of some other sort--namely, the sense that one discovers that the theory is a good one, and the theory in fact imputes properties of that sort to the items it invokes. For on Sellars' view there is no reason to suppose that anyone would know anything at all about thoughts independently of some specification of some of the properties of thoughts on the basis of theoretical reasoning involved in framing the explanations in question. Thus, on Sellars' view, it would be mistaken to suppose that one could discover that thoughts exhibit intentional properties by having thought episodes available for inspection, and on inspection coming to see that they exhibit features analogous to the semantical properties of linguistic items.

If a view of the sort discussed in the preceding two paragraphs is what Chisholm has in mind in making the claim in question, then, it would seem clear, there is a substantive disagreement between him and Sellars. On the basis of what Chisholm writes, however, it is difficult to evaluate whether this is, in fact, what he has in mind, for it does not seem that he anywhere makes his claims as explicit as the formulation of the view which has been given above. And since he neither is explicit concerning whether such a view is what he wishes to propound, nor indicates any way in which such a view l might possibly be defended, it is difficult to see how one might go about evaluating the issue which, if Chisholm does have this view in mind, would separate him from Sellars. If Chisholm's endorsement of this view was (at least part of) what resulted in his finding Sellars' position unacceptable, then it seems clear that there could be a fruitful continuation of an interchange of the sort represented by Chisholm's and Sellars' correspondence. It seems clear that a defense of Sellars' position would rest, as sketched above, with an appeal to the idea that our knowledge of thoughts, as we talk about them by means of current psychological discourse, is dependent on the role which thoughts are conceived to play in the explanation of non-parrotingly verbal behavior and intelligent nonverbal behavior. Since, on the other hand, there seems to be no indication by Chisholm of how the view discussed above, if it were endorsed by Chisholm, would be defended, it does not seem possible to evaluate the merits of the endorsement of such a view in connection with a projected continuation of the interchange.


[Table of Contents][Go to Chapter 5]

Notes:

{1} In a letter dated November 8, 1965, Sellars suggests speaking of utterings and mark-makings instead of speaking of marks and noises as a way of avoiding this terminological difficulty. And with much the same point in mind, he writes in [Sellars (6) 151] that utterances are not simply the acoustic phenomena which occur and are produced by language speakers; rather "utterances are people uttering." [Back]

{2} Although Sellars uses the expressions 'imply' and 'logically imply' in this context, I shall discuss his claim as amounting to the claim that there is a sense of 'entail' for which the relevant entailment relations hold. [Back]

{3} The three sorts of qualification in question might be formulated, in order to be neutral with respect not only to the two cases considered above but also to other cases which we might consider, as follows. (a) Expressions contained in both sentences (or statements) under consideration must be a part of the language in question; (b) the way in which such expressions are applied is regarded as not subject to any radical changes and (c) we imagine as ruled out any possible empirical discoveries which would result in our rejecting the kind of discourse represented by either sentence (or statement) as not applying to anything whatever. It is clear that this notion of entailment (with qualification) is in many respects vague, but it seems to satisfy the purpose of allowing us to explicate the sense of 'entail' which, it might be expected, Sellars would invoke in defense of his claim. [Back]

{4} In the letter which Sellars writes in reply to that letter of Chisholm's, Sellars remarks that he believes that he and Chisholm "would both have reservations about" the claim that nothing other than psychological and semantical items is intentional, although, Sellars continues, "none that is relevant to our problem." [Chisholm and Sellars 535] What Sellars seems to have in mind here is that it is reasonable to speak of items of overt behavior which, not being part of any language are not properly speaking semantical items, nonetheless play roles analogous to those played by semantical and psychological items, as intentional. [Back]

{5} In altering this first clause from 'even if there were no semantical items' to 'even if sign designs were not intentional', it does not appear that we are doing more than expressing more rigorously what Chisholm seems to have had in mind. For it seems clear that if the condition expressed by the latter clause is satisfied, then the condition expressed by the former must be: there could be no semantical items unless the items which we are calling sign designs are intentional. Moreover, there seem to be two ways in which the former condition could be satisfied: (i) there could be no sign designs in the first place; and (ii) there could be sign design but none would be intentional. Since it seems that this latter circumstance is the one which Chisholm wishes to deal with, the alteration in question seems warranted. [Back]

{6} Chisholm does not, except in adopting Sellars' means of formulating the matter, use sentences of this type, but he seems to have this sort of sentence in mind when, in "Sentences about Believing," he expresses concern about sentences of the form 'People use such and such a word to mean so-and-so.'. [Chisholm (17) 517] [Back]

{7} Cf. [Sellars (6) 154], where the same distinction is presented. [Back]

{8} A similar conclusion seems to be reached by Bernstein, who argues that just as Sellars wishes to claim that "the categories of intentionality are nothing more nor less than metalinguistic categories in terms of which we speak epistemically about overt speech," so Chisholm might appropriately be understood as arguing that "the metalinguistic categories in terms of which we speak epistemically about overt speech are nothing more nor less than the categories of intentionality." [Bernstein 312] Accepting this as a plausible view of Chisholm's position, Bernstein concludes that it is not clear that the two positions are diverse in the way in which the inconclusiveness of the correspondence might lead one to believe. (Ibid.) Although the arguments presented in the foregoing sections seem to support Bernstein's contention, it does not seem that his conclusion is based on any sort of detailed examination of the claims the truth of which Chisholm urges in his letters. [Back]

{9} In particular, as we have seen in a footnote in Chapter III, section VI, there is one special way in which, on Sellars' view, our knowledge about the thought episodes whose existence is established on the basis of Jones' theoretical reasoning does result in an increased understanding of the role which semantical items play in our general pattern of behavior. For pre-Joneseans are, presumably, able to report noninferentially on their dispositions to engage in certain sorts of linguistic behavior, but without psychological discourse as constructed on the basis of Jones' theory, they are unable to account for or explain this ability of theirs. It seems clear, however, that this is not the only respect in which knowledge about thoughts results in an increased understanding of the nature of linguistic behavior. For not only does such knowledge provide an account of the conviction that pre-Joneseans may be supposed to have, that when they engage in intelligent nonverbal behavior this behavior would be accompanied by appropriate sorts of non-parroting verbal behavior if certain circumstances were to obtain (Cf. [Sellars (6) 153] and especially [Sellars (6) 153, fn. 8].), but it also provides an account for the fact that intelligent nonverbal behavior is (sometimes) accompanied by appropriate sorts of non-parroting verbal behavior. [Back]


[Table of Contents][Go to Chapter 5]