Invited Introduction: Finding Psychology Simon Blackburn The Philosophical Quarterly, Vol. 36, No. 143, Special Issue: Mind, Causation and Action. (Apr., 1986), pp. 111-122. Stable URL: http://links.jstor.org/sici?sici=0031-8094%28198604%2936%3A143%3C111%3AIIFP%3E2.0.CO%3B2-0 The Philosophical Quarterly is currently published by The Philosophical Quarterly.
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/about/terms.html. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/journals/philquar.html. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.
JSTOR is an independent not-for-profit organization dedicated to and preserving a digital archive of scholarly journals. For more information regarding JSTOR, please contact
[email protected].
http://www.jstor.org Sun Jun 17 04:43:15 2007
The Philosophical Quarterly Vol. 36 No. 143 ISSN 0031-8094 $2.00
INVITED INTRODUCTION: FINDING PSYCHOLOGY
The philosophy of mind is largely a chase after fugitive facts; facts about meaning, thought, agency, consciousness, whose nature escapes our understanding. They escape it for two reasons. Firstly, what we would allow as success in understanding them is constrained by various principles, and the constraints are very tight. Misinterpreted even a little, they become impossible to satisfy. They include:
(1) That we should not deny firstlthird person asymmetries - yet we should not make knowledge of other minds impossible.
(2) That we should not abandon a physicalist ontology - yet we should not make intentionality (agency, etc.) impossible.
(3) That we should not ignore "the holism of the mental" - yet we should not deny that beliefs (etc.) have their own content and identities. (4) That we should respect the apparent identity of belief over identically appearing ("doppelganger") possible worlds - yet we should make proper place for genuine reference and intentionality in describing psychologies. Each member of these four pairs stands in some tension with the other; perhaps more philosophers are keen to stress allegiance to one than have succeeded in exorcizing the pull to the other. These tensions are familiar. But the second main obstacle, I believe, to progress is more abstract. It is that we come to these issues with a set of doubtful categories. These include certain distinctions (causal vs. noncausal explanations; realist vs. instrumentalist construal of theories) and technical terms (notably, I suggest, 'truth-condition', 'representation', and most treacherously, 'state'). In this brief introductory article I shall say little about the particular constraints on a proper psychology, but more about the importance of understanding the categories we bring to it. If I have a theme, it is this: in today's climate, philosophical psychology has ousted philosophy of language as the ~ u e e nof the philosophical sciences - yet what it needs is a better philosophy of truth and of fact. My theme is intended to impinge on assumptions which seem to be present in many writers in the "cognitive
112
SIMON BLACKBURN
psychology" tradition, regardless of differences amongst them. Indeed, those differences (for example between representative and syntactic theories of mind, or instrumental and realist theories of intentional ascription) seem to me a response to pressures which can actually be released by better understanding of truth and fact.
In recent years the best hope for forging an understanding of mind has been functionalism. This starts with the observation that we use psychological attributions to explain and predict the things each other does. So common-sense psychology is a "folk-theory", or attempt at explanation. If not itself a theory, at least it is a kind of proto-theory, or framework for fuller theory. It specifies links internally amongst psychological "states", and between them and inputs ("stimuli") and outputs ("behaviour"). Furthermore (goes the dominant story) these explanations are causal: it must be because I believe that the cat is under the bed that I am inclined to say that it is there, or feel uneasy about its being there. A complete psychological model of a subject would provide the entire dynamics of this system; the suggestion is not that folk psychology ever enables us to do this for an actual subject, but that it contains enough principles and concepts for us to be able to frame the outline of such a theory. Its rough generalizations can be related both to an evolutionary understanding of ourselves, and to norms such as those of decision theory or logic. For instance, common sense might suppose that the belief that there is a cat in front of me is typically caused (using normal vision and hearing in normal circumstances) when and only when there is indeed a cat in front of me; or that a tendency to believe that p normally follows upon a tendency to believe that p & q: a teleological/evolutionary story is in the wings waiting to say why such things should be true. Creatures set to act on cats under the bed when there are none, or not set so to act when there is one, do worse. But there will also be occasions where (for natural reasons) our systems are set to form inappropriate beliefs and desires, and these too can be identified. T h e liberating effect of this conception of psychology is well known: it gives psychological terms the freedom from strict particular behavioural connections that logical behaviourism wrongly required, and since it can be quite silent about the precise "realization" of the causal pathways, it can be charitable to machines and Martians, giving psychology the right degree of autonomy from neurophysiology. Or, it may seem to do so. But now, it may be suggested, on this causal version of functionalism, we only believe, desire, etc. if there exist states with very particular places in our causal organization. For instance, as Stich
'
INVITED INTRODUCTION: FINDING PSYCHOLOGY
113
points out, in common sense theory my believing &at the cat is under the bed explains both why I say that it is, and why I reach down to pull it out.' So it will be a promise of causal functionalism that there exists one state causally responsible for both the saying and the doing. But then (the argument continues) physicalism ensures that the only kind of state able to explain both those things would be at a neurological node somewhere in the head, and there may be no such node. It might turn out, for example, (Stich and Churchland cite empirical work pointing tentatively in this direction) that nothwithstanding what it feels like from the inside, behaviour control and verbal control could be under the command of quite different systems, with yet another part of the brain ensuring some kind of co-ordinati~n.~ The physical investigation of the brain would reveal no such thing as "the state causally responsible for both the saying and the doing", and in that case the folk-theory implies a falsehood. I think this is a good specimen of the kind of argument that leads theorists to see folk psychology as "scientifically doubtful". Does it work? Perhaps it seems plain that if folk psychology says that there is something that causes both A and B, yet science shows that A has a different cause from B, then either folk psychology is wrong, or it needs to abandon any pretence of descriptive accuracy. Yet consider this analogy. Folk meteorology tells us that crops fail to grow, and people feel miserable, because of the bad weather. And this 'because' is causal. So, it is a promise of folk meteorology that there is one state causing each of these effects. But good empirical work could well show that in fact one state (persistent cold, say) causes the crops not to grow, whereas another state (persistent wet, say) causes people to be miserable. The best full physical story of the atmosphere and its effects may acknowledge no one state with both consequences. Similarly perhaps folk geology talks of earthquakes causing buildings to fall. Perhaps physical geologists separate different things (transverse from vertical wave movements in rock, say) each with different effects (shaking and rolling, say). Do these discoveries undermine folk meteorology or geology? Didn't folk-theory say that one thing or state caused the shaking and the rolling? Here is the natural solution to this strained problem. So long as the two wave movements, or the wet and the cold, or the tendency to say and the tendency to do, typically occur together, the folk-terms have a role. They tell of a syndrome, and their utility depends on the general co-ordination of its different aspects. The concepts would lose their propriety (and utility) if, typically, the unity of the phenomena broke up. (Would harmless shaking without rolling be an earthquake?) But the realization that the phenomenon is complex does not break up this unity. If earthquakes (bad weather, beliefs)
' Stephen Stich, From Folk Psychology to Cognitive Science (London, 1983), p. 231. Stich, op. cit., p. 231 ff. Paul M. Churchland, Matter and Consciousness (London, 1984), p.79.
114
SIMON BLACKBURN
typically involve different aspects which science can separate, we learn that the folk-story can be filled out further, but in no case are we apt to abandon it. A chair causes me to maintain my present posture regardless of whether some of its properties are irrelevant to that power.' There are pressures which cause thinkers to resist this simple reconciliation. It is as if we think: well, if the vertical movements brought down the buildings and the transverse movements are harmless then it wasn't really the earthquake (not that overall state, or event) that caused the damage. The villain of this piece is the notion of a state, and a change of state. We too easily slide into a billiard-ball model of the notion: a proper state should be physically identifiable and located, causing things by the exertion of scientifically respectable forces. How else are we to accommodate the physicalist pressure? Agglomerated states like bad weather, earthquakes, the presence of a chair or of a belief, become simple obstacles to clear identification of the causal lines. If it seems unlikely that so innocent-sounding a term as 'state' should play us false, consider how the flight into the head started. Belief states explain both acting and saying, so they must "consist in" states far back along the causal chains science finds. This is what leads to difficulty if a causal isomorphism with neurological states fails. But is common sense embroiled in any such theory? Certainly it allows that I am often in a psychological state -but it does not insist or even allow that psychological states are often in me. In folk-theory, being in a state is like being in a tizzy or a bother. I believe that the cat is under the bed; very well, I am in a state of believing that the cat is under the bed. But is this state a locatable part of me, causally influential through emitting gravitons and electrons? Is it amenable to tweezers and microscopes? Obviously not. It is right to say that I am in such a state, because I am set to control myself appropriately in response to information as from the cat. States involved in that control, either in the acceptance of the information or in the subsequent modifications of my actions, gain their title - their right to be thought of as "representing" the presence of the cat, in virtue of their role in this agency. It is the whole person which has a psychology, not internal bits of him. The trick is to realize that although the neurophysiological truth shows us how internal states cause and explain behaviour, the direction which illuminates the philosophical and psychological problems of intentionality - and which is fully consistent with functionalism - is the reverse. It is only in virtue of its being a part of the signals in neural pathways in the whole active person that an electrical or chemical modification can be said to represent ? The concept 'chair' need not be any part of the final scientific Book for it to be true that the presence of chairs explains things. Chairs are physical, even if to physics they form a class with which it has no truck.
'
INVITED INTRODUCTION: FINDING PSYCHOLOGY
115
the impending tiger or the death of Hastings. The grounds for saying that the signals encode information about the tiger or the death of Hastings is that they play some role in the dynamic organization of the whole person. One could not detect what is encoded from any investigation of the internal, intrinsic properties of the configurations of any particular part of the animal.'
Consider another philosophical argument for internal "representations". The propositional attitude idioms are on the face of it relational, so that ifwe suppose that they identify literal truths we ought to find a real object to which the subject is being re1ated.j If a subject is related to something when he believes that p, then since relating us to abstract propositions offends against metaphysical constraints, the best bet becomes an inner "representation", a "sentence in the head" - the internal representative thing which, somehow, stands for the cat being under the bed. This would be something there in the brains, with representative properties. It is true that its capacity for these properties is a little puzzling ("Of the semanticity of mental representations we have, as things now stand, no adequate account" Fodor6), but theorists bent on rescuing the propositional attitudes have struggled to assure themselves that this "semanticity" is going to be there. The scene is set for people to recognize that the properties of things in the head (cells and pathways) are the first instance connected with other things in the head, and only mediately and accidentally with parts of the outside world - there is no real (spatially) inner state which is essentially connected with the cat, in that had the cat not been there, or had there been no cathood in the world, it could not have existed. But a belief-state involving a cat must, essentially, involve it. But since spatially configured states are all that there are (physicalism), there is no such thing as belief. My alternative is simple: if belief-states both "have semanticity" and cause behaviour, then they must be thought of as states of an animal in an environment - something which includes as well the feedback from action. Like Evans, I think we can have the scientific and psychological insights of the computational theory of mind, and also keep a firm grip on the need to explain intentionality from outside in - to form a naturalistic psychology which cites the typical causes and functions of the agent.7 I think that it is literally true that we believe and desire, and represent the world to ourselves: W e r e 1 have been especially helped by conversations with Derek Bolton. Stich, op. cit., pp. 24, 78. J. A. Fodor, "Fodor's Guide to Mental Representation", Mirut' 94 (1985), p. 88. Fodor, op. cit., p. 99. Gareth Evans, "Comment on Jerry A. Fodor's '.Methodological Solipsism Considered as a Research Strategy in Cognitive Psychology"', Collected Papers (Oxford, 1985), p. 400.
116
SIMON BLACKBURN
the relational expressions are entirely proper, yet they imply no "representations", no sentences in the head or inner bits advantageously thought of like sentences on a cathode-ray tube, whose presence somehow carries the semantic burden. It is no offence against physicalism to suppose that when we relate people to propositions, the only relatum is abstract (indeed, the offence would be if that were not so: physicalism only fails if there are real causally acting things which escape the physicist's ontology). The proposition is literally an abstraction from the psychological similarities of all those whose mental lives are organized with that proposition as a focus - doubters, believers, hopers alike. The content of a belief is not an item in space and time to which a subject is related. It is an abstraction from facts we want to know about the organization and awareness of the subject. This is what functionalism needs to say, and can say. Propositional description of people is precisely that which invents an object to stand at that point. Telling how someone is related to a particular proposition is telling something literal about him, but the proposition needs no spatial and temporal identity to function as it does in this story. Of course, in saying this I cheerfully stand opposed to the dreary tradition of finding sentences, utterances, representations - anything, so long as it be not abstract - to stand as the object of the demonstratives introducing 'that'-clauses. But, comes the reply, this avoids the ontological problem only at the cost of abandoning "realism" about the intentional. Propositions become "abstracta" - "calculation-bound entities or logical constructs", like lines in a parallelogram of forces, or like directions on a Fregean constru~tion.~ And how can such things play any role in genuine causal explanation? The answer of course is that by themselves they don't: propositions do not bump into people or emanate light rays. What does play a role in explanation is possession of a propositional attitude - believing, hoping, etc. that p. Similarly "velocities" play no role in causal explanation - only the fact that particular things are travelling at particular velocities. Possession of a propositional attitude plays a role in causal explanation, because a subject can be so typed only if he is in a state - and his being in that state can be cited in explanation. But, to repeat, his being in such a state makes no reference to any state which is in him. Perhaps another opposition to this may stem from allegiance to Quine's ontological attitudes. If propositions are values of variables in our best theories of people - if people believe, doubt, assert and deny these things, then we who hold these truths about them are "ontologically committed" to the propositions which we mention. Physicalism then forces us to bring them down to earth. But this is far too crude a picture of reference. It is just not the case that a belief has to be a physically identifiable thing for it Stich, op. cit. p. 243.
.
INVITED INTRODUCTION: FINDING PSYCHOLOGY
117
to be true that people have beliefs and act because of them - even in a purely physical universe. This is not so even if, having related people to propositions, we also find it convenient to quantify over the ubstructa: say that there is one thing which two people each believe, and so on. All that has to be so is that the real, gross, physically identifiable things - the people - get into dispositional or other states, which play a part in explaining their actions, and which we recognize as such. Science may tell us that when such states change, so do the internal patterns of electrical and chemical impulses, but folk psychology is gloriously silent about that: there is simply no "substantive commitment about underlying internal mechanism^".^
We escape inner representations by reminding ourselves of the external, natural, relationship between the subject and how he is best to behave in his world. Now there is a dispute which, as far as I can see, cross-cuts this, but which tends to get involved with it, and I hope readers will forgive me if I digress to say why it should not. This is the issue of whether there is a distinct category of thoughts which are essentially individuated by reference: the "singular thoughts" or "Russellian thoughts" of current debate.1° In my terminology, "universalists" are sufficiently impressed by "doppelganger" cases to suggest that no thoughts should be essentially classified by reference (ways of reporting thinkers however, using names, and indexicals, are of course bound to their actual referents)." My universalists defend themselves by pointing to functional equivalences between persons and their doppelgangers; Gareth Evans and John McDowell think that this idea is only tempting if we already sympathise too much with the flight inside the head, and this in turn will lead either to the "darkness within" following the collapse of representational theories, or to distressing Cartesiani~m.'~ The hardest case is that of the perceptual demonstrative. It seems useful to distinguish at least three different aspects involved in an ordinary case of using such a demonstrative. There is the identity of the object; the identity of the features whereby it is perceived; and there is the initial way it is taken by the subject, or "what he makes (perceptually)" of the scene. It is possible to rotate these against each other: the same object may appear differently, or the appearance may be taken differently although it is the same, and Stich denies this, p. 244-5. Evans, The Varieties ofReference (Oxford, 1982), ch. 4. The universalist does not deny that singular terms used in 'that'-clauses have a distinct semantic role - he only queries whether this role is to introduce a szri generis singular thought. See my Spreading the Word, (Oxford, 1984), ch. 9.2. l 2 John McDowell, "Singular Thought and The Extent of Inner Space", forthcoming in Subject, Thought, and Context, ed. John McDowell & Philip Penit (Oxford, 1986). 'O
118
SIMON BLACKBURN
different objects can display identical features. We thus get thought-experiments of the kind I call "spinning the possible worlds", providing substitute and empty cases, and no doubt others, and we can ask what stays constant and what changes as we do so. It is clear that paying attention to these thought experiments in no way prevents us from accommodating the cognitive differences between, say, concentrating upon a perceptually presented object, and merely describing one. The mental state of a subject who takes himself to be in the presence of an object and who is, for instance, paying attention just to it, has to be distinguished sharply from that of someone thinking generally, e.g. by using definite descriptions ila Russell. It was, I think, absolutely right of Evans to hammer this point home in chapter 6 of his book. Nothing seems to follow about the identity of these mental states, and in particular there is no reason yet to associate their identity with that of the object, forcing just one answer to the question: if it had not been that object, could the subject yet have been thinking the same, or would it follow that he would be thinking differently? It was too quick of Evans to associate the perfectly correct points about the quite different nature of the thinking involved in perceptual demonstrative use, with any doctrine of the individuation of what is going on psychologically - of thoughts. It may be that we ought not to pursue the question of individuation - that it already concedes too much to the notion of a "mental state" attacked above, so that in thinking of such states as having definite, if controversial, identity conditions we are already in error. If we do pursue the question of individuation, it must depend upon what we want to individuate thoughts for -which aspects of sameness and difference across mental lives are to matter? One common aim is to satisfy the Fregean principle that it must not be possible without irrationality to hold rationally conflicting attitudes to one and the same thought. But this is hard to satisfy, and doubly hard if the identity of the object thought about is brought in. It is hard to satisfy if'mode of presentation' is independently intelligible (it can of course be introduced simply as "whatever it is which enables thoughts to satisfy Frege's constraint"). In any ordinary sense we can have the same object presented in the same way to someone occupying the same perspective, and thought of in the perceptual demonstrative way - that object - whilst contradictory ascriptions of predicates are rationally made. Someone might be in a position in which he rationally supposes that it is a different object, although in fact it is not: object and mode of presentation are the same, but contrary predicates get applied (e.g. I am told by a conjuror that I will be presented with a succession of boxes, one of which will have a ES note in it. Whilst I look at the one visible box, I rationally come to believe that a different one has been substituted as I blink, and think that although that one did not have a note in
INVITED INTRODUCTION: FINDING PSYCHOLOGY
119
it, this one does. One might try saying that this is a different perspective, because of the passage of time, but then the difficulty will be to connect senses cut this fine with the theory of communication. There is no independent motivation, that I am aware of, for denying that the phrase 'that box' introduces the same thought (certainly: the same modification of the subject's mental life) on each occasion. In this case, it is true, the passage of time is connected with something which can be called a change in the subject's perspective, or the mode of presentation, because he remembers the earlier (stages of the) box, and perhaps his perspective on the box changes as time goes by. But the same case can arise without that. A man may be on a production line: thinking "this glass is perfect" . . . then (remembering the criteria better) . . . "this glass is flawed"; but unknownst to him the supervisor has sent round the same glass again. Here we have the same glass, same public appearance, same way of taking it, and no perspectival change brought about because of a memory of the first judgement - but rationally different propositional attitudes. Frege's principle needs a colder look than it normally gets. But in any case, once the theoretical need is identified, the individuation of thoughts will follow on. It is hard to make more of this issue than that, and this does nothing to connect the recognition of singular thoughts with the attainment of a genuinely "anti-Cartesian" theory of mind. In a fully Cartesian picture we have: (i) the autonomy of the mental realm; (ii) the transparency of the realm to the introspection of the subject (this means: infallibility); (iii) the correspondingly problematic access of subjectivity to the rest of the world. Is a universalist either committed to these three doctrines, or covertly motivated by them? T h e issue is whether there is a theoretically important truth, or way of characterizing the mental life of the subject, which has the thought the same in either substitute or empty cases. It is admitted on all sides that something is the same: how things seem to the subject. Even McDowell admits that "by itself there is nothing dangerous about the idea that how things seem to one is a fact, knowable in a way that is immune to the sources of error attending one's capacity to find out about the world about one". One might have expected, then, that by itself there is nothing dangerous about the idea that how one is thinking is a fact knowable in ways which are immune to the sources of error attending one's capacity to find out about the identie of objects in the world about one. In other words, the issue about thinking could be settled by association with the admittedly harmless concept of seeming (in other words not taken Cartesianly). The deluded subject,
120
SIMON BLACKBURN
thinking "It's coming to get me" when in the grip of a hallucination, would indeed be thinking in the characteristic, attention-focussing way, and so describing him would be like describing how things seem to him - and this is admitted to be legitimate. McDowell seeks to make the legitimate notion of seeming the same somehow disjunctive (so that entirely different facts make things seem the same on different occasions). But even if this move had anything to commend it, which I rather doubt, it could be made about thoughts as much as about experience. In an hallucination case the functioning, as normally specified (propositionally), needs a thought slot, whereas on the Evans and McDowell conception, no singular thought is entertained. And it may need the same slot where on that conception, a different thought is being entertained (the unnoticed substitution of an identically appearing object, for instance). Still, there is a difference between substitute possibilities and empty ones: the doppelganger faced with a substitute but identically appearing object is possibly functionally identical with the original subject; one who hallucinates so that things seem to him exactly as they do to the original subject, is surely not. One aspect of his functioning may be the same, but another is different - the one causing the hallucination. Once again, this reminds us that questions of identity of thought need a theoretical background to come alive. I do not doubt that theorists such as McDowell would accept this: it is not a conflict of mere intuition. But I do doubt whether the theoretical need which forces us to give prominent place to the singular thought has yet been identified. Evans and McDowell share the naturalistic aim of placing psychology in the life of the agent in the world. But if we can have a position which is as genuinely opposed to methodological solipsism, as genuinely naturalistic about psychology, but does not incur the counter-intuitive costs of singular thoughts, it is obviously important to separate it from its vulnerable associates. Is the debate important to epistemology? McDowell has again claimed, doubtfully: "allowing intrinsic object-dependence, we have to set whatever literally spatial boundaries are in question outside the subject's skin or skull. Cognitive space incorporates the relevant portions of the 'external' world. So its relations to that world should not pose philosophical difficulties in the Cartesian Style". It is essential, on my view to deny that any literal spatial (or temporal) boundaries are in question.I3 The right response to those setting the boundaries of the mind inside the head is not to set them instead some civilized disrance outside it. And as far as epistemology goes, the only result ' j Literal spatial boundaries would be in question only if the concept of a state attacked above were in order. Considered as a metaphor, however, the spatial vocabulary needs very careful interpretation - much better go for the literal truth, expressed in terms of possibilities and chances.
INVITED INTRODUCTION: FINDING PSYCHOLOGY
121
of identifying thought by a criterion which includes the identity of external objects, is to make one's own cognitive space an object of possible doubt: if substitutes and empty possibilities were more common, one would literally not know what one was thinking.
Let us return to the suggestion that we identify belief, desire, thought, in the life the subject lives in his world. I have tried to diagnose some reasons why this kind of naturalism is so easy to overlook. The final, perhaps the basic reason lies in the theory of truth. If the ontology is one of physical atoms in their places, exerting their forces, then it must be some distribution of these that gives the "truth condition" for someone's believing that p. It takes a spatial configuration to represent spatial configurations. This is the Tractarian conception of the relation of language and the world, and notwithstanding Wittgenstein's recantation, it still dominates our thinking. One can imagine the thoughts I have sketched about the way to relate psycholbgy and the science of configurations in the neural pathways as "later Wittgensteinian", and hence unscientific; or one can imagine them being taken to licence an instrumentalist view of psychological description, in the manner of Dennett.I4But I deny that at this point we have a good distinction between "taking up an intentional stance" towards a person, and "describing a set of truths about him". Sometimes, surely, there is such a distinction which does real work (in the theory of morals, and modals, for instance). But folk psychology is genuinely explanatory, and where we accept explanations, it seems to me, we are not at liberty to turn around and say that we do so but only in an "anti-realist" spirit. Seeing the psychological as essentially large-scale, dealing with facts about the organization of the whole subject, is not, of course, denying any of the scientific interest of discovering how information is encoded, channelled, and dealt with in the head. Nor is it taking an a priori stance on the likely success of computational research programmes in empirical psychology (although it is, perhaps, offering a way to wean ourselves away from the model of the digital processing of something thought of as formulae). Certainly there is no a prion' reason to expect this level of science to be conducted in linguistic categories. Electrical impulses are not sentences even if we can regard them as representative in the same way that we can regard impulses in a telephone-line as representative of a conversation. Believing,,as I do, that semantics rules syntax (the syntactical organization of a sentence or indeed the identity of anything as a sentence, is a reflection of
-
l4
D. Dennett, "Intentional Systems" in Brainstorms (London, 1978).
.
122
SIMON BLACKBURN
the kinds of semantic power of its various components) I think it is equally artificial to talk of the encodings possessing a "syntax". Traps still wait. One development of a "macro" approach to psychology is to think exclusively in terms of our attribution of psychologies to others. The interpreter and his principles (charity, humanity, etc.) is certainly an alternative to the telephone engineer. Concentrating on the interpreter tells us a good deal about the epistemology of psychological attribution. But it must be doubted whether it brings similar success in thinking about its metaphysics or ontology. For once more it proves difficult to keep hold of the idea that the interpreter is discmering genuine truths, and not just imposing a scheme of description. And in any case, from the standpoint of the metaphysics of the mental, there is always the background residue - the unspoken reliance on the interpreter's own powers (the "frog in the mug" problem). T h e closer one studies those writings which concentrate upon interpretation, the more the matter turns out to hinge on preconceptions about the powers the interpreter brings to his task. Unless we have a regress, a "disappearance" theory of mind, in which the only way to think of those powers is to postulate another interpreter, there must be a point of escape, in which we consider just the thinker, his organization, and his world. So why start on the regress at all?
Pem broke College, Oxford