David M. Rosenthal,
Intentionality: A Study of the Views of Chisholm and Sellars

I: CHISHOLM'S VIEWS (1)

In this chapter, I begin my consideration of Chisholm's and Sellars' views on intentionality by a detailed discussion, leading to a rejection, of Chisholm's claims in "Sentences about Believing"--an attempt to present a logical mark which is exhibited by all and only sentences which are about psychological phenomena. This attempt, the best known of a number made by Chisholm,{1} represents a useful way of beginning an examination of this topic because one of the important ways in which philosophers have, in the context of Anglo-American philosophy, attempted to discuss what is involved in psychological phenomena is by way of discussing what peculiarities, in addition to their subject matter, locutions which are about psychological phenomena might exhibit.

In particular, an important strand of the problem of intentionality consists in the attempt to discover whether there are any important differences between psychological sentences and non-psychological sentences, independently of their differing subject matter. For, it is believed, if such independent differences can be shown to exist, then the differences between the subject matters of such sentences might arguably be different in some philosophically significant way.

This chapter will argue that there are difficulties with the claims advanced by Chisholm which require certain revisions; that if all the required revisions are introduced, then the claims, while not transformed beyond recognition, nonetheless become exceedingly complex in various unexpected ways; and that certain revisions result in our having good reason to reject any modified versions of Chisholm's claims. We shall begin, then, with an attempt to isolate the assertions which Chisholm is concerned to defend and which give rise to these sorts of difficulties.

I. In his article, "Sentences about Believing," [Chisholm (17)] Professor Roderick Chisholm advances several claims with the aim of establishing a certain connection between the set of sentences each member of which describes psychological phenomena, and the set of sentences each member of which exhibits at least one of certain logical properties. In this chapter, I shall attempt to give adequate formalizations of these claims, to examine the logical relationships among the formalized versions of the claims, and to discuss certain consequences of adopting these formalized versions.

Chisholm claims that we may identify as intentional, among simple declarative sentences, all and only those which exhibit at least one of-three logical characteristics.{2} A compound sentence, Chisholm continues, may be said to be intentional if, and only if, it has at least one simple sentence component which, in its declarative form, is intentional. A declarative sentence, then, is intentional if, and only if, it has at least one of several logical properties; but for the purpose of this paper, we may simplify this claim and say that a declarative sentence is intentional if, and only if, it exhibits a certain (disjunctive) logical characteristic.{3}

Chisholm goes on, then, to advance the following thesis about the use of intentional sentences:

Let us say (1) that we do not need to use intentional language when we describe non-psychological, or "physical," phenomena; we can express all that we know, or believe, about such phenomena in language which is not intentional. And let us say (2) that, when we wish to describe certain psychological phenomena--in particular, when we wish to describe thinking, believing, perceiving, seeing, knowing, wanting, hoping and the like--either (a) we must use language which is intentional or (b) we must use a vocabulary which we do not need to use when we describe non-psychological, or "physical," phenomena. [Chisholm (17) 512]{4}

We shall refer to the claim which Chisholm advances in this passage as the thesis of intentionality or, for short, as the thesis. And in order to distinguish the thesis of intentionality from the logical mark which Chisholm adopts in order to identify intentional sentences, we shall call the adoption of this mark the definition of intentionality or, simply, the definition. For the sake of brevity, finally, we shall call the logical characteristic by means of which Chisholm identifies intentional sentences the characteristic.

We may say that the definition of intentionality is expressed by the following sentence:

(1) All and only intentional sentences exhibit the characteristic.

Let us note, however, that Chisholm is careful not to state the thesis of intentionality in parallel fashion, as:

(2) All and only sentences which describe psychological phenomena are intentional.{5}

for it is clear that these two sentences jointly entail the following sentence:

(3) All and only sentences which describe psychological phenomena exhibit the characteristic.

But Chisholm explicitly recognizes the existence of counter-examples to this last sentence.

The particular counter-examples which Chisholm considers are the sentences

The patient will be immune from the effects of any new epidemics.

and

John is cat-perceptive.

For, as he points out, the former, although it describes no psychological phenomenon, nevertheless exhibits the characteristic; while the latter fails to exhibit the characteristic, even though it does describe a psychological phenomenon.{6} On the basis of the existence of these sentences, we may conclude that

(4) There exist sentences which describe psychological phenomena and do not exhibit the characteristic, and there exist sentences which do not describe psychological phenomena and do exhibit the characteristic.

It is clear that each conjunct of sentence (4) entails the denial of sentence (3). Given the truth of sentence (4), then, we are forced to relinquish our assertion either of sentence (1) or of sentence (2).

As we have seen, however, Chisholm does not assert either sentence (2) or sentence (3). Our problem, then, is to seek an accurate formulation of the thesis and of the definition such that our formulation, while capturing Chisholm's intended meaning, does not entail the denial of sentence (4). We shall find clues to finding such a formulation by considering Chisholm's own discussion of the counter-examples in question, although, as we shall see, certain new problems of formulation will arise.

II. It is clear that whatever Chisholm wishes to claim must be compatible with the existence of sentences which both exhibit the characteristic and fail to describe psychological phenomena. (We shall henceforth speak of sentences which describe psychological phenomena as psychological sentences, referring to all other (declarative) sentences as non-psychological sentences.) Chisholm asserts, of such sentences, that we may "readily transform them into conditionals which are not intentional." [Chisholm (17) 512] Thus, he proposes, we may transform our example on the previous page into

If there should be any new epidemics, the patient would not be affected by them.

And in general Chisholm argues,

any other ostensibly non-psychological sentence which is intentional can be transformed, in an equally obvious way, into . . . a sentence of one of two possible types: either (a) it will no longer be intentional or (b) it will be explicitly psychological. [Chisholm (17) 512]

If we are able so to transform a given non-psychological sentence which exhibits the characteristic, then, Chisholm appears to be arguing, the existence of that sentence is compatible with the claims he wishes to make.

With this in mind, let us take another look at Chisholm's formulation of the thesis. Part (1) of the thesis asserts that we need not use intentional sentences in order to describe non-psychological phenomena.{7} We might express this by saying that any non-psychological sentence which is intentional (and thus exhibits the characteristic) is expendable in favor of a non-psychological sentence which is non-intentional (and thus fails to exhibit the characteristic). This raises the question, however, as to how we are to tell whether a given sentence is, in fact, expendable in favor of another. The following suggestion seems plausible. We may say that a given sentence is expendable in favor of another if, and only if, the two sentences are equivalent in meaning. This suggestion is supported by the fact that as we have seen above, Chisholm argues that if we show, of a given non-psychological sentence, that we can transform it into another sentence, then we have shown that we do not need that given sentence in order to describe non-psychological phenomena.

Further support for this proposal may be seen in Chisholm's remark that "anyone who understands the language can readily transform" the sentences in the manner indicated. [Chisholm (17) 512] This suggests that no more than a knowledge of the meanings of the component terms is required. In another context, moreover, Chisholm appears to use as interchangeable the two locutions 'B is a transformation of A.' and 'Instead of saying A, one may say B.'. (Cf. [Chisholm (17) 511]) This suggests that one sentence is a transformation of another if, and only if, one says the same thing by using one sentence as one does by using the other. This, in turn, supports the view that a sentence is expendable in favor of another if, and only if, they are equivalent in meaning. In the absence of another plausible account of expendability, then, we may accept one in terms of meaning-equivalence. What particular account is adopted, however, is not relevant to the issues discussed below so long as we have some means of understanding the claims Chisholm is making.

These considerations suggest the following modification of our earlier formulation of the thesis of intentionality. Instead of claiming that a sentence is non-psychological if, and only if, it is non-intentional, we may say that a sentence is non-psychological if, and only if, it is either non- intentional or expendable in favor of a non-intentional one. We shall have to examine this proposal more closely in section III, but we must first turn to a consideration of sentences which are psychological but do not exhibit the characteristic, in order to see what modification is suggested by Chisholm's discussion of such sentences.

Chisholm's strategy seems to be the following. If we consider a sentence like

John is cat-perceptive.

we shall find that, although it does not exhibit the characteristic, it contains a technical term, namely, 'cat-perceptive'. Of such technical terms, Chisholm asserts that "if we wish to tell anyone what . . . [they] mean, we must use intentional language again." [Chisholm (17) 513] Chisholm appears, moreover, to be arguing that if we can show that a sentence which is psychological contains such a term, then the existence of that sentence is compatible with the claims he wishes to make.{8} Instead of claiming, then, that a sentence is psychological if, and only if, it is intentional, we may say that a sentence is psychological if, and only if, either it is intentional or it contains a technical term of the kind indicated.{9}

It is important for Chisholm to be able to provide some means of identifying these terms, independent of the fact that they occur in psychological, non-intentional sentences. A suggestion for such a means of identification may be found in Chisholm's remark that we must use intentional language in order to tell someone what the terms mean. For in imposing this requirement, Chisholm appears to be claiming that if we are to express what any sentence containing a given technical term means without using the term, we shall have to use at least one intentional sentence. Thus, the following test for such terms may be tentatively proposed. A term is of the sort indicated if, and only if, any sentence which contains it entails some intentional sentence. {10}

As it stands, however, this means of identifying such terms is inadequate. Let us consider the case of a sentence constructed by disjoining two sentences, one of which is intentional and the other of which is analytic. Since a compound sentence is intentional if, and only if, it has at least one simple sentence component which in its declarative form is intentional, the sentence we are considering will itself be intentional. But it is clear that this same sentence will also be analytic. Now since an analytic sentence is entailed by any sentence whatever, it is clear that any sentence whatever will entail the particular intentional, analytic sentence which we have constructed. But it follows from this that any sentence whatever trivially satisfied our necessary and sufficient condition for a sentence to contain a technical term. It appears that difficulty can be met by slightly revising our necessary and sufficient condition, as follows. A term is of the sort indicated if, and only if, any sentence which contains it entails some simple intentional sentence. {11}

From examining Chisholm's discussion of his putative counter-examples, we have come up with suggestions for modifying our earlier formulation of the thesis. As we remarked above, we have in effect retained our original formulation of the definition of intentionality. Our aim is to formulate both the thesis and the definition in such a way that they will be compatible with sentence (4). With this aim in mind, we might have chosen to modify our earlier formulation of the definition, retaining that of the thesis. The passages of Chisholm's text which have been considered, however, appear strongly to favor a modification of our original formulation of the thesis, and a retention of that of the definition.{12}

III. Having done some preliminary groundwork, we are now in a position more closely to examine the required modifications of our original formulation of the thesis of intentionality. The present section will examine what modifications of the thesis might be required in addition to those already made. Section IV will then propose an approach which is an alternative to that of the present section, and section V will introduce still further modifications of the proposals advanced in sections III and IV.

In order most easily to formulate our modifications, it will be convenient to adopt the following conventions. We shall use the letter 'P' to stand for the predicate 'describes some psychological phenomenon', 'S' for ' is a sentence', 'I' for 'is intentional', 'C' for 'exhibits the characteristic', 'E' for 'is expendable in favor of', 'K' for 'contains' and 'T' for 'is a technical term of the sort mentioned by Chisholm'. Using these conventions, we may rewrite our formulation of the definition of intentionality as

(5) (x) [Sx --> (Ix <--> Cx)]

and our original formulation of the thesis as

(x) [Sx --> (Px <--> Ix)]

The two sentences which represent our modification of this first formulation of the thesis may then be written as follows:

(6) (x) [Sx --> (Px <--> (Ix v (Ey)(Ty & Kxy)))]

(7) (x) [Sx --> (~Px <--> (~Ix v (Ey)(Sy & ~Iy & Exy)))]

We may, finally, rewrite sentence (4) as a conjunction of the following two sentences:

(8) (Ex) (Sx & ~Px & Cx)

(9) (Ex) (Sx & Px & ~Cx)

(As we shall see, it will be convenient in what follows to treat these two conjuncts separately.)

We are now in a position to say that sentence (5) represents the definition, and that sentences (6) and (7) jointly represent the thesis. Our problem may then be put in terms of the following question: Is the conjunction of (5), (6), (7), (8) and (9) consistent? If it is not, then we must seek another formulation which does meet at least this condition.

It is convenient to take the two conjuncts of the thesis separately, and to ask whether each is compatible with the conjunction of (5), (8) and (9). If we do so, we find the following. From the conjunction of sentences (5), (6) and (8), we can derive the inconsistent sentence

(Ex) (Px & ~Px) .

It follows that this conjunction of sentences is itself inconsistent. Similarly, from the conjunction of sentences (5), (7) and (9), we can derive the same sentence

(Ex) (Px & ~Px),

from which it follows that this second conjunction is also inconsistent. Since neither conjunct of the thesis is compatible with the conjunction of (5), (8) and (9), it follows that we must look for a different formulation.

It seems clear that our best policy is to seek an adequate formulation of the thesis, having accepted our original version of the definition and the existential claims made by sentences (8) and (9). It might be suggested, then, that what we really wish to assert by means of our thesis is that a sentence is psychological if, and only if, either it is intentional or it is not intentional but contains a technical term; and, in parallel fashion, that a sentence is not psychological if, and only if, either it is not intentional or, although intentional, is expendable in favor of a non-intentional one. Our modified thesis would then read as follows:

(6') (x) [Sx --> [Px <--> (Ix v (~Ix & (Ey) (Ty & Kxy)))]]

(7') (x) [Sx --> [~Px <--> (~Ix v (Ix & (Ey) (Sy & ~Iy & Exy)))]]

It can be shown once again, however, that the sentence '(Ex) (Px & ~Px)' is deducible from a conjunction either of (5), (6') and (8), or of (5), (7') and (9).{13}

A third possibility seems to present itself, however, which may hold more promise. We may require, as necessary and sufficient conditions for a sentence to be psychological, that either it is intentional and is not expendable in favor of a non-intentional sentence, or it contains a technical term. In parallel fashion, we may require as necessary and sufficient conditions for a sentence to be not psychological, that either it is not intentional and does not contain a technical term, or it is expendable in favor of a non-intentional sentence. Using our abbreviated notation, we may express these requirements as follows:

(6'') (x) [Sx --> [Px <--> ((Ix & ~(Ey)(Sy & ~Iy & Exy)) v (Ey) (Ty & Kxy))]]

(7'') (x) [Sx --> [~Px <--> ((~Ix & ~(Ey)(Ty & Kxy)) v (Ey) (Sy & ~Iy & Exy))]]

If, now, we ask of these two sentences whether each is compatible with the conjunction of sentences (5), (8) and (9), it appears that we may answer affirmatively. For we seem unable, from the conjunction of these five sentences, to derive any inconsistent sentence, such as '(Ex) (Px & ~Px)'.

Sentences (6") and (7") do not, however, provide us with a suitable formulation of the thesis of intentionality. Let us again consider the sentence

John is cat-perceptive.

This sentence contains a technical term, namely 'cat-perceptive'. But it also is expendable in favor of a non-intentional sentence, for it is equivalent in meaning to a non-intentional sentence, namely itself. (And in any event, other technical terms could be invented, with appropriately specified meanings, in order to construct different sentences containing these terms which would be equivalent in meaning to the sentence in question.) That sentences exist which both are expendable in favor of non-intentional sentences and contain technical terms may be asserted by

(4' ) (Ex) (Ey) (Ez) (Sx & Sy & ~Iy & Exy & Tz & Kxz).

It is easily shown, however, that from the conjunction of (6"), (7") and (4') we may generate inconsistency, for we may derive from this conjunction the sentence '(Ex) (Px & ~Px)'. Since, as we have seen, we cannot reject the truth of sentence (4'), we must reject the conjunction of (6") and (7").

A fourth modification of the thesis, however, may be constructed by combining features of sentences (6') and (7') with those of sentences (6") and (7"), namely

(6''' ) (x) [Sx --> [Px <--> ( (Ix & ~(Ey) (Sy & ~Iy & Exy)) v (~Ix & (Ey)(Ty & Kxy)))]]

conjoined with

(7''' ) (x) [Sx --> [~Px <--> ((~Ix & ~(Ey) (Ty & Kxy )) v (Ix & (Ey)(Sy & ~Iy & Exy)))]].

Like the conjunction of sentences (6") and (7"), this version of the thesis is compatible with the conjunction of sentences (5), (8) and (9). Unlike the earlier version, however, the conjunction of sentences (6''') and (7"') is also compatible with sentence (4'), and thus avoids the difficulty which arose from the truth of sentence (4'). It is not required, however, that we use the conjunction of (6''') and (7'") in formulating the thesis. For it is easily shown that these two sentences, unlike the conjuncts of our earlier formulations, are interdeducible, and from this it follows that each, taken alone, is interdeducible with the conjunction of both. For this reason, we may choose one of the sentences, say (6'''), as our modified formulation of the thesis.

We shall return to further consideration of this modified version of the thesis in sections V and VI. Before further discussion of it, however, we may ask whether there is a different, and simpler, modification of our original formulation which also avoids incompatibility with the conjunction of sentences (5), (8), (9) and (4'). In the search for such an alternative formulation of the thesis, it will be useful to return once more to an examination of the passage in which Chisholm asserts the thesis. The following section is devoted to these tasks.

IV. We may begin anew by expressing our original formulation of the thesis as a conjunction of two universally quantified material conditionals, and rather than ask how we may satisfactorily modify a single universally quantified biconditional, we may ask instead how best to modify each of our two conditionals. Thus, we may start with

(x) [Sx --> (Px --> Ix)]

and

(x) [Sx --> (~Px --> ~Ix)]

It is clear that neither conditional is true as it stands, for

(Ex) (Sx & Px & ~Ix)

and

(Ex) (Sx & ~Px & Ix)

entail, respectively, the denials of our conditionals. Our task, then, is to modify each conditional so as to avoid incompatibility with our existentially quantified sentences. Following the tactics employed in section II, the following two conditionals are suggested:

(10) (x) [Sx --> [Px --> (Ix v (Ey)(Ty & Kxy))]]

(11) (x) [Sx --> [~Px --> (~Ix v (Ey) (Sy & ~Iy & Exy))]].

If we now consider the conjunction of sentences (5), (8), (9), (10), (11) and (4'), we find we cannot infer from it that (Ex)(Px & ~Px). It appears that the conjunction of these five sentences is a consistent sentence.

Sentences (10) and (11) are simpler in structure than either (6''') or (7'''), and, since each set is compatible with (5), (8), (9) and (4'), the former seems preferable as a formulation of the thesis. It is worth asking, then, whether the text supports this weakened reading. Chisholm asserts that "when we describe non-psychological, or 'physical,' phenomena," the sentences we must use satisfy a certain condition; and that "when we wish to describe certain psychological phenomena," the sentences we must use satisfy another, and different, condition. [Chisholm (17) 511] He does not, however, make the stronger assertion that a sentence satisfies the first condition if, and only if, it is non-psychological; nor that a sentence satisfies the second condition if, and only if, it is psychological. In short, he does not explicitly mention any sufficient conditions for a sentence to be psychological, or for it to be non-psychological.{14} It seems, therefore, that we may be unwarranted in representing the thesis by either sentences (6"') or (7'''). It remains, then, for us to ask whether there may be respects in which our last version is, after all, unacceptable. In particular, we may ask whether this version is compatible with the existence of certain sorts of cases which Chisholm would wish to exclude.

The conjunction of sentences (10) and (11) is interdeducible with

(12) (x) [Sx --> [(Px v ~Ix v (Ey) (Sy & ~Iy & Exy)) & (~Px v Ix v (Ey)(Ty & Kxy))]],

so that we may look for items which sentence (12) fails to exclude, even though Chisholm would wish them ruled out. Two sorts of items seem to satisfy this condition: a sentence which at once is non-psychological, is non-intentional and contains a technical term, and a sentence which is at once psychological, intentional and expendable in favor of a non-intentional sentence. It is clear that each of these is compatible with the condition expressed by (12), and that they are therefore compatible with the conjunction of (10) and (11). Would Chisholm wish to exclude these sentences? Since a sentence of the first sort contains a technical term, we know that if we wish to avoid using this term (or some other technical term), we must use intentional language instead. And since a sentence of the second sort is expendable in favor of a non-intentional sentence, we know that its psychological subject matter could be expressed, equally well, by a non-intentional sentence. It would therefore follow from the existence of such sentences that there are non-psychological sentences which, to be re-expressed, require the use either of intentional language or of technical terms; and that there are psychological sentences which can be re-expressed by means of non-intentional language. Without further qualification, therefore, such cases seem to run counter to the purport of Chisholm's thesis.

In section V, we shall consider similar types of cases which it seems likely that Chisholm would wish to exclude, but at this point we must ask whether there is some convenient modification of sentences (10) and (11) which would effectively rule out the cases discussed above. Once again, we might be tempted to propose the following alteration:

(10') (x) [Sx --> [Px --> (Ix v (~Ix & (Ey)(Ty & Kxy)))]]

(11') (x) [Sx --> [~Px --> (~Ix v (Ix & (Ey)(Sy & ~Iy & Exy)))]].

It can easily be seen, however, that the existence of the sorts of sentences in question is compatible with what is expressed by (10') and (11'). Moreover, the conjunction of these two sentences is interdeducible with the conjunction of (10) and (11), just as the conjunction of (6) and (7) was with that of (6') and (7'). It is still open, however, to adopt a modification parallel in structure to (6") and (7"), namely,

(10") (x) [Sx --> [Px --> ((Ix & ~(Ey)(Sy & ~Iy & Exy)) v (Ey) (Ty & Kxy))]]

(11'') (x) [Sx --> [~Px --> ((~Ix & ~(Ey) (Ty & Kxy)) v (Ey)(Sy & ~Iy & Exy))]]

This proposal appears to satisfy our requirements, for the conjunction of these two sentences entails the denial of

(Ex) [Sx & ~Px & ~Ix & (Ey) (Ty & Kxy)]

and of

(Ex) [Sx & Px & Ix & (Ey) (Sy & ~Iy & Exy)].

We have proposed two sentences, (6''') and the conjunction of (10'') and (11''), each of which has a good claim to represent accurately and effectively the content of Chisholm's thesis of intentionality. Both sentences are jointly compatible with the conjunction of the definition of intentionality and sentences asserting the existence of certain sentences which Chisholm admits do, in fact, exist. Each formulation, moreover, rules out the existence of certain sentences which it seems plausible to suppose that Chisholm wishes to have ruled out. And because the second formulation is both simpler and weaker than the first, it seems that the second has some claim to be preferred. For we have seen that Chisholm nowhere makes explicit mention of sufficient conditions for a sentence to be psychological, or for a sentence to be non-psychological.

One feature of both proposed formulations seems especially deserving of mention. In each case, the formulation exhibits a structure of considerably greater complexity than might have been expected from an examination of Chisholm's article. In particular, the occurrence, in each proposal, of denials of existentially quantified expressions might not have been expected from a perusal of Chisholm's own formulation. For it is tempting to suppose, given Chisholm's words, that if we have shown of a sentence either that it is both psychological and intentional, or that it is both non-psychological and non-intentional, we have thereby shown the sentence to be consonant with the thesis. To suppose this would be mistaken, however, if either (6'"), or the conjunction of (10") and (11"), is an adequate formulation of the thesis. For to show of some particular sentence that it is consonant with either version of the thesis, it is not enough to show that it is both psychological and intentional; one must show also that there exists no non-intentional sentence in favor of which the sentence is expendable. Similarly, to show of a non-psychological and non-intentional sentence that it is consonant with the thesis, one must show that it contains no technical term.{15}

V. In the previous section we had reason to discuss sentences which are non-psychological, are non-intentional, and contain instances of our special vocabulary; and sentences which are psychological, intentional and expendable in favor of non-intentional sentences. It was in an attempt to rule out the possibility of such cases that we arrived at the conjunction of sentences (10'') and (11'') as a proposed formulation of the thesis. We may now ask whether there are certain other sorts of sentences which, although allowed by our formulations as they stand, we should prefer to rule out.

Let us consider a sentence which is both non-psychological and intentional. We say infer, by means either of (6"') or of (11''), that this sentence is expendable in favor of a non-intentional one. But let us suppose that this non-intentional sentence itself contains a technical term. It would follow that we have succeeded in showing only that our non-psychological, intentional sentence is expendable in favor of a sentence which, in order to be re-expressed, requires the use of intentional language all over again. In order to rule out this possibility, we may propose the following alteration of sentence (6'''):

(13) (x) [Sx --> [~Px <--> ((~Ix & ~(Ey)(Ty & Kxy)) v (Ix & (Ey)(Sy & ~Iy & Exy & ~(Ez)(Tz & Kyz))))]]

Similarly, if we consider the formulation proposed in section IV, it is clear that an identical modification, having the same effect, can be made by altering sentence (11") so as to read

(14) (x) [Sx --> [~Px --> ((~Ix & ~(Ey)(Ty & Kxy)) v (Ey)(Sy & ~Iy & Exy & ~(Ex)(Tz & Kyz)))]]. {16}

There seems to be no other means of dealing with the problem raised by the existence of such sentences than that of adopting such a modification. We may therefore conclude that whichever formulation we accept, the restriction described here must be introduced.

This required restriction may be introduced in one case by simply rejecting the conjunction of (10") and (11") as a possible formulation of the thesis, and using the conjunction of (10'') and (14). In the case of our other proposed formulation, however, we have been using sentence (6''') taken alone, since, as we have seen, sentences (6'") and (7''') are interdeducible. Our restriction may be introduced here, then, either by rejecting (6''') in favor of the conjunction of (6''') and (13), or by rejecting it in favor of (13) taken alone. It remains to ask which of these moves seems more satisfactory.

A sentence which is at once psychological, intentional and expendable in favor of a non-intentional sentence would be a counter-example to (13) only if the sentence in favor of which it is expendable contains no technical term. If the sentence in favor of which it is expendable does contain a technical term, it would not provide a counter-example to (13), although it would nonetheless provide one to (6'''). Since there seems to be no reason to rule out such a case, sentence (13) taken alone seem preferable to the conjunction of (6''') and (13) as a formulation of the thesis. For the latter, unlike the former, will be incompatible with the existence of such cases.

A similar consideration may lead us slightly to modify sentence (10''). For just as we saw with (6'''), a sentence which is psychological, intentional, and expendable in favor of a non-intentional sentence containing a technical term will provide us with a counter-example to (10"). Since such a sentence would not, however, be a counter-example to

(10''') (x) [Sx --> [Px --> ((Ix & ~(Ey)(Sy & ~Iy & Exy & ~(Ez)(Tz & Kyz)) v (Ey) (Ty & Kxy))]],

the latter seems preferable as a conjunct of a formulation of the thesis.

The argument for modifying sentences (11'') and (6"') rested on a consideration of certain sorts of non-psychological, intentional sentences. Is there any parallel modification which should be introduced because of similar problems arising from the existence of psychological, non-intentional sentences? The above alteration consisted in modifying our conditions for the expendability of one sentence in favor of another. Let us ask, then, whether, given the existence of psychological, non-intentional sentences, any modification is required of our test for technical terms. For by using either sentence (10''') or sentence (13) we can infer of a psychological, non- intentional sentence that it contains a technical term.

The means proposed in section II for independently identifying technical terms was the following: a term is of the sort indicated if, and only if, any sentence which contains it entails some simple intentional sentence. Given this understanding, we could everywhere replace the expression

(Ey) (Ty & Kxy)

by the following (open) sentence:

(Ey) [y is a term & Kxy & (z)[(Sz & Kzy) --> (Ew) (Sw & Iw & w is simple & z entails w)]].

Suppose, however, that certain sentences which contain such a term entail only sentences which, although intentional, are also expendable in favor of non-intentional ones. It would follow that although our condition is satisfied, the so-called technical term does not require the use of intentional language for its explication. For the fact that certain entailed sentences were expendable in favor of non-intentional ones would allow us to use these non-intentional sentences for our explication.

How might we so modify our necessary and sufficient condition for a term to be a technical term as to avoid this difficulty? The following addition to our earlier statement seems adequate: A term is of the sort indicated if, and only if, any sentence which contains it entails some simple intentional sentence, and this simple intentional sentence is not expendable in favor of a non-intentional one. Given this modification we can everywhere replace the expression

(Ey) (Ty & Kxy)

by

(15) (Ey)[y is a term & Kxy & (z) [(Sz & Kzy) --> (Ew)(Sw & Iw & w is simple & z entails w & ~(Ev)(Sv & ~Iv & Ewv))]].

There are several additional objections to this last proposed test for technical terms which must be discussed. In section II we found it necessary to introduce the restriction that the entailed sentence mentioned in our test be a simple sentence. For if the entailed sentence were permitted to be compound, then, since any sentence whatever entails the disjunction of an analytic sentence and an intentional one, our test would be trivially satisfied by any sentence at all. The restriction introduced in section II is not, however, adequate to prevent our test being trivially satisfied in this fashion.

Let us coin a two-place predicate expression 'glub' which, we shall stipulate, means the same as 'is tall or is not tall or is thinking of'. It follows that the sentence

George glub Vienna.

is, although simple, both intentional and analytic, for it will mean the same as the sentence

George is tall or is not tall or is thinking of Vienna.

Thus if we wish to avoid our test being trivially satisfied by any sentence whatever, the entailed sentence mentioned in the test must be not only simple, but non-analytic as well. This restriction can be introduced by requiring that there is some sentence which does not entail the entailed sentence in question. For any analytic sentence will fail to satisfy this condition. Thus sentence (15) might be expanded so as to read:

(15') (Ey) [y is a term & Kxy & (z) [(Sz & Kzy) --> (Ew)(Ev) (Sw & Iw & w is simple & z entails w & Sv & ~(v entails w) & ~(Eu)(Su & ~Iu & Ewu))]].

But if any sentence containing a technical term is to entail some simple, non-analytic, intentional sentence, we encounter the following difficulty. Since no analytic sentence entails any non-analytic one, a result of this restriction is that no technical term is contained by an analytic sentence. Since this is clearly not the case, we must further revise our sentence as follows:

(15") (Ey) [y is a term & Kxy & (z) [(Sz & Kzy & (Ew)(Sw & ~(w entails z)) --> (Ev)(Eu)(Sv & Iv & v is simple & z entails v & Su & ~(u entails v) & ~(Et)(St & ~It & Evt))]].

This revised sentence is not, however, proof against a similar objection. For consider the term 'tall', and a sentence containing it, say

George is tall.

This sentence entails the (compound) intentional sentence

George is tall or George is thinking of Vienna.

But if we coin a two-place predicate expression 'blub' which we stipulate to mean the same as 'is tall or is thinking of', then our sentence will also entail the simple intentional sentence

George blub Vienna.

Since similar devices can be employed for any other sentence which contains the term 'tall', it seems to follow that 'tall' is a technical term, since any sentence containing it entails some simple intentional sentence. It is not clear just how this objection might be met, for it is not clear that we can avoid the consequences of coining predicates for the purpose of constructing such entailed sentences. {17}

There is a third objection to our test for technical terms, as contained now in sentence (15"). As we have seen, the sentence

(16) John is cat-perceptive.

contains a technical term, and it also non-trivially entails a simple intentional sentence, namely

John perceives a cat.

As our test now stands, a term is technical only if the entailed intentional sentence is simple, non-analytic, and not expendable in favor of some non-intentional sentence. In the case under consideration, however, the entailed intentional sentence is expendable in favor of a non-intentional sentence, namely (16). For we may understand the entailed sentence to mean the same as sentence (16). We must therefore revise this restriction so as to provide for such cases. The case which we have considered suggests our doing this by requiring that if the entailed intentional sentence is expendable in favor of some non-intentional sentence, then that non-intentional sentence contains some technical term. Thus we might revise sentence (15") as follows:

(15''') (Ey) [y is a term & Kxy & (z) [(Sz & Kzy & (Ew)(Sw & ~(w entails z)) --> (Ev)(Eu) (Sv & Iv & v is simple & z entails v & Su & ~(u entails v) & ~(Et)(St & ~It & Evt & ~(Es) (Ts & Kts)))[[.

This open sentence is intended to be an adequate replacement of the expression

(Ey) (Ty & Kxy).

As a result of our last modification, however, sentence (15''') ) itself contains the expression '(Es)(Ts & Kts)'. For this reason, (15''') is inadequate to provide a test for a sentence containing a technical term. It is not at all clear that there is any way in which this circularity can be avoided. For if we wish to show that a given term is technical, then we must show that if we are to avoid using the term in question, we must use either intentional language or some other technical term. (Cf. [Chisholm (17) 513]) We shall return to this point in section VII, but we shall first turn to a related discussion, for the purpose of which we shall provisionally accept sentence (15"') as our test for technical terms.

VI. Each formulation of the thesis which we have examined exhibits, as does the definition of intentionality, a universally quantified form. For this reason, the thesis must be understood as asserting something about an indefinitely large number of sentences. It is therefore logically impossible conclusively to verify it.{18} For conclusive verification would require that we show, of each of an indefinitely large number of sentences, that it satisfies the thesis. Conclusive falsification, on the other hand, would require only that we show, of some given sentence, that it fails to satisfy the thesis. This leads us to ask whether, given one or another of the formulations discussed, it is logically possible conclusively to falsify the thesis.

Let us for the moment ignore the difficulties we encountered in identifying a term as a technical one, and suppose that we are given an effective procedure for picking out technical terms. (Because the number of terms which any sentence contains is finite, such a procedure will allow us effectively to determine, of any sentence, whether it does, or does not, contain a technical term.) If, now, we examine each of the sentences (6'''), (13), (10'''), (11") and (14), we find that it is logically possible to falsify each but the last two. For in order to show that some sentence failed to satisfy either (11") or (14), we would have to show, respectively, either that

(Ey)(Sy & ~Iy & Exy)

or that

(Ey)[Sy & ~Iy & Exy & ~(Ez)(Tz & Kyz)]

is not true of the sentence. This cannot be done, for it would involve showing something to be true of an indefinitely large number of sentences. Sentences (11'') and (14), however, are used to formulate the thesis only in conjunction with (10'''). Since it is logically possible to falsify (10'''), it is also logically possible to falsify any formulation of the thesis which contains (10''') as a conjunct. For this reason, our inability to falsify either (11'') or (14) does not prevent us from being able to falsify formulations in which one or the other occurs.{19} Given an effective procedure for picking out technical terms, then, it is logically possible to find counter-examples to any of our formulations of the thesis, and thereby conclusively to falsify them.

If we take account of the difficulties we encountered in devising an adequate test for technical terms, however, the situation is importantly altered. For the above conclusions were based, in part, upon our assumption that we were able to determine whether or not '(Ey) (Ty & Kxy)' was true of a given sentence. If we adopt the test for technical terms suggested in the previous section, however, then if we wish to determine whether or not '(Ey)(Ty & Kxy)' is true of a certain sentence, we must first determine whether or not the (open) sentence (15''') is true of that sentence. In order to show that (15''') is true, or that it is not true, of a given sentence, we must show that the universally quantified (open) sentence embedded in (15''') is true, or that it is not true, of a given term. To show this embedded sentence to be true of some term, however, would involve showing something about indefinitely many sentences, which we cannot do. Similarly, because we cannot conclusively show the consequent of this embedded conditional to be false, we are also unable to show, of any term whatever, that (15''') is not true of it.{20} Let us ask, then, what effect this limitation has on the logical possibility of finding counter-examples to our several versions of the thesis.

In order conclusively to falsify sentence (6'''), it is sufficient simply to find a sentence which is both psychological and expendable in favor of a non-intentional sentence, and since it is possible to do this, it is also possible to falsify (6'''). If we wish to find a counter-example to any of our other five sentences, however, we must be able to show either that some sentence contains, or fails to contain, a technical term, or that there is no non-intentional sentence in favor of which a given sentence is expendable. Failing an alternative independent test for technical terms, however, we can show none of these things. It follows that it is logically impossible, failing such a test, to falsify any of our five sentences except (6''').

A requirement of conclusive falsifiability, then, rules out each formulation of the thesis except that expressed by (6'''). As we have noticed, however, this formulation has several disadvantages over the others. For, unlike either (13) or the conjunction of (10''') and (14), it is unable effectively to deal with non-psychological, intentional sentences. Furthermore, it seems unlikely that such a formulation, with its embedded biconditionals, captures Chisholm's intent as accurately as one containing conditionals instead. In the absence of an alternative independent test for technical terms, however, it is the only one of our proposals which logically allows for the possibility of counter-examples.

This limitation is not the only one which results from adopting sentence (15''') as an explication of the expression 'x contains a technical term'. For just as we are prevented from conclusively showing that a certain sentence is a counter-example to any of our formulations but one, so we encounter similar difficulties in trying to show instead that a certain sentence satisfies one of the formulations. For we are unable to show that there is no non-intentional sentence in favor of which a given sentence is expendable, and if we adopt sentence (15''') we are unable to show that a given sentence contains, or fails to contain, a technical term. If we can show none of these things, however, then we are (logically) unable to show of any sentence whatever that it satisfies (13); or, of any psychological sentence, that it satisfies either (6''') or (10''') or, of any non- psychological sentence, that it satisfies (14). If a sentence is to satisfy a formulation consisting in a conjunction, it clearly must satisfy both conjuncts. It is therefore impossible, failing a satisfactory alternative to sentence (15'''),{21} conclusively to show of any sentence whatever that it satisfies any formulation except the conjunction of (10''') and (11"), or (6"') alone. And we are unable to show that any but nonpsychological sentences satisfy these formulations.{22}

VII. Chisholm's article represents one attempt to discover some non-analytic connection between linguistic expressions which describe psychological phenomena and linguistic expressions which exhibit specifiable logical features. The difficulties which face this attempt result initially from the particular set of logical characteristics which Chisholm uses in framing his definition of intentionality.{23} For if this definition is adopted, then one cannot truly assert that all and only psychological sentences exhibit at least one of the characteristics. Chisholm formulated his claims with a view to taking account of this fact. In the course of introducing necessary qualifications on the connection which we originally supposed Chisholm wished to establish, however, we arrived at a set of connections each of which exhibits considerable complexity. Because of this complexity, even if we were able to establish such a connection, it would not exhibit any intuitively clear relationship between the logical features in question and psychological subject matter. The value of establishing such a connection would appear to lie in the possibility of increasing our understanding of psychological phenomena by logically characterizing the language used to describe such phenomena. But what such a complex connection might tell us about psychological phenomena is not at all clear.

The non-intuitive character of the connection which must be proposed if one adopts this definition of intentionality is not, as we have seen, the only difficulty which an attempt using this definition faces. For if we use the test for technical terms suggested by Chisholm's discussion, then we run into obstacles in trying even to establish the connection. Such an attempt would have to consist in showing that the connection holds in a large number of sample cases. But we are able conclusively to show this only of sentences which fail to describe psychological phenomena. And the relationships which we can establish, in the case of non-psychological sentences, are not, as we have seen, satisfactory. For they fail to exclude the possibility that a non-psychological, intentional sentence may be expendable only in favor of a sentence which contains a technical term, and which therefore entails, in turn, an intentional sentence. If we are able to establish a relationship only between non-psychological sentences and sentences exhibiting certain logical features, it is even less clear what such a connection might tell us about psychological phenomena.

Most of the arguments which Chisholm presents in "Sentences about Believing," however, lend support to his thesis not directly, but rather by dealing with putative counter-examples. Thus Chisholm writes: "I do not pretend to be able to show that . . . [the thesis] is true in its application to believing. But I think that there are serious difficulties . . . which stand in the way of showing that it is false." [Chisholm (17) 512] As we have seen, however, if we have no alternative to the test for technical terms suggested by Chisholm's remarks, then we are logically prevented from showing that the thesis is false in any but the least acceptable of our modified versions. For the modifications which we were forced to introduce resulted in a connection for which it is impossible to find conclusive counter-examples. And since we are logically prevented from finding counter- examples to any but the least satisfactory of our formulations, arguing against putative counter-examples cannot lend the plausibility to the thesis which it otherwise might.{24} It appears, then, that if any illuminating connection of this general sort is to be discovered, it must be one involving a different set of logical characteristics.{25}

Finally there are the difficulties which we have encountered in the course of attempting to develop an adequate test for technical terms. As we have seen, Chisholm introduced the notion of technical terms in order to account for certain descriptions of psychological phenomena which exhibited none of the logical characteristics with which he was concerned. The last problem which faced us in this connection was that in order to determine whether or not a given term is technical, we must make reference not only to Chisholm's logical characteristics, but also to a class of terms which we already know to be technical. It appears, then, that we cannot determine which terms are technical apart from determining the whole family of such terms. And it is not clear that we have any way of doing this apart from making reference to whether or not a given term appears to us to be ordinarily used in describing psychological phenomena.

Given anything like our understanding of Chisholm's notion of technical terms, we have seen that any sentence which exhibits any of Chisholm's logical characteristics will also contain some technical term. If we had an adequate test for technical terms, therefore, our logical characteristics would be superfluous. For our thesis could then provide simply that all and only psychological sentences contain technical terms.{26} Our inability to find a satisfactory independent test for technical terms, therefore, represents the most important difficulty which any program similar to Chisholm's must overcome.


[Table of Contents] [Go to Chapter 2]

Notes:

{1} Cf. [Chisholm (3)], [Chisholm (4)], [Chisholm (7)], [Chisholm (9)], [Chisholm (10)], [Chisholm (11)], [Chisholm (12)], and [Chisholm (16)]. Several of these alternative attempts will be discussed in the course of Chapter Two. [Back]

{2} Chisholm formulates these three logical characteristics as follows. A sentence exhibits the first "if it uses a substantival expression--a name or a description--in such a way that neither the sentence nor its contradictory implies either that there is or that there isn't anything to which the substantival expression truly applies." [Chisholm (17) 510] A sentence exhibits the second "if neither the sentence nor its contradictory implies either that [the indicative form of] the phrase following the principle verb is true or that it is false." [Chisholm (17) 510] Finally, a sentence exhibits the third "if it contains a name (or description) which has an indirect reference in that sentence," where we understand "a name (or description) of a certain thing . . . [to have] an indirect reference in a sentence if [and only if] its replacement by a different name (or description) of that thing results in a sentence whose truth-value may differ from that of the original sentence." [Chisholm (17) 511] Chisholm does not explicitly state that one or another of these three logical marks is necessary for a simple declarative sentence to be intentional. He does, however, seem to presuppose this throughout the balance of his paper, and, following Cornman, we shall take for granted that Chisholm has this in mind. (Cf. [Cornman (1) 49].) [Back]

{3} Although Chisholm does not discuss sentences which are not declarative, it seems open to suggest the following device for handling such sentences: a non-declarative sentence is intentional if, and only if, the sentence would, in its declarative form, exhibit the logical characteristic. This way of dealing with interrogative, imperative and exclamatory sentences is suggested by Chisholm's means of determining the intentionality of compound declarative sentences. [Back]

{4} If we were to adopt some means of classifying non-declarative sentences as intentional or not, we might wish to make some corresponding alteration in the wording of this thesis. For example, we might speak not about sentences which describe psychological or non-psychological phenomena, but rather about sentences the subject matter of which is psychological or non-psychological. These moves will not be discussed, however, in what follows. [Back]

{5} Chisholm speaks of sentences which describe non-psychological phenomena, rather than of sentences which fail to describe any psychological phenomenon. If, however, we confine our attention to declarative sentences, that is, to sentences which are used to describe, then these two classes become co-extensional. This restriction will be presupposed throughout, in what follows. [Back]

{6} That is, neither the former sentence nor its denial entails either that there will, or that there will not, be any new epidemics. This sentence exhibits, therefore, the first of our several logical marks. (Cf. footnote 2.) None of these marks, however, is exhibited by the latter sentence. [Back]

{7} Compare an earlier passage in which Chisholm talks of "the 'intentional use' of language" as one use "which we need to . . . [make of] language when we talk about certain psychological states and events. . . . It is a kind of use we can avoid when we talk about non-psychological states and events." [Chisholm (17) 510] Since Chisholm nowhere adds further stipulations concerning what he (as against Brentano) means by 'intentional,' it is open for him to seek a set of necessary and sufficient conditions for a sentence to be intentional which will nevertheless fail to be necessary and sufficient for a sentence to be psychological. [Back]

{8} Thus Chisholm is specifying a condition which, if satisfied by a psychological, non-intentional sentence, renders the existence of that sentence compatible with his thesis. It is important to distinguish this move from one of a different kind which Chisholm makes when, later in his article, he discusses the American "New Realists," Ayer's Thinking and Meaning, and James' pragmatic theory of truth. Chisholm (17) 514-519] In these pages Chisholm considers instances of non-intentional sentences which have been advanced as expressing the same thing as certain psychological sentences, and argues that, as a matter of fact, they do not. Similarly, Chisholm hopes to get around the fact that "the sentences we use to describe the meanings and uses of words are ordinarily intentional" [Chisholm (17) 516] not by claiming that these sentences are expendable in favor of non-intentional ones, but rather by arguing that they are, in fact, psychological. [Chisholm (17) 517] (Thus he is here using the second of the two strategies indicated in the passage quoted near the beginning of section II.) These later discussions are not, therefore, relevant to a consideration of Chisholm's treatment of sentences which are, in fact, either psychological and non-intentional, or non-psychological and intentional. [Back]

{9} Again, this modification is supported by a closer reading of Chisholm's own formulation of the thesis. For part (2) of this formulation mentions two conditions which are disjointly necessary for a sentence to be psychological. One is that it be intentional, and the other is that it contain some term belonging to "a vocabulary which we do not need to use when we describe non-psychological, or 'physical,' phenomena." [Chisholm (17) 512]

Chisholm's description of the logical characteristic invokes a distinction between simple and compound sentences. It is therefore worth asking whether the adequacy of the characteristic would be affected by coining new predicates so as to be able to construct simple sentences equivalent in meaning to certain compound ones. Let us suppose that 'plub' is a coined two-place predicate meaning the same as 'is tall and is not tall and is thinking of'. Since a self-contradictory sentence entails any sentence whatever, the sentence

George plub Vienna.

will entail that Vienna exists (and that it does not). Similarly, if we were to stipulate that 'plub' means the same as 'is tall or is not tall or is thinking of', then our sentence would be analytic and its denial would entail both that Vienna exists and that it does not. Thus in each case our sentence, although psychological, fails to exhibit the characteristic.

We can, further, suppose 'slub' to be a different two-place predicate meaning the same as 'is tall and is not tall and thinks'. Thus the sentence

George slub that the cat is on the mat.

will entail

The cat is on the mat,

and our sentence, although once more psychological, will fail to exhibit the second of Chisholm's logical marks. And since our sentence will also entail and be entailed by any sentence of the form

George slub that --- is on the mat.

it will also fail to exhibit the third logical mark. For our sentence will have to have the same truth-value as any other sentence having such a form. Such cases do not, however, constitute counter-examples to Chisholm's claims; for it is clear that we may construe each of our coined two-place predicates as technical terms.

A somewhat different sort of case is possible. Let us stipulate that the term 'zlub' is a two-place predicate meaning the same as 'is tall or hit'. Then neither the sentence

George zlub Paul.

nor its denial entails either that Paul exists or that he does not. So in this case the sentence exhibits Chisholm's logical characteristic in spite of its clearly failing to be psychological. The sentence again does not provide us with a counter-example to Chisholm's claims, however, for it is clear that the sentence is expendable in favor of a non-intentional sentence, namely

George is tall or George hit Paul.

[Back]

{10} It might be argued that using this notion of technical terms in so modifying the thesis leads to certain counter-intuitive results. This one might argue as follows. The word 'table' is used to characterize and pick out a certain sort of object not only in terms of the physical features of such objects, but all in terms of the intentions and beliefs of persons who build and use such objects. Thus any sentence which contains the word 'table' will entail some sentence which makes reference to certain intentions and beliefs of builders and users of tables. Since these entailed sentences will be intentional, it follows that the word 'table' is a technical term. It does not appear, however, that a sentence such as 'Tables are often rectangular in shape.' is psychological.

In advancing such an argument, however, we have accepted as a premise that when we use the word 'table' we are, in effect, making reference to the intentions and beliefs of builders and users of tables. If, therefore, we wish to show by means of this argument that 'table' is a technical term, then we must accept that sentences containing it are implicitly psychological. If, on the other hand, we do not accept that 'table' is a technical term, then there is no need for us to maintain that sentences containing it are psychological. (Cf. Chisholm's remark quoted at the middle of p. 7, supra.)

It might be argued that if we do accept that sentences about tables make reference to the intentions and beliefs of builders and users of tables, then the distinction between psychological subject matter and non-psychological subject matter collapses. But even if we adopt the view that sentences about artifacts are implicitly psychological, this leaves us free to argue that sentences about, say, stones and solar systems are not. Thus the argument outlined in the first paragraph of this footnote does not seem to commit one to the rejection of this distinction, but rather a non-standard (and possibly counter-intuitive) way of drawing the distinction. [Back]

{11} An alternative revision of our original condition would require that any sentence containing a technical term entail some non-analytic intentional sentence. This alternative seems, however, unduly restrictive; for there may be sentences containing technical terms which do entail simple intentional sentences, all of which latter sentences are analytic. Examples of such sentences, namely, ones which are simple, intentional and analytic, might be 'Stones never think of Vienna.', and '~(Ex) x is thinking of Vienna and not thinking of Vienna at the same time)' (construing 'is thinking of Vienna and not thinking of Vienna at the same time' as a single predicate expression). Since any analytic sentence entails no non-analytic sentence, it would seem that any intentional sentence that is entailed by an analytic sentence containing a technical term will itself be analytic.

According to the construction of 'technical term' which we have adopted, not only such terms as 'cat-perceptive', which are (according to Chisholm) coined specifically in order to avoid the use of intentional language, will be technical, but also such terms as 'think', 'believe', and 'perceive' which ordinarily render the containing sentence intentional. This is not, however, a disadvantage because it does not interfere with the independence of the test. In fact, the admission of such "ordinary language" terms as members of Chisholm's special vocabulary may be useful in dealing with certain putative counter-examples to Chisholm's claims. Kenny, for example, asserts that 'Diogenes knows an honest man.' is a counter- example to Chisholm's thesis because it is both psychological and non-intentional. [Kenny 198] Similarly, Brown claims that 'John is thinking.', since it is both psychological and non-intentional, runs counter to the thesis. [Brown (2) 130] One might answer these claims by arguing that 'knows', in the relevant sense, and 'thinking', are technical terms. To show this one would have to show that every sentence in which either term occurs entails some intentional simple sentence. One might argue, for example, that the two sentences entail, respectively, 'Diogenes knows that he has met an honest man.' and 'John is thinking about some item.', each of which is intentional.

The adequacy of this proposal will be examined further in sections V and VI. What is important for our present purpose is simply that we have some test which satisfied the condition of independence. [Back]

{12} For this reason, Chisholm's thesis of intentionality must be distinguished from the thesis defended by Kenny, who is concerned "to find what formal property, if any, belongs always and only to descriptions of psychological phenomena." [Kenny 194. Emphasis supplied.] For the fact that a sentence contains a technical term, or the fact that it is expendable in favor of a non-intentional sentence, is not, for Kenny, a formal property belonging to the sentence. Thus Brown appears to be mistaken in supposing that Chisholm, like Kenny, wishes to find a set of formal necessary and sufficient conditions for a sentence to be psychological. [Brown 126] It does not run counter to Chisholm's claims, therefore, to point out that not all and only psychological sentences exhibit the characteristic. [Brown 127-130]

While O'Conner, in contrast with Brown, appears to take into consideration Chisholm's claims about technical terms and the expendability of non-psychological, intentional sentences in favor of non-intentional ones, O'Conner goes on to claim that "whether or not such claims are true is a question quite separate from Chisholm's account of intentional language." [O'Conner 175] Thus O'Conner proceeds to discuss Chisholm's thesis without any further reference to these claims. Since Chisholm makes explicit mention of these claims in formulating his thesis, it is difficult to see why O'Conner believes them independent of "Chisholm's account of intentional language." [Back]

{13} It can also be shown without difficulty that the conjunction of sentences (6) and (7) is interdeducible with the conjunction of sentences (6') and (7'), from which it is clear that no logical advantage can be gained by using the latter pair instead of the former.[Back]

{14} It is clear that Chisholm's assertions do entail that certain conditions are sufficient for a sentence to be psychological, and others for a sentence to be non- psychological. For by contraposition of Chisholm's claims, we can infer that if a sentence is intentional and not expendable in favor of a non-intentional sentence, then it is psychological; and that if a sentence is not intentional and contains no technical term, then it is non-psychological. This does not justify us, however, in asserting that Chisholm believes that the material equivalence expressed by (6''') (and by (7''') holds. [Back]

{15} The troublesome sorts of sentences which have led us to reject (10) and (11) in favor of (10'') and (11'') were sentences which are non-psychological, and non-intentional, and contain instances of our special vocabulary, and sentences which are psychological, intentional and expendable in favor of non-intentional sentences. The strategy adopted in so modifying (10) and (11) has been simply to rule out the possibility of such sentences. An alternative strategy would be to stipulate special conditions which such sentences, if they exist, must satisfy, and thereby render their existence no longer contrary to the purport of Chisholm's thesis. The following conditions might be proposed. If a psychological sentence is expendable in favor of a non-intentional one, then that non-intentional sentence contains a technical term. Similarly, if a non- psychological sentence contains a technical term, then there is a non- intentional sentence containing no technical term in favor of which the sentence in question is expendable. Thus we might propose

(x) [Sx --> [Px --> ((Ix v (Ey) (Ty & Kxy)) & (z)((Sz & ~Iz & Exz) --> (Ew)(Tw & Kzw)))]]

as an alternative to sentence (10''), and

(x) [Sx --> [~Px --> [(~Ix v (Ey)(Sy & ~Iy & Exy)) & (z) ((Tz & Kxz) --> (Ew)(Sw & ~Iw & Exw & ~(Eu)(Tu & Kwu)))]]]

as an alternative to sentence (11"). For these sentences provide special conditions for our troublesome cases instead of simply ruling them out.

Although Chisholm's own formulations do not lend support to our using one rather than the other of our strategies in dealing with these troublesome cases, each alternative appears to have advantages and disadvantages. The existence of a psychological, intentional sentence which is expendable in favor of a non-intentional sentence would provide a reason for preferring a formulation of the thesis containing our alternative to sentence (10"), rather than either (10'') itself or (6'''). For the existence of such a sentence is incompatible with the latter, but not with the former. An example of such a sentence might be 'John perceives a cat.' which, one could argue, is expendable in favor of the non-intentional sentence 'John is cat-perceptive.'. (For the meaning of 'cat-perceptive' could be so specified as to render these two sentences equivalent in meaning.) The existence of a non-psychological, non-intentional sentence containing a technical term, similarly, would provide a reason for preferring a formulation of the thesis which contained our alternative to sentence (11'').

It is worth pointing out that if a non-psychological, non-intentional sentence containing a technical term existed, we would be unable to adopt even our alternative to sentence (11" ) unless we were prepared also to reject either our test for technical terms or our test of expendability. For let us suppose that A is a non-psychological, non-intentional sentence containing a technical term. If we accept the truth of our alternative to sentence (11''), we can infer that A is expendable in favor of some non- intentional sentence containing no technical term. If we accept our two tests, however, we can infer that any sentence in favor of which A is expendable is equivalent in meaning to A, and therefore entails A. Since A, moreover, contains a technical term, it also entails some intentional sentence. It follows that any sentence in favor of which A is expendable entails some intentional sentence, from which it follows in turn that any such sentence contains a technical term. While our alternative to sentence (10'') appears to enjoy a certain advantage over (10'') and (6''' ), unless a different test for technical terms or a different test for expendability is forthcoming, it does not appear that our alternative to sentence (11'') can be used in formulating a workable version of the thesis. For this reason, our alternatives to sentences (10'') and (11'') will not be discussed in the remaining sections. [Back}

{16} If we accept that every sentence is expendable in favor of itself, it can easily be shown that if the first disjunct of sentence (14) is satisfied, then the second disjunct is as well. Given this result, it can be shown that sentence (14) is interdeducible with the following simpler sentence:

(x) Sx --> [~Px --> (Ey)(Sy & ~Iy & Exy & ~(Ez)(Tz & Kyz))]].

We shall, however, retain sentence (14) since it, unlike this simpler formulation, is parallel in structure to our other modified sentences. [Back]

{17} It might be argued that the device of coining such terms is not an important objection to our test because in artificially coining such terms we are in effect changing the language under consideration by expanding its vocabulary. Two points may be made in this connection. First, Chisholm initially introduces the notion of technical terms just in order to handle certain sorts of artificially coined expressions. (Cf. Chisholm (17) 513].) And second, to argue in this fashion would presuppose that our thesis is intended to advance claims only about a given language (say, English) as it stands at a given time. Such a restriction would, in effect, deprive the thesis from having any general philosophical importance, for it could then be argued that if the thesis is true, that fact reflects no more than a linguistic accident. [Back]

{18} Strictly speaking, it is possible, by choosing suitable predicate expressions, to construct certain universally quantified sentences which, although they must be understood as asserting something about an indefinitely large number of things, nevertheless admit of conclusive verification. An example might be the sentence

(x) [x is a marble in this bag --> x is purple].

For this sentence, although conclusively verifiable, asserts of everything whatever that either it is not a marble in this bag or it is purple. The universally quantified formulations of the thesis do not, however, contain predicate expressions which allow for conclusive verification in spite of their universally quantified form. [Back]

{19} The force of this point is exhibited by our ability to reexpress, for example, the conjunction of (10''') and (11"') as the single universally quantified expression

(x) [Sx --> [Px --> [((Ix & ~(Ey)(Sy & ~Iy & Exy)) v (Ey)(Ty & Kxy))) & (~Px --> ((~Ix & ~(Ey)(Ty & Kxy)) v (Ey)(Sy & ~Iy & Exy))]]]. [Back]

{20} These restrictions apply even if we adopt the weaker test for technical terms originally proposed in section II. [Back]

{21} Given an effective procedure for picking out technical terms, the situation would be considerably more encouraging. For in such a case, we would no longer be logically prevented from showing of any given sentence that it satisfied any one of our formulations. [Back]

{22} One might suppose that these remarks provide us with arguments for simply rejecting these formulations and accepting, instead, the conjunction of the alternatives to sentences (10") and (11'') discussed in the final footnote of section IV. By arguments parallel to those above, however, it is easily shown that the same difficulties face this alternative formulation. For to show, of any sentence whatever, that it satisfies, or that it fails to satisfy, this conjunction, we would be required to show either that something is true of indefinitely many sentences, or that a given sentence contains, or fails to contain, a technical term. [Back]

{23} As Chisholm has formulated his claims the logical characteristics which he invokes include certain entailment relations among sentences holding or failing to hold, the expendability of certain sentences in favor of others, and the fact that certain sentences contain terms of a specifiable sort. In the course of our attempt to make these claims more explicit, we have, in effect, proposed to account for both expendability and the notion of a technical term solely in terms of entailment relations among sentences. Thus the logical characteristics with which we have been concerned have all been, at bottom, a matter of entailment relations. Although such a reduction may be unwarranted or uncalled for, some way of making these notions explicit is required which does not itself make reference to the use of linguistic expressions to describe psychological phenomena. [Back]

{24} It is clear that Chisholm's strategy is a reasonable one given his unmodified version of the thesis as expressed, for example, by the conjunction of (6) and (7), or of (10) and (11). It is only after certain required modifications have been introduced that the strategy no longer seems workable. [Back]

{25} Several different proposals for a definition of intentionality have been advanced by Chisholm since the publication of "Sentences about Believing," two of which will be discussed in sections III and IV of Chapter Two. [Back]

{26} Cf. Quine's formulation of the thesis of intentionality as the thesis that "there is no breaking out of the intentional vocabulary by explaining its members in other terms." [Quine (3) 220] We may understand Quine's intentional vocabulary as consisting in the set of all and only technical terms. [Back]


[Table of Contents] [Go to Chapter 2]