Content and Consciousness

by Daniel C. Dennett

  The nineteenth-century psychologist-philosopher Franz Brentano, in his Psychologie vom Empirischen Standpunkt, claimed to have formulated an exact and useful distinction between mental phenomena and physical phenomena, and it is this distinction that forms the basis of the argument. Mental phenomena, according to Brentano, exhibit Intentionality,1 a term he revived from medieval philosophy. ‘Every mental phenomenon is characterized by what the scholastics of the Middle Ages called the Intentional (and also mental) inexistence (Inexistenz) of an object (Gegenstand), and what we would call, although in not entirely unambiguous terms, the reference to a content, a direction upon an object (by which we are not to understand a reality in this case) …’2 Brentano’s thesis divides roughly into two parts – although how clearly Brentano saw the division is hard to say. Some mental phenomena are ‘directed upon’ an object (and these objects have unusual characteristics), and other mental phenomena are related to a content or proposition or meaning. There is some difficulty, as we shall see, in welding these two parts into a single characteristic of Intentionality, and yet intuitively Brentano’s insight is about one characteristic, and an important one.

  Subsequent investigators have annexed to this distinction the claim that no statement or statements about non-Intentional phenomena can have the same truth conditions as any statement about Intentional phenomena. Since, roughly speaking, the domain of statements about Intentional phenomena is the domain of statements expressed in what I have called the mental language, and the domain of statements about non-Intentional phenomena includes all discourse in the physical sciences, the force of this claim is that no systematic correlation of sentences of the sort envisaged in the previous chapter is logically possible. As Brentano’s characterization of the distinction between Intentional and non-Intentional is clarified and modified, the strength of this claim will become evident, and unless a way of refuting it is found, the metaphysical brush-clearing of the previous chapter will have been to no avail, since the Intentionalist thesis, quite independently of any arguments about the ontological status of mental entities, proclaims an unbridgeable gulf between the mental and the physical.

  Brentano’s point about direction upon an object is this: One cannot want without wanting something, imagine without imagining something, hope without hoping for something, and yet the object in all these cases does not, or need not, exist in the fashion of objects of physical actions, such as lifting, touching, sitting upon. Thus if I want a wife it not only does not follow that there is a wife I want, but also does not follow that there is a woman whom I want to marry, any more than it would follow from the fact that I want a space ship that there is a space ship that I want. On the other hand, if I hit a wife or take a ride in a space ship, it follows that there is a wife I hit and a space ship in which I ride. The objects of wanting are thus said by Brentano to have Intentional inexistence, and the same holds true for the objects of imagining, remembering, hoping and so forth. I can imagine a sphinx, remember the dead, and hope for a cure to the common cold, none of which objects exist – at least in the ordinary way. Brentano says these objects have ‘inexistence’ but it is not altogether clear whether Brentano meant by his prefix ‘in-’ that these objects enjoy a form of non-being, or existence in the mind, or both (Cf. Anselm’s ‘in intellectu’) – but in any case it is a queer sort of existence he had in mind.

  His point about relation to content is that in addition to believing in ghosts (a case of direction upon an object) we also believe that …, and hope that …, and in these cases there is not so much an object directed upon as a proposition related to. It was, perhaps, but certainly should not have been, Brentano’s thesis that these propositions stand in the same relation to mental phenomena as the Intentionally inexistent objects mentioned above, that believed propositions enjoy the same kind of queer existence in relation to mental phenomena as hoped-for rains and believed-in ghosts. A moment’s reflection on the different status of the believed-in ghost and the believed proposition ‘there are ghosts’ should convince us that this is a blind alley. In fact, Brentano’s proclivity to talk in terms of strange objects is in general more trouble than not, and an effort will be made to make his point entirely in terms of relation to content.3

  The first step in reformulations of the Brentano thesis has usually been to raise the subject level of the discussion from phenomena to talk about phenomena, turning the distinction into a matter of how we describe or allude to certain phenomena in our ordinary language. Thus Chisholm says that ‘we can formulate a working criterion by means of which we can distinguish sentences that are Intentional, or are used Intentionally, in a certain language from sentences that are not’4 (my italics). This procedure coincides nicely, of course, with our deliberate ontological blindness; we do not suppose that there are any actual phenomena (thoughts, beliefs, desires) for Intentional sentences to be about. Chisholm proposes three independently operating criteria for Intentional sentences.5

  (1) A simple declarative sentence is Intentional 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.

  For example, neither ‘I want a space ship’ nor ‘I do not want a space ship’ implies either that there is or that there is not a space ship, and hence both sentences are Intentional.

  (2) Any noncompound sentence which contains a propositional clause … is Intentional provided that neither the sentence nor its contradictory implies either that the propositional clause is true or that it is false.

  For example, neither ‘I hope that it will rain’ nor ‘I do not hope that it will rain’ implies that it will or will not rain, and hence both sentences are Intentional.

  (3) If A and B are two names or descriptions designating the same thing or things, and sentence P differs from sentence Q only in having A where Q has B, then sentences P and Q are Intentional if the truth of one together with the truth that A and B are co-designative does not imply the truth of the other.

  A familiar example is Quine’s: although Tully is identical with Cicero (‘Tully’ and ‘Cicero’ name the same individual), from ‘Tom believes Cicero denounced Catiline’ it does not follow that ‘Tom believes Tully denounced Catiline’ is true, since Tom may not know or believe that Tully and Cicero are one.

  Chisholm’s three criteria come close to reproducing Brentano’s distinction, but a few alterations must be made to meet a host of apparent counterexamples.6 First, sentences containing such verbs as ‘hunt’ and ‘search’, which are not obviously mental terms, are usually construed to meet (1), as

  (4) Ponce de Leon was searching for the Fountain of Youth shows.

  The usual line with these, which I shall follow, is to reconstrue ‘mental’ rather more broadly, or replace it with ‘psychological’, and accept such sentences as falling within Brentano’s notion of Intentionality and on a par with the sentences with a more obviously mental subject matter. A point of contact can be seen here between philosophical and psychological policy. ‘The rat is searching for an escape route’ is as much to be avoided by the rigorous behaviouristic psychologist as ‘The rat hopes or believes …’, for ‘search’, in virtue of its Intentionality, is a non-observational term, just as much as the more obviously mental terms, and hence has no place in the ‘pure’ data language of the behaviourists. If a term like ‘search’ is to be used at all by these psychologists, it must be defined in observational terms, all of which are non-Intentional. The difficulty psychologists have had in providing these definitions comes as no surprise to the Intentionalist, of course, for he has an argument to show such definitions to be strictly impossible.7

  Counter-examples that force a different adjustment in Brentano’s thesis are the case of clearly mental expressions which fail to meet the criteria. Thus

  (5) Tom perceives the cat

  implies the existence of a cat, vio
lating (1), and

  (6) John knows that Smith is a lawyer

  implies that Smith is a lawyer, violating (2).8 There are other kindred cases, such as ‘discover’, and ‘recognize’; and ‘is under the illusion that’ and ‘hallucinates’, which imply the falsity of the clause or non-existence of the object. ‘Fears’ might be seen as a fence-sitter, since ‘Tom fears the bogyman’ might be held to be true without implying the existence of non-existence of any bogyman, while ‘Tom fears dogs’ or ‘Tom is afraid of the dark’ might be seen to imply the existence of dogs and the dark. ‘Foresee’ in the present tense does not imply the truth of the clause (‘She foresees that there will be a great disaster’) but ‘Noah foresaw that there would be a great flood’ might be held to imply the occurrence of the flood.

  The idiosyncrasies of these ordinary expressions and their failure to meet our criteria should neither surprise nor discourage us – nor for that matter should they engage our interest for long. These are clearly mongrel terms, part mental or psychological, part contextual or epistemic. Thus ‘know’ has in addition to its implications regarding the psychological state or attitude of the knower the implication that what is known is true. And ‘John sees the dog’ speaks not only of John’s state of mind, but also of his position in the physical environment. We could further muddy the waters by inventing new expressions which would be similar to such paradigm Intentional expressions as ‘believe’ and ‘want’ but violate the criteria. Thus if ‘fwant’ had the implication that what was fwanted was in one’s pocket all along, it would violate (1), and if ‘beglieving’ was restricted so that people beglieved that you could square the circle or travel faster than light, it would violate (2).

  The best way with these is to allow that they are Intentional in virtue of the fact that they have Intentional implications. ‘John knows that p’ implies ‘John is under the impression that p’, and ‘John perceives x’ implies ‘John seems to himself to perceive x’, and these meet Chisholm’s criteria.9 This treatment is in some ways parallel to that of Husserl and later Phenomenologists, who speak of the epoché or ‘bracketing’ of external, real-world implications, a practice that is held to get us to the true mental phenomena. In addition to the mongrel terms mentioned, there are a host of other expressions with less direct Intentional implications. Thus ‘John signed the contract’ implies a number of things about John’s beliefs, as do ‘John was introduced to Mary’ and ‘Tom knows Mr. Smith’. The picture that emerges is one of an arsenal of quasi-mental, quasi-psychological terms with Intentional components, and these components can generally be given a satisfactory analysis in terms of the ‘pure’ Intentional terms, such as ‘believes’.

  Failure to recognize this situation can lead to spurious philosophical puzzles. If ‘know’ is held to refer purely and simply to a psychological state, we get the following absurd exchange: ‘I know that John is in Boston’ ‘But he isn’t. He changed his mind and left Boston yesterday’ ‘That’s strange. I could have sworn that I knew that, but I see now that I didn’t. I must have misidentified my mental state. I must be careful in the future to keep a close watch for the true earmarks of states of knowledge.’ It can be argued that the entire Cartesian epistemology is bedevilled by this mistake. The absurdity of treating the mongrel terms as referring neat to psychological states can be brought out even more clearly in this exchange: ‘I hate Stalin’ ‘But Stalin is dead’ ‘Oh! Then I guess I don’t hate Stalin, since that would imply his existence.’

  Another challenge to our criteria is Geach’s example:

  (7) I owe John a horse10

  the truth of which does not imply the existence or non-existence of a horse, satisfying (1). This can properly be viewed as Intentional, however, in virtue of its psychological implications, concerning promising, stating, believing and so forth. We are not through the counterexamples yet, however. Quine offers several more:11

  (8) ‘Tully was a Roman’ is trochaic

  does not admit substitution of ‘Cicero’ for ‘Tully’ thus apparently satisfying (3), and by no stretch of the imagination is (8) about the mental life. (8) also apparently satisfies (2) as

  (9) ‘Tully was a German’ is trochaic

  indicates, provided we allow the phrase in inverted commas to count as a propositional clause. As Quine points out, however, direct quotation is best viewed as making names of expressions out of expressions, so that the ‘Cicero’-‘Tully’ substitution is an illicit alteration within a name, and (8) and (9) are then simply sentences about two different trochaic sentences. This effectively rules out the family of counterexamples based on direct quotation, but others are generated by the alethic modalities of necessity and possibility.

  (10) 9 is necessarily greater than 7

  is true, but substitution via the identity ‘9 = the number of planets’ gives us

  (11) the number of planets is necessarily greater than 7

  which is false. So (10) meets criterion (3), but is not about mental phenomena. Unlike the others, this counterexample cannot be legislated away without controversy. On the one hand there are the various systems of modal logic which would construe (10) and its kind in such a way that the implication to (11) would be blocked while still preserving in some form the principle of substitution of codesignative expressions salva veritate.12 On such an analysis necessity statements would no longer meet (3).13 Alternatively one can be sceptical of modal logic and take the way out suggested by Quine.14 One can claim that if ‘necessarily’ means anything, it qualifies statements or sentences, not, as in the example, relations like greater than. (10) then, must be revised to read

  (12) ‘9 is greater than 7’ is necessarily true

  and by our convention for direct quotation the proposed substitution for ‘9’ is prohibited. If neither of these suggestions can be made to stick, we shall simply have to enlarge the Intentionality thesis to include sentences not about psychological phenomena. This would mean abandoning Brentano’s claim to have discovered the hallmark that distinguishes the mental from the physical, but the crucial argument – that no truth criteria for Intentional sentences can be formulated in the terms of physical science – will not be abrogated; its conclusion will merely be extended to cover areas outside the mind-body problem.15

  To sum up, the effect of all these counterexamples is to bend the Brentano thesis into something quite unlike the original. As adjusted, Intentionality is not a mark that divides phenomena from phenomena, but sentences from sentences – and whereas Brentano associated Intentionality with the mental, we have given it a broader association with the psychological, and are prepared even to abandon this claim if a suitable home cannot be found for the non-psychological modalities. It will help to bend the thesis just a bit more, in an effort to unify Brentano’s dual point about objects and content. Chisholm’s three criteria preserve this duality: (1) is about direction upon an object, and (2) and (3) are about relation to a content. Criterion (1) could be dropped and the distinction of Intentionality could be made entirely in terms of the truth and reference of propositional clauses, if we could replace object-sentences with sentences of propositional attitude. Some object-sentences transform quite gracefully. ‘I am hoping for peace’ becomes ‘I am hoping that there will be peace’ and ‘I believe in goblins’ becomes ‘I believe that there are goblins’. Others are at best awkward. ‘I want a wife’ might go easily into ‘I wish that I had a wife’, except wishing and wanting are distinguishable, so ‘I want it to be the case that I have a wife’ is preferable but cumbersome. One could as well use ‘I want that I should have a wife’, which is reminiscent of Damon Runyan (‘You want I should lean on him, boss?’) and so is not entirely non-ordinary; or, even more barbarously, one could say ‘I want that I have a wife’. It is not to be expected that our ordinary language will provide natural expressions to serve as translations in every case. Still less is it to be demanded that for every mongrel Intentional expression there must be an ordinary language expression that
satisfactorily analyses the Intentional component of the mongrel expression, for this would require a sort of systematic perfection in ordinary language that there is no reason to suppose must exist.

  Some object-sentences cannot be translated into single propositional-attitude-sentences at all. No propositional paraphrase of ‘John hates spinach’, for example, is remotely convincing as a translation. ‘John believes that spinach nauseates him’ and ‘John wants that he is not served spinach’ say both less and more than the object-sentence.16 A large enough collection of such partially successful paraphrases might serve, either in some strict alternation and conjunction system or in the looser ‘family’ way, as a suitable transformation of the sentence, but there is no particular payoff in setting out to elaborate these collections for each and every object-expression.