RICHARD

SWINBORNE

THOUGHT

(Received 19 November, 1984) The aim of this paper is to investigate the nature of thought, and in particular the nature of those passive thoughts which suddenly strike a person and which he recognizes as expressing a view of his about the world - such as that today is Monday or that there is no money left in the bank. My contention is that thought is unanalysable in other terms, and in particular in terms of sensations (auditory, visual or whatever). Man differs from a complex machine not merely in having sensations but in making judgments about the world.

In order to argue m y thesis, I need briefly to set up a taxonomy of the mental life, which will enable me to distinguish thoughts from the other items which it contains. I suggest that we distinguish the mental as the realm of those states, events, etc, to which the subject has privileged access. By the subject having privileged access, I do not mean that he has incorrigible knowledge, but only that however well informed an outsider is about the subject's mental life, the subject is always able to know better what are his mental states and events than can any outsider. Thus, plausibly with respect to any sensation, such as pain, whatever ways others have of finding out whether he has a pain (by study o f behaviour and brain states), the subject can also use; but he has a further way - through experience of the pain. An event, state, episode, etc, is a mental state etc. if its subject has a way of knowing about it (of logical necessity) unavailable to others. Within the realm of the mental we may then make a distinction between items o f two kinds - mental states, in which a subject may be while not at the time being in any way aware that he is in those states; and conscious episodes, o f which the subject must be in some degree aware while he is the subject of them. The two obvious examples of mental states are beliefs and desires. My belief that Edinburgh lies to the North of London is something Philosophical Studies 48 (1985) 153-171. 0031-8116/85.10 1985 by D. Reidel Publishing Company

154

RICHARD

SWINBU RNE

which I have had for many years, including while I am unconscious. Desires for fame and fortune, success and revenge, likewise last for a long period including times while their possessor is totally unaware of them or anything else. A person's being in such a mental state clearly involves his doing certain actions, or having certain experiences from time to time or under certain conditions. My believing that my home lies in a certain direction involves my proceeding in that direction when I have the purpose to go home. But the belief still exists when there is no occasion to manifest it. Beliefs and desires are mental in that they are not just dispositions to act, but states of which the subject can become aware by asking himself what he believes about so and so, what he desires in some respect, and the answer which he naturally gives to himself puts him in a better position than others to know about his beliefs and desires. At least, so I would contend. If any one denies this and claims, for example, that desires are solely matters of the patterns of public behaviour, then he is claiming that they are not mental states in my sense. This will not effect my distinction between mental states and conscious episodes, which is my primary concern; it will simply mean that my examples of the former are unsatisfactory. Conscious episodes by contrast are goings-on of which a man is necessarily to some extent aware while he has them. Thus a pain is a conscious episode because a pain would not be a pain unless its subject was in some degree aware of having it. I distinguish among conscious episodes between the active, which consist in the agent intentionally bringing about something, and the passive, those which happen to a person and do not involve the subject's agency. A person trying to do something is an active conscious episode, but most conscious episodes are passive. They consist in something happening to a person. Visual, auditory etc. sensations are like this. And further, most conscious episodes are not merely passive, but are involuntary - i.e. not brought about intentionally by the subject. Most of the auditory sensations of which I am aware are not merely passive, but involuntary. By contrast my having a visual image, although passive (consisting in something happening to me) may sometimes be volunatry, in the sense that I bring it about that it does happen to me. Among passive conscious episodes we must distinguish those, the true descriptions of whose content are referentially transparent and those which are not. Sensations, I claim, are of the former kind, thoughts of the latter. Paradigm cases of my sensations are my experiencing the patterns of colour in

THOUGHT

155

m y visual field, sensations of taste or smell, mild aches and pains, sensations o f hot or cold in parts of my body; together with their pale imitations in m y remembering memory images or m y imagining imagined images. From any true descriptions of the content o f m y sensations (what I experience) you can get another true description by substituting terms with the same reference and logically equivalent sense. If m y sensation is o f six rows of four dots, then m y sensation is of twenty four dots in six rows. Further, there may be aspects to m y sensations of which I am quite unaware until my attention is drawn to them (aspects which are nothing to do with the relations o f the sensation to something outside itself) - e.g. that the lines in m y imaged figure are o f equal length, or that so and so is written on the third line of m y eidetic image of the page o f a book. Thoughts, unlike sensations, have a propositional content (in the sense that their content can be captured by a sentence of some language, not necessarily a language known to the subject), and this content is referentially opaque. My thought that today is Monday is not a thought that April 16th is the second day o f the week, even if, as a matter of fact (unknown to me) today is April 16th and Monday is the second day of the week. For the content of m y thought is simply that of which I am aware. Many states or events which count as mental on m y definition are, complex in that they include distinguishable conscious episodes and mental states or the acquisition or loss of these, often together with elements which are public in the sense that the subject does not have privileged access to them. Thus emotions, I suggest, consist o f a complex mixture of sensations, desires, beliefs and thoughts. Again, normal perception seems to me to include both a public element, a sensory element, and the acquisition o f a belief. My perceiving a table believing that I do (in contrast to the derivative case of m y perceiving a table, believing it to be something else), consists, I suggest, in the table (a public thing) causing me to have certain visual sensations which I believe to be caused by the table. The exact analysis o f perception is of course a contentious matter. I give m y preferred analysis only to suggest by an initially plausible example that if we separate off from some mental states and events public elements, we get mental states and events o f purer kinds.

iI My next concern is to distinguish the kind o f thought with which I am con-

156

RICHARD

SWINBU RNE

cerned, and which I shall call occurrent thought; from other kinds o f thought, and to bring out their relations to occurrent thought. My concern is with the thoughts which are occurrent mental events. They are the thoughts that occur to a subject at some particular moment of time. Crucial among such thoughts are those which he sees as expressions of his belief, and which, following philosophical tradition, I shall call 'judgments'. They are the thoughts which strike a subject, which he sees as "welling up" in him as his comment on the world - such as, for example, the thought that today is Monday, that I ought to be in London now, that it is m y wife's birthday, or that Labour is going to win the next election. Not all occurrent thoughts are judgments. There are the thoughts which express mere possibilities which the thinker entertains, or questions which occur to him or commands which he feels to be impinging on him. Judgments if captured in a sentence will be captured in a declarative sentence (that something is the case), but not all occurrent thoughts have a declarative sentence as their linguistic expression. There are also, as well as judgments the occurrent thoughts caused to occur in the thinker as the (believed) content of the spoken or written words of others. Your saying the words 'today is Monday' will cause me to have the occurrent thought that today is Monday, without that thought being a judgment of mine. Occurrent thoughts may also be brought about in a subject by his own intentional action either as a private act or as the content o f a public communication. If I ask a question or issue a command aloud, or tell a lie, I normally have the corresponding occurrent thought, but it is not a judgment. We must distinguish occurrent thought o f any of these kinds, first from the thought in which a man may be said to think this or that - that the Earth is flat, or that 5 is the cube route o f 35 - even while he is totally unconscious. In that sense 'thought' is a synonym o f 'belief'. 'He thinks that the John is his brother' is synonymous with 'he believes that John is his brother', and describes a state in which a man may be when he has no relevant occurrent thoughts. Occurrent thoughts typically express beliefs, but the beliefs still exist when there are no thoughts occurring which express them. If the judgment, the occurrent thought which the agent sees as an expression o f his belief, is that so and so, then that is plausibly a primary criterion of the man's belief that so and so. (For that reason, such judgments have often been called - not very happily - occurrent beliefs.) Some philosophers might indeed regard

THOUGHT

157

judgment as the only criterion o f belief; for most other philosophers (as I noted earlier) behaviour is a criterion in addition to judgment, and for some philosophers behaviour is the only criterion. (For the latter philosophers, how a subject sees his beliefs, has no relevance to what those beliefs are.) Secondly, we must distinguish occurrent thought from that 'thought' which characterizes a man's public performance. 1 When we say of a man 'he thinks about what he is doing' we may mean only that he does what he does intelligently. His attention is on his job, and he performs those actions which secure his ends without undesired consequences. When we say that some man thinks what he is doing when he is gardening, we may mean only that he doesn't tread on garden plants, uproots weeds rather than garden plants, waters what is dry and fertilizes what is barren. We are describing the successful character o f his public performance, not goings-on additional to that, such as occurrent thoughts. _Thirdly we must distinguish occurrent thought which is passive from the active process of thinking, which is the process of actively bringing about occurrent thought. Active thought may be of three kinds. First there is the case where an agent produces in himself a thought as a bye-product of communicating it to others in speech or writing. Secondly, there is the case where an agent, thinking privately, repeats to himself some familiar thought or train of thoughts. I may say to m y s e l f - on purpose, to remind myself, to drum the thought home - that today is Monday, or there is no money left in the bank. Or I may run through the steps o f a familiar argument in order to keep it in mind for when I try to lecture on the subject without notes. Such thought is a matter of a man intentionally bringing about the occurrence o f the occurrent passive thoughts. Since, as we shall see, occurrent thoughts are normally mediated by sensations, in the case of private thoughts imaged (usually auditory) sensations, o f words which express the thought, active thought will normally involve the bringing about of such (imaged) sensations. But note that, neither aloud nor privately, could I intentionally bring about an unfamiliar or new thought; for I can only intentionally bring about that which I know how to bring about and so that with which I am already familiar. Thoughts have to occur to a man unexpectedly, before he can rehearse them actively. The third kind o f active thought is when a man intentionally thinks about some subject or problem, normally privately, for a certain period of time. He may choose or decide to think about philosophy for half an hour. He may

158

RICHARD

SWINBU RNE

think about how to spend the summer or about the solution to an equation because he chooses or sets himself to do these things. What is going on here is that he is intentionally bringing about certain passive thoughts (in the way described above), normally by saying to himself certain things, or conjuring up certain images, in the hope that there will occur to him new and unexpected thoughts in the right subject area (e.g. philosophy), or thoughts which provide answers to the questions which concern him (e.g. how to spend the summer). The way to think about how to spend the summer is to ask yourself such questions as 'How do m y friends spend the summer?', 'Where have I spent earlier summers?' etc. in hope that some o f the answers (e.g. 'John goes to Italy') will spark off a thought which you do not actively bring about, but which is brought about by some process over which you have no control (e.g. 'I ought to go to Italy') which constitutes a solution o f your problem. I conclude that active thought (of these three kinds) is analysable in terms o f other more basic constituents of the mental life - intentional bringing about, and occurrent thought.

III

To revert now to m y primary concern - occurrent thoughts. Such thoughts occur all the time, frequently uncaused by the subject. My contention with regard to them is that they are mental events o f whose content-the subject has infallible knowledge (i.e. his beliefs about their content are necessarily true), and which are not further analysable. Clearly thoughts are mental events. If others have means of discovering m y thoughts (by investigating m y brain states or behaviour or environment), I could use the same methods - and yet I could know better than they what I was thinking because I was doing the thinking and so have an additional means of discovering m y thoughts. A subject has privileged access to his thoughts. That the subject has infallible knowledge of his thoughts follows from the very nature o f thought. We pick out some event as a thought (which occurred at a stated time) only if it was something of which the subject was conscious or aware. And what makes a conscious episode a thought is that the subject is aware of it as having a certain content; and the thought just is the thought that has that content o f which he is aware. And whatever might be the case with other o f an agent's beliefs, an agent's beliefs about his thoughts are

THOUGHT

159

surely given by his judgments which state the contents of those thoughts which he is aware o f them as having. This claim that the subject's beliefs about the content of his thoughts are necessarily true, needs however to be guarded by two qualifications. The first is that, although necessarily a subject who has the thought that p, believes that he has the thought that p, he may refuse to admit it - publicly or even to himself. He may refuse to let the thought 'I have the thought that p ' occur; if he allowed his judgments about the content of his thoughts to occur, that would be the judgment which did occur. But the belief that he has the thought that p is a belief which the agent supresses from his consciousness; it becomes subconscious in the sense that he refuses to admit it to himself, not in the sense that he is unaware of it .2 The second qualification t o the claim that a subject's beliefs about the content of his thoughts are necessarily true, is that any description which he may give of them in any language to himself or others, while attempting to describe them correctly, which depends for its correctness as a description on anything outside the thought itself may be a wrong description. In particular any description in some language which depends on the subject's beliefs about the sense or reference of words, or force o f sentences, in that language may be mistaken. I may attempt to report m y thought that John is distinguished by using the words 'John is notorious'. This description of my thought depends for its correctness on m y belief that 'notorious' means 'distinguished'; m y belief is false, and so I have misdescribed m y thought. But I do not have a mistaken belief about the content of m y thought; I did not have the thought that John is notorious, and I did not believe that I did. I had the thought that John is distinguished. I may even have had the thought that John is correctly described as 'notorious'. I am not in error in my beliefs that I had those thoughts. Any description of a thought which depends for its correctness on the subject's beliefs about the reference o f his referring expressions may also be mistaken, if these beliefs are false. Having the thought that the shimmering lake before me is beautiful, I may describe that thought as the thought 'that water is beautiful'. This description depends for its correctness on my belief that 'that water' refers to the shimmering lake in front of me. But it doesn't if the lake is not a lake of water. But again I do not have a mistaken belief about the content of my thought. I do not have the thought that that water is

160

RICHARD

SWINBURNE

beautiful, and I do not believe that I do. For the very reason why I could not be having that thought - that there is no water in front o f me and so 'that water' refers to nothing - also prevents me from having the belief that I have the thought that that water is beautiful. To the extent to which the content of thought is referentially opaque, to the same extent is the subject's belief about that content. The thought which I have is the thought that the lake is beautiful. I may even have the thought that there is water in front o f me which is beautiful. If I have those thoughts I am not in error in m y beliefs that I have those thoughts. But what then is the content to the claim that an agent has an infallible belief about the content o f his thought, if it does not entail that he will give a correct description of it? It is this - that if he knows the meaning and reference o f relevant words and force o f relevant sentences he cannot but give a correct description o f his thoughts if he so chooses: When I misdescribe m y thought, and am given the relevant information about language and its reference, then I m a y redescribe the thought. I will recognize the redescription as no more than that - a redescription o f what I knew already as the content of m y thought. I do not see myself as coming to know more about the content o f m y thought - in the way that I might come to know more about the content o f a sensation (say, an eidetic image) by examining it more closely; only as coming to know better how to express it in language. 3

IV

Finally, I claim, thoughts are not analysable in terms of anything else, actual or hypothetical, i.e. they are not a matter o f something else occurrent or what would happen under various circumstances. Analyses in terms of public phenomena are not plausible. Just as sensations are not analysable in terms of brain events o f certain electro-chemical kinds, neither are thoughts. Whatever view one may take about the contingent identity of mental events with brain events, it is a fact about a subject that he has some thought, additional to any fact about some physical occurrence in his brain. 4 Nor are thoughts analysable in terms of public behaviour (or the brain events which cause this or that public behaviour). It may occur to me 'it must be hot in Singapore'. Yet I have no occasion for going to Singapore, and the thought may not affect m y conduct actual or hypothetical in any way. I may refuse to tell you what I was thinking about and m y having had that thought m a y influence m y

THOUGHT

161

conduct in no other way. Even if m y belief that it is hot in Singapore makes a difference to my behaviour, what difference it will make depends on my purposes. (If I want to go where it is hot, I will go to Singapore; otherwise, not. And so on). My purposes are not visible on the surface of m y behaviour. And beliefs are not the same as thoughts. Any attempt to analyse thought in terms of the difference it would make to public behaviour, for given purposes, proves viciously circular. It may be suggested: 'my thought was a thought that p if and only if when asked if p was m y thought, and I was seeking to tell the truth, I would say " y e s " ' . But in this analysis 'seeking to tell the truth' must be not seeking in general to tell the truth, but seeking to provide a true answer to the question. Yet how can I possibly seek to provide a true answer to the question, unless there is already an answer which is true, and whose truth does not depend on how ! respond to the question? The description of the disposition already presupposes that thoughts exist prior to and independently of the disposition to manifest them in this way. It is somewhat more plausible to attempt to analyse thoughts in terms of mental events of other kinds, viz. sensations. Among sensations are the imaginings of sounds or sights which a man conjures up or which occur to him spontaneously. And without doubt when a subject has a thought he often has auditory images o f many of the words which would occur in a sentence which expressed the thought. When I have the thought that some President is dead, auditory images of the word 'President' and 'dead' flit through my mind. So it is natural to suggest that having a thought consists in the occurrence of images (normally auditory or visual) of words of a sentence which expresses that thought. One way for me to have thought that the President is dead is for the words 'the President is dead' to 'flit through my mind'. I argue that this is neither sufficient nor necessary if I am to have thought in question. It is not sufficient. To start with, the words which flit through my mind may be words of a language which I do not understand. I may have liked certain sounds of a foreign language, and they may occur to me again without my having the least idea what they mean and so without their occurrence constituting the occurrence to me of a thought. How often must the words of the Latin Mass have passed through the mind of pious Roman Catholics o f old (when the mass was said throughout the world in Latin) without those words being the vehicle of any thought. Let us suppose the restriction is added that the words be of a language which I understand, and also (to meet

162

RICHARD

SWlN BURNE

an earlier point) that I have true beliefs about the meaning and force and any reference made by those words. There are then three harder difficulties. First, it is not enough for me to know English for the occurrence to me of the words 'the President is dead' to express a thought of mine that some President is dead. They too may be the words of a song which passes through my mincl without being the vehicle of a thought. A man is usually careful not to say aloud words which he does not mean (for fear of what his hearers might think), but he may say to himself much that he does not mean. Secondly, even given that the words which occur to a subject are words of a sentence which he understands as expressing a thought, the words may often not be enough to determine the meaning of the sentence. In particular they may be quite inadequate to determine the intended referent (for reasons which have nothing to do with the actual surrounding circumstances - see earlier). I speak of 'The President', but which President - of the USA, or of China, or of my local history society? The sentence does not fully capture the thought, for in having the thought I know of which President I am thinking, but the sentence does not make that clear. And, finally, even given that we can determine which thought the words express, we still do not know whether they express a judgment of the subject or whether he is recalling some past conversation or imagining some future one in which another person says something to him and which he understands in a certain way, but which in no way represents his own current belief. Imagined conversation is different from one's own judgments. "The President is dead" may be words which I recall someone saying to me in the autumn of 1963 after the assassination of President Kennedy. Or they may even be words which I recall myself uttering on some previous occasion. And yet the subject knows whether he is making a judgment or merely imagining a past conversation; but the pattern of his sensations does not show this. So which thought he is having(let alone whether he is making a judgment thereby) is not determined by the pattern of the subject's sensations. Something else beside the words in necessary for the thought. Let us see if it can be supplied without making thoughts special unanalysable events. There are two possible such theories in terms of actual context and in terms of hypothetical context. Let us take actual context first. I understand by the actual context of his imaged words the surrounding pattern of a subject's mental events, including his intentions and his perceptions of what is going on around him in the world and other visual images. The suggestion

THOUGHT

163

then goes that what determines whether the words 'the President is dead' are meant as opposed to being just the words of a song would be factors such as whether t have previously been talking to somebody about 'the President' or had a picture in my mind of (say) President Reagan; or by contrast whether 'The President is dead' were the words of a song known to me and whether I was often in the habit of repeating to myself words of songs. What determines that the intended reference is (say) to Harold Jones, President of the local history society, is whether I have a picture of him in mind when I imagine the words. What determines whether the words were a judgment of mine, as opposed to words which I was imagining someone saying, was whether I had visual images of someone out of whose mouth I seemed to hear the words coming. It should be apparent that this kind of analysis will not work. Often there is no context to a man's thought. Thoughts falsh through my mind when I have been thinking of nothing before - men often wake up suddenly with a nagging thought - 'he's dead'. And before any subsequent train of thought could show who 'he' is, the thinker may be distracted by the crockery being dropped downstairs. But just as much as one's own thoughts, there may flash through one's mind the utterances of others, or conversations one would like to have had. And again the sequence may be too short for any context to show whether they are one's own judgments or not. Thought is often so casual and irrational that the context could not show which (if any) thought is expressed by the words of which one has images. We do not always think in paragraphs. Sudden thoughts strike us and then our attention is diverted to something else. I am walking along thinking of nothing in particular, having only the sensations involved in perception of my surroundings. The words flash through my mind 'You are old'. Before more words can flash through my mind, my attention is diverted by catching sight of a colleague. There are many different interpretations which could be given to the words. They might be the beginning of 'You are old, Father William'. They might be part of an imaginary conversation with an elderly colleague and so on. But the crucial point is that I know what is the interpretation of the words, and the surroundings do not dictate the interpretation. Nor can visual images of faces determine the reference of referring expressions. I often imagine people saying things or refer to them without having visual images of them. Or if I do have visual images the images are so vague that in themselves they could be images of almost anybody; what determines

164

RICHARD

SWINBU RNE

of whom they are intended images is not something in themselves but the way I construe them (i.e. the role they play in my thought). The'thought determines of whom the image is an image; the image does not determine of whom the thought is a thought. It may be suggested that it is not the actual context of a subject's imaged words, but the hypothetical context - e.g. how the sequence of imaged words would have continued if the subject's attention had not been diverted. Having a thought would thus be (in part) a dispositional property - like fragility or malleability. I do not think that any such dispositional account of thought will work. For although it may be the case that if I have a certain thought, I must have a certain disposition, and conversely; the description of the disposition already presupposes the thought to be a separate thing from itself and so cannot be what having the thought consists in. In this sense, the account o f thought in terms of having a disposition to have certain mental events (as was a previous account in terms of a disposition to evince certain public behaviour) is viciously circular. Most accounts of this kind are obviously false. The words occur to me: 'You are old'. If the suggestion is made that they constitute the thought that I myself are old if and only if, when my attention is not diverted to something in the outside world (e.g. my catching sight of my colleague), they are followed by imaged words forming sentences expressing propositions about such matters as my hair being grey, my children growing up, etc., the account if false. I may have the thought that I am old, casually, without brooding on it; I may not mind. If alternatively the suggestion is made that the words constitute the thought that I am myself old if and only if when I continue to have imaged words forming sentences but not sentences concerned with a quite different subject then I will have words forming sentences expressing propositions about such matters as my hair being grey, my children growing up, etc., the use of the words 'a quite different subject' makes the account viciously circular. For the account presupposes that there is an independent criterion of what I was thinking about originally. The account cannot be applied unless we can determine by some independent means when the subject of my thought changes; but we cannot do that unless we can determine what I was thinking about in the first place - which was what the account was seeking to elucidate. And, further, both suggested analyses presuppose that we can determine by some independent means which thought propositions are expressed by the sentences which occur to me subsequently.

THOUGHT

165

Thus, suppose that if after having auditory images of the words 'you are old', I had gone on having auditory images of words, I would have had auditory images of the words 'grey-headed'. Now if I continued to think about the same subject, and if the words 'grey-headed' expressed the proposition that I was myself grey-headed, then indeed m y original words constituted the thought that I was myself old. But the analysis presupposesan independent criterion of my continuing to think about the same subject, and which thought is constituted by the words 'grey-headed'. The latter words may have constituted the thought about some colleague that he was grey-headed. I may have been thinking about him all the time, or I may have changed from thinking about me to thinking about him. It is a remarkable fact that despite a considerable philosophical industry devoted to the philosophy of mind, there has been relatively little philosophical writing in the past fourty years on the subject of thought as opposed to belief. To my mind, the most important major study was Professor H . H . Price's Thinking and Experience. s Price adopts a form of the dispositional account of thought. "There is", he writes, "a form of thinking in which nothing else is present to the mind except a succession of (token) words". 6 In such verbal thinking, concepts "manifest themselves by determining the direction which the flow of words takes". Which concept I am expressing by my words, is not just, he holds, a matter o f the actual words used, but of my readiness to substitute other words. "Our concepts manifest themselves in a readiness to produce alternative verbal formulations". 7 If, to use Price's example, when I say to myself "a dodecahedron is a solid figure", I am thereby manifesting our concept of a dodecahedron, "I shall be able to produce other utterances, I shall be able to put it another way. Instead of saying a dodecahedron is a solid figure, I can say 'it is a three dimensional shape' and add 'it has twelve corners', or I can translate my original remark into French or Latin; or again I can draw a rough sketch of a dodecahedron on a piece of paper", s This "readiness" to put it another way, is interpreted by Price not, as I have done, simply as a counter-factual conditional ("if I had been asked, I would have said so-and-so") but, Price stresses, 9 as "a felt readiness". We feel guided or directed, as it were by our thought, to say so and so and to regard so and so as a correct translation of it. However,although there may be a felt tendency for spoken words to continue in a certain way, there may be no felt tendency to repeat other sentences (either to ourselves or publicly) as paraphrases of the original sentence, or to continue the se-

166

RICHARD

SWINBURNE

quence of words in any particular way. A passing thought may be so boring that I have no "felt" tendency to repeat it, aloud or to myself. Nor need I have any "felt" tendency to give some paraphrase to myself or others, if asked; or to acknowledge that a suggested paraphrase does give the content of my thought. For I may forget my thought; and some thoughts are such that I do not wish to acknowledge them even to myself. My "felt" tendencies may be to lie (publicly or even to myself) about my thought. Price's version of the dispositional theory is no improvement on the one which I considered earlier. Such continued failures of dispositional theories suggest that a thought is what it is, independently of what I subsequently do about it. What tempts philosophers to suppose that the meaning of imaged auditory or visual sensations is given by an actual or hypothetical context of some kind is a totally false analogy between the public meaning of a public sentence and the private meaning of a private sentence. A public sentence only has a meaning to the extent to which it sets up similar clear thoughts in hearers who have learnt the language in question. There are various context-dependent devices which determine the reference, sense, and force of public utterances and so ensure that this happens despite any unclarity of sense or reference in the words uttered taken by themselves. I can pick out objects which it may be difficult to describe, by means indexical expressions - i.e. expressions such as 'I', 'he', 'this', 'over there', 'now', 'here', whose reference is determined not just by what is said but by the context in which it is said (e.g. the reference of 'you' depends on which person I am talking to and the reference of 'over there' depends on the direction in which I am pointing.) The context will include the previous sentences uttered by a speaker or the person to whom he was talking. The claim made by 'what you just said is false' will depend on what you just said; and the reference of 'The President' will depend on which President we had previouslybeen discussing. The meaning of a public sentence is a public matter and there are public criteria for determining what it is. But the meaning of words which a subject utters to himself or finds occurring to him is not a public matter; the meaning is whatever the subject supposes it to be; and the fact that other people would understand those words differently if they were said aloud is quite irrelevent. Likewise the private thought had by a speaker or hearer of a public sentence may be a perfectly clear one, even though the public sentence lacks a clear meaning. 1~ Not merely are the auditory or visual sensations of the words of a sentence

THOUGHT

167

of a language known to the subject not a sufficient condition for the expression of a thought, they are not even a necessary condition. For, first and crucially, if a thought is expressed in words, the thought may be expressed in far fewer words than are needed for a grammatical sentence. The word 'old' alone said to myself may serve to express the thought that I am old. Secondly, as I noted earlier, a thinker may attempt to express his thought in a sentence of a public language which does not express that thought. He may use the words 'the water is beautiful', and yet the words of that sentence express no proposition and so can express no thought, when there is no water in front of the speaker. But the words by which he attempts to give it public expression may also be the words, the auditory images of which were the vehicle of his thought, which can only be given expression in a public language by a different sentence. And thirdly it does seem very plausible to suppose that having thoughts is not confined to language users. There are cases of those who learn to use words late in life yet report having thoughts earlier) x And animals often behave just as if they are struck by some thought and yet they have no language. Such thought may well be mediated by picture images in a way described by Price in a dispositional account of thought in picture images. 12 But pictures, no more than words, can bring out fully the content of the thought. Maybe I can have the thought 'crocodiles are dangerous' by having an image of a crocodile darting through the water and catching me by the leg. What makes that image an expression of that thought is for Price, partly 13 a matter of which other images I am prepared readily to substitute for the given image; such other images must be, in the words of Hume which Price quotes, 'present in power'. Although my crocodile image may resemble a particular crocodile very closely, my thought does not concern him alone because I am prepared to substitute very different crocodile images. Nor does it concern me alone, because I am prepared to substitute images of other people getting caught. Yet the same difficulties remain for Price's account of thought in picture images, as for his account of thought in words. I may not have the 'felt' tendency in question - the thought may be too boring. If I asked to give an 'image paraphrase', I may be unable to do so, having forgotten the thought; or I may stop myself doing so,because of a dislike for the thought. Picture images may express a thought, yet there is more to a thought than picture images in context (as we have seen, there is more to it than word images in context.) These considerations so far bring out that there is more in the way of

168

RICHARD

SWINBURNE

occurrent events to a thought than imaged words or other symbols which express it, and that we do not need much in the way of the latter. Do we need imaged words or other sensations at all as the vehicle of thought? I s u g g e s t that we do not; a subject can have a thought without either word images or any other imaged sensations to express it. Since clearly thought goes far beyond the vague sensory content which expresses it, there seems no good philosophical argument to show that it needs the latter at all. In favour of this, I present two further empirical considerations. First, there is the familiar phenomenon of having an idea for which even if we try we cannot find anything like the right words; later the words come, which we recognize as capturing the thought. Secondly, there is the work of the Wurzburg school of psychologists in the first decade of this oentury who got subjects to introspect carefully and report the thoughts which they had and what went on mentally when they had these thoughts. (Often they were asked to report on the thoughts which they had between being given some question and making some reply} a) Subjects continually reported that thoughts came unmediated by sensations. True, reports of introspection are very suspect, liable to be overinfluenced by the philosophical prejudices of the subject or investigator. Nevertheless, there is a wealth of empirical testimony to "imageless thought", which the trend of my argument suggests to be logically possible. If indeed, thoughts are occurrent mental events distinct from sensations of their imaged expressions in spoken words or any other expressions, we need an account of why many, perhaps most, thoughts are expressed in words; and an account of why, although there can be thoughts which a man does not express in words, it is a little odd to suppose that he has thoughts, which he would totally be unable to express in words, albeit imperfectly. It would be a little odd to suppose some New Guinea Tribesman to have the thought captured in Einstein's equation, E = m c 2 , when he could not even do simple arithmetic in spoken or written symbols. So what is the connection between talk and thought? It is, I suggest, this. As we learn a language, we learn to utter public sentences which have a meaning determined by public conventions and we learn to understand public sentences uttered by others, the meaning of which is again determined by public conventions. In uttering and understanding public sentences we come to have the thoughts expressed by those sentences. More generally, in learning the words of a language and the ways of connecting

THOUGHT

169

them, we acquire the ability to have the thoughts which can be expressed by those words and those ways of connecting them. We also acquire the ability to have thoughts in non-linguistic ways. Through learning to react, as animals do, to certain objects and properties, we acquire the ability to have thoughts about those objects and properties often mediated, no doubt, by pictures of them. It is not logically necessary that we acquire the ability to have thoughts in this way, but it is practically necessary. We can only have thoughts which are initially fairly close to our ability to express them in language or practice. But having acquired the ability to have thoughts of certain kinds, we may forget or not need all or even any of the public words which express them. Most of us much of the time do need tags (of single imaged words or vague picture images rather than full sentences) in order to have one thought distinct from some other. But for some of us, especially when the objects of our thought are very familiar ones, the tags become less and less necessary, is In claiming that thoughts are unanalysable mental events, I do not wish to be thought to be denying various obvious truths. First, a thought, being essentially propositional, wouldn't be a thought unless it could be expressed by some language or other. But it does not follow that it can be expressed in a language currently available to the subject. Secondly, I do not wish to deny that thoughts affirm relations between things thought about (be they objects, properties or concepts) which will be picked out by separate words in sentences which express the thoughts. In claiming that thoughts are unanalysable, I mean only that the mental events which are thoughts do not have noncognitive parts. Nor, thirdly, do I wish to deny that we cannot have certain thoughts unless we have certain beliefs. To take Davidson's example, we cannot believe and so have the thought that the gun is loaded, unless we have beliefs such as that "a gun is a weapon, and that it is a more or less enduring physical object"J ~ But this is because a man's thought about a gun is a thought about a weapon which is an enduring physical object and so his thought contains such thoughts, and hence his expression of it is an expression of those beliefs too. A man's thought would be wrongly described as the thought that 'the gun is loaded' unless he had those beliefs. Nor finally do I wish to deny that there is public evidence about what are a man's thoughts in the pattern of his public behaviour and speech. But despite all these obvious points, m y Claim remains that thoughts are unanalysable mental events. It is the role of public language (which we can image privately in sensory images), in developing our private thought, which makes us suppose that thought just

170

RICHARD SWINBURNE

is sub-voce talk. But t h a t is o n l y o n e m o r e e x a m p l e o f t h e g e n e t i c fallacy; c h i l d r e n are n o t t h e same as t h e i r p a r e n t s . T h o u g h t s m a y b e c o m e distinCt f r o m t h e language w h i c h t h e m b i r t h . 17

NOTES 1 Which G. Ryle drew so prominently to our attention in The Concept o f Mind, London, 1949. 2 For self-deception as refusing to spell out to yourself your beliefs, see H. Fingarette, Self-Deception, Roufledge and Kegan Paul, London, 1969, Ch. 3. It is of course Putnam who has drawn our attention in recent years to the point that what a thinker can truly be described as thinking depends on what there is in the world and how it is related to ourselves, as well as what is going o n 'in the thinker's mind' (see e.g.H. Putnam, Reason, Truth, and History, Cambridge University Press, Cambridge, 1981, Chs. 1 and 2.) Putnam uses "believe" and "think" interchangeably, but he would, I believe, regard this thesis as applying both to what I am calling "thought" and to what I am calling "belief", and that seems right. He suggests the introduction of Husserl's device of "bracketing" in order to distinguish the truly mental element of thought, which is my concern in this paper: "If we 'bracket' the belief that there is a glass of water on the table, then what we ascribe to John is simply the mental state of an actual or possible person who believes that there is a glass of water on the table (in the full 'unbracketed' ordinary sense.) ...We will say that he has the bracketed belief that (there is a glass of water on the table). In effect the device of bracketing subtracts entailments from the ordinary belief locution (all entailments that refer to the external world, or what is external to the thinker's mind)" - o p . cir. p. 28. Similarly Dennett describes a subject's "notional attitude" as the element of belief which remains constant in different surroundings. See D. Dennett, 'Beyond belief', in A. Woodfield (ed.), Thought and object, Clarendon Press, Oxford, 1982, e.g. pp. 80f. Dennett and also Searle (J. R. Searle, Intentionality, Cambridge University Press, Cambridge, 1983, Ch. 8) bring out that there is no such thing as de re belief, in the sense of a belief of which the object referred to is itself a component, merely a continuum of ways of describing beliefs with varying degrees of external commitment. 4 From which it is, I suggest, correct to conclude that thoughts, like sensations, are distinct events from brain events. But I do not argue that here. See my 'Are mental events identical with brain events?' American Philosophical Quarterly, 1982, 19, pp. 173-181. s Hutchinson, London, 1953. 6 p. 347. p. 349. 8 p. 349. 9 pp. 350ff. 10 That language is a device for expressing publicly intentional states (such as beliefs and thoughts) which may exist independently of language is a major theme of J. R. Searle, Intentionality, Cambridge University Press, Cambridge, 1983. Searle understands by Intentionality (p. 1) "that property of many mental states and events by which they are directed at or about or of objects and states of affairs in the world". He writes (p. 26)': ''it is not possible to give a logical analysis of the Intentionality of the mental in terms of simpler actions, since Intentionality is, so to speak, a ground floor property of the mind, not a logically complex feature built up by combining simpler elements". And (p. 27): "An utterance can have Intentionality, just as a belief has Intentionality, but whereas the Intentionality of the belief is intrinsic the Intentionality of the utterance is derived."

THOUGHT

171

11 See Wittgenstein's doubts about such a case reported by William James, Philosophical

Investigations, 342. ~2 op. tit., especially chapters 8 and 9. 13 Price holds that the image may be made to represent, not a particular crocodile but crocodiles in general, not merely by the fact that I could easily substitute very different crocodile images for it, but by my crocodile-image having a generic character, like a composite photograph. ~4 For a description of their work in this field, see G. Humphrey, Thinking, Methuen, London, 1951, chapter 2. ~s The doctrine of the 'language of thought', that humans are born with an interior language in which they think private thoughts and then communicate with others by translating in and out of that private language, is obviously relevant here. But just how it is relevant depends on two matters which expositions of the doctrine sometimes leave obscure. The first is whether 'thought' means belief, or thought in my sense. Usually, I think, it means belief. But then it is not too clear what it is for a belief to be in this language as opposed to that at all. The second matter is the relation between the "thought" however construed and the brain states which constitute the thought. Is it identity or is it mere one-many correlation. As I understand most expositions of this doctrine it amounts to a claim that beliefs (and perhaps also thoughts) have brain-correlates (whether or not identical with them) such that these "sentence-analogs have significant grammatical structure" (Harty H. Field 'Mental representation' in N. Block (ed.), Readings in Philosophy o f Psychology, Methuen, London, 1981, Vol. 2, p. 85). However it seems a physiologically rash hypothesis to suppose that the brain operates in this way, and one quite unnecessary in order to express the view that beliefs are brain-states - seeC. Peacocke, Sense and Content, Ch. 8, and D. Dennett, 'A cure for the common code' in his Brainstorms, Harvester, Sussex, 1979. Yet even if there are such brain correlates, there seems to me no reason to suppose that thoughts consist of mental items which are other than images of words of public languages or items (perhaps pictures) corresponding to observable features which are important in an agent's public behaviour. And against any hypothesis of an innate language ready for the expression of thoughts, before the agent has much experience of the public world is the fact on which I commented in the text - that the sort of thoughts and beliefs people have are very much conditioned by the language and practice of their society. Relatively seldom do they seem to have thoughts and beliefs not so expressible; and the beliefs and thoughts of which they are capable grow with their understanding of the language and participation in the practice of their societies. This would be surprising if they had a whole language of thought already existing independent of the public language. All this is not to deny Chomsky's claim that humans have an inbuitt capacity for learning public language - it is merely to doubt that they have much of a language apart from the public language (which in time they come to use privately). 16 D. Davidson, 'Thought and talk', in S. Guttenplan (ed),Mindand Language, Oxford, 1975, p. 10. 1~ Other recent writers, as well as Seaxle, who seek to distinguish between thoughts and their expression, and whose views echo in part the views expressed in this paper include: W.J. Ginnane, 'Thoughts', Mind, 1960, 69, pp. 372-390; P. T. Geach, Mental Acts, Routledge and Kegan Paul, London, 1957 (see especially, w 23); Zeno Vendler, Res Cogitans, Cornell University Press, Ithaca, NY, 1972, Chapter 3; and Colin McGinn, The Character o f Mind, Oxford, 1982, Chapter 4.

Oriel College, Oxford OX1 4EW, U.K.