Content and Consciousness

by Daniel C. Dennett

  7 The reader is invited to form other natural sentences affirming the existence of voices. He will find that we very rarely assert flat out the existence of a voice or voices. In this respect a voice has a much slimmer claim to reality than, say, a twinkle in the eye.

  8 See note 3 above.

  9 This position differs substantially from Ryle’s. He supposes that it will suffice to talk of different types or categories of existence (op. cit., pp. 22ff.). With regard to the view proposed above, it cannot be denied that there is also something counterintuitive in holding these odd sentences to be ill-formed, not logically false. For if there is something queer about ‘I can sit on an opportunity’, it seems that a very natural, ordinary, oft-heard way of alluding to this queerness would be to say: ‘but you can’t sit on an opportunity; it isn’t that sort of thing.’ This must also be ill-formed, however, in spite of intuitions. Such locutions can be kept as short cuts, however, for the more proper: ‘ “Opportunity” is not accepted into the context “can sit on …” ’, and I do this several times in this chapter.

  10 My use of ‘referential’ is only close to that of Quine in Word and Object, but I chose the term for these affinities. Another term, ‘syncategorematic’, is close to my ‘non-referential’, but was rejected since in its established use it generally describes adjectives, not nouns, and stresses class determination over existence; thus ‘expectant’ is syncategorematic in ‘expectant mother’ since the class of expectant mothers is not the subclass of mothers who are expectant (see Quine, op cit., p. 103). The distinction I wish to mark is not that between ‘horse’ and ‘centaur’ (there are horses, but no centaurs; in that sense ‘centaur’ does not refer), but that between ‘horse’ and ‘when’; ‘when’ is non-referential in my sense, which does not mean that whens are mythical or extinct; Quine would say ‘when’ was not a term.

  11 Karel Lambert has tried to persuade me that an adequate logical language for maintaining the ontological neutrality I require here would be a two-quantifier logic such as those developed by Leonard, Van Fraassen, and himself (H. S. Leonard, ‘Essences, Attributes and Predicates’, Proceedings and Addresses of the American Philosophical Association, Vol. XXXVII, 1964, pp. 25–51. B. Van Fraassen, ‘Meaning Relations among Predicates’, Nous, I, 1967, pp. 161–79. R. K. Meyer and K. Lambert, ‘Universally Free Logic and Standard Quantification Theory’, Journal of Symbolic Logic, XXXIII, No. 1, 1968, pp. 8–26. B. Van Fraassen and K. Lambert, ‘Quantifiers, Meaning Relations and Modality’ in K. Lambert, ed., Philosophical Development in Non-classical Logic; Modality, Existence and related areas). These languages were developed to deal with rather different problems, in particular the ‘possible objects’ so handy to modal logicians, and adapting my position to them would require allying voices, thoughts and minds to centaurs and gryphons (if not to round squares), and this would clearly be a distortion of the view I present. Whether in the end it would be a logically or philosophically undesirable distortion is not yet clear to me.

  12 In other words we start by treating the sentences in the fashion of the p and q of the propositional calculus; what then confronts us is whether their parts can also be brought under the quantifiers of the predicate calculus (viewed, in Quine’s fashion, as ontologically committing).

  13 H. Putnam, ‘Psychological Predicates’ in W. H. Capitan and D. D. Merrill, eds., Art, Mind and Religion, Pittsburgh, 1967, pp. 37–48. It might be argued that I have been unfair to the identity theorists, and that Smart and others have in fact made use of the ontological points I have raised. Smart, after all, in response to the objection that an after-image was not a brain process, replies ‘I am not arguing that the after-image is a brain process, but that the experience of having an after-image is a brain process’ (‘Sensations and Brain Processes’, revised version in Chappell, op. cit., p. 168). One might interpret this as holding that ‘after-image’ was not the referential atom, but rather ‘having-an-after-image’, a fused idiom. Nagel (op. cit., p. 341) goes even farther: ‘Instead of identifying thoughts, sensations, after-images, and so forth with brain processes, I propose to identify a person’s having the sensation with his body’s being in a physical state or undergoing a physical process.’ Nagel’s position is very close to mine in that he has simply taken a mental language sentence whole, turned it into the appropriate gerund nominalization, and put it next to an identity sign flanking a similarly altered physical entity sentence. But Putnam’s objection holds as well against Nagel’s more circumspect identities; the correlation Nagel supposes must hold is still too strong to be plausible. Moreover, hasn’t one lost the point of identity theory once one begins treating whole sentences as names in effect of situations or states of affairs which are then proclaimed identical with other situations or states of affairs?

  2

  INTENTIONALITY

  1 I shall capitalize Brentano’s term and its derivatives to distinguish them from the somewhat related and etymologically similar family of more common terms, ‘intend’, ‘intentions’, ‘intentionally’, which will be examined in detail in Chapter 9. It seems clearest to carry this practice into quotations as well.

  2 F. Brentano, Psychologie vom Empirischen Standpunkt, Leipzig, 1874, Vol. I, Book II, Chap. i, ‘The Distinction between Mental and Physical Phenomena’ – a selection in R. Chisholm, Realism and the Background of Phenomenology, Glencoe, 1960, trans. D. B. Terrell.

  3 Brentano called these objects ‘fictions’, indicating his own refusal to take seriously a metaphysical class of ‘inexisters’. For an excellent discussion of Intentional objects, see G. E. M. Anscombe, ‘The Intentionality of Sensation: A Grammatical Feature’, Sec. 1, in R. J. Butler, ed., Analytic Philosophy (Second Series), Oxford, 1965, pp. 158–68.

  4 R. Chisholm, Perceiving: a philosophical study, Ithaca, 1957, p. 170.

  5 Ibid., pp. 170–1.

  6 Chisholm himself has attempted in a number of papers to reformulate his criteria to meet objections. (See especially his ‘On some psychological concepts and the “logic” of Intentionality’, in H. Castañeda, ed., Intentionality, Minds and Perception, Detroit, 1967, pp. 11–57.) Since our aims diverge, however, there is no point here in recounting his various modifications.

  7 C. Taylor, in The Explanation of Behaviour, London, 1964, Part II, provides an excellent survey of behaviourists’ so far fruitless efforts to produce ‘operational’ definitions of such terms as ‘desire’.

  8 (6) meets (3) however, since if Smith, unbeknownst to John, is his uncle, the substitution will make the sentence false.

  9 Cf. S. Körner, Experience and Theory: an essay in the philosophy of science, London, 1966, p. 200.

  10 P. T. Geach, ‘Intentional Identity’, Journal of Phil. LXIV, 20, 1967, p. 629.

  11 Quine, op. cit. These and kindred examples are discussed in his Chs. 4–6.

  12 The hope of modal logic is that it will be possible to adjust the entities referred to in modal statements, or restrict the ways in which entities may be described, so that substitution – where it is permitted – still preserves truth. For example, one might claim that the only sense of ‘9’ that renders (10) true renders the identity ‘9 = the number of planets’ false; or in the same vein, that the sense of ‘the number of planets’ needed to make the identity true makes (11) true as well (a distinction being seen between (11) and ‘there are necessarily more than 7 planets’.

  13 The situation is not that simple, however. In the view of some modal logicians, any success with the alethic modalities will bring in its wake parallel successes in the other modalities, including the psychological or Intentional modalities. In this event the logic of Intentionality would be merely part of a unified modal logic, and instead of modal logic providing a solution to our counter-example, it would go on to provide something very like a refutation of the Intentionalist thesis of non-reducibility. This eventuality seems remote to me, as should become clear in what follows.

  14 See Quine, op. cit., Ch. 6.

 
15 Chisholm himself accepts the conclusion that (1)–(3) do not serve to distinguish the psychological from other modalities, and attempts to make the distinction in another way (Castañeda, op. cit., p. 11).

  16 Cf. ibid., p. 33.

  17 For an example of this sort of difficulty, see my ‘Geach on Intentional Identity’, Journal of Philosophy, LXV, II, 1968, pp. 335–41.

  18 R. Carnap, in Meaning and Necessity, Chicago, 1947, uses ‘intensional’ in a narrower sense to refer to a subclass of non-extensional sentences.

  19 Cf. Taylor, op. cit., p. 200.

  20 Quine, op. cit., p. 220.

  21 As a response to the request to utter a three-word sentence in English, saying ‘it is raining’ would be a case where (16) was true and (17) false.

  22 Quine, op. cit., pp. 220–4; Chisholm, Perceiving, Ch. 11.

  23 Cf. Körner’s distinction between dogmatic and methodological behaviourists (op. cit., Ch. XIII). It is the difficulty with Intentional language that turns methodological behaviourists into dogmatic behaviourists, but Körner claims this is not the only refuge.

  24 Quine, op. cit., p. 221.

  25 K. S. Lashley, Brain Mechanisms and Intelligence, Chicago, 1929.

  26 There is an incisive and systematic survey of the frustrations of the behaviourists in Part II of Taylor’s Explanation of Behaviour. Although, as he points out, his critique is open-ended, the troubles so far encountered exhibit ‘fairly reliable signs’ of futility. In particular there is the accumulation of inelegant ad hoc props and provisos that make stimulus-response theories so stupefyingly complex (p. 272). D. Shwayder’s Stratification of Behaviour analyses similar obstacles to stimulus-response theory, but Shwayder does not hold it to be in principle impossible for the behaviourists to produce a theory complex enough to deal with the problems he raises.

  27 For example, an animal in a problem-solving experiment will be ‘trained to criterion’, which means: given enough training trials to bring his performance up to an arbitrary standard of success, say 9 out of 10. But 9 out of 10 what? The ‘strictly behavioural’ criterion of learning is in actuality hedged with a ceteris paribus clause; what it means is: 9 times out of 10 the animal finds the goal box it was looking for, it achieves its goal.

  28 There could be, of course, ‘autonomous sciences of Intention’ without all the trappings of the Phenomenological movement, but with many centrally shared features.

  29 Arguments designed to show roughly this are found in Taylor, op. cit., A. I. Melden, Free Action, London, 1961 and Anscombe, Intention, Oxford, 1957.

  30 Cf. Taylor, op. cit., p. 44.

  31 Taylor says: ‘The notion of adaption is of course implicit in the ordinary language teleological form of explanation where action is frequently explained in terms of its propitiousness for certain purposes, i.e. by its “adaptiveness” in respect of these ends. The aim of S-R theories on the other hand is to explain behaviour without using a notion of this kind. Thus for Hull it is one of the tasks of a molar science of behaviour to explain why behaviour is adaptive, why “it is successful in the sense of reducing needs and facilitating survival”, a task separate from though closely related to that of explaining why the behaviour of different organisms is as it is. Adaptiveness is thus an explicandum for S-R theory. It is not a principle to be used in the explicans’ (op. cit., p. 115). Cf. I. Sheffler, The Anatomy of Inquiry, New York, 1963, p. 92.

  32 C. S. Pittendrigh, ‘Adaptation, Natural Selection and Behavior’ in A. Roe and G. G. Simpson, eds., Behavior and Evolution, New Haven, 1958, p. 395.

  33 Cf. Taylor, op. cit., pp. 107–8. Taylor points out (p. 271) that Freud was a centralist, hoping (rather faintly) to find a physiological basis for such Intentional phenomena as (subconsciously) intending to do something, fears, hates and repressed memories.

  3

  EVOLUTION IN THE BRAIN

  1 D. M. MacKay, ‘Towards an Information-Flow Model of Human Behaviour’, British Journal of Psychology, XLVII, 1956, pp. 30–43. For another description of storage see Taylor, op. cit., p. 108.

  2 J. Y. Lettvin, et al., ‘What the Frog’s Eye Tells the Frog’s Brain’, Proceedings of the Institute of Radio Engineers, 1959, pp. 1940–51.

  3 The example is D. Wooldridge’s in The Machinery of the Brain, New York, 1963.

  4 B. F. Skinner, The Behavior of Organisms, New York, 1938.

  5 M. Arbib, Brains, Machines, and Mathematics, New York, 1964, p. 56.

  6 Deprivation experiments are designed so to limit an animal’s sensory experience from birth that if the animal performs some perspicuous act, the ability to do this can be explained only as an inherited, instinctual capacity, rather than a learned one.

  7 See Taylor, op. cit., for a critique of these concepts, esp. pp. 170ff.

  8 This formula gives a teleological characterization of the conditions of existence of some neural structures. The question of whether teleological explanations are eliminable in favour of non-teleological explanations is the question of whether such a formula is subject to further explanation or must just stand as it is, a brute and unaccountable fact. My claim that the formula is explicable in terms of the operations of natural selection – a process that can be given a non-teleological description – is thus a claim in favour of elimination of the teleological. Cf. Taylor’s discussion of teleological explanation, op. cit., Part One. As Taylor also points out, the question of the elimination of the teleological is intimately bound up with the question of the reducibility of the Intentional.

  9 A. Newell and H. A. Simon, ‘GPS, a Program that Simulates Human Thought’ in H. Billing, ed., Lernende Automaten, Munich, 1961, reprinted in E. A. Feigenbaum and J. Feldman, Computers and Thought, New York, 1963, pp. 279–92.

  4

  THE ASCRIPTION OF CONTENT

  1 See, for example, D. H. Hubel and T. N. Wiesel, ‘Receptive Fields, Binocular Interaction and Functional Architecture in the Cat’s Visual Cortex’, Journal of Physiology, 1962, pp. 106–54; Lettvin et al., op. cit. and ‘Two Remarks on the Visual System of the Frog’ in W. A. Rosenblith, ed., Sensory Communication, New York, 1961, pp. 757–76; W. R. A. Muntz, ‘Vision in Frogs’, Scientific American, 210, 1964, pp. 110–19, and D. H. Hubel, ‘The Visual Cortex of the Brain’, Scientific American, 209, 1963, pp. 54–74.

  2 The analogy can be reversed in the case of information theory properly so called, of which parts may be considered relatively uninterpreted but amenable to a variety of different ‘meanings’ or applications or physical realizations. In this case, it should be noted, the information being considered (e.g., in terms of ‘bits’) is not intelligently used information. That is, one is concerned with the reliable transmission of impulses, dots, ons and offs, letters of the alphabet, codes, but not with the understanding of messages.

  3 Frank McGuinness has pointed out to me that a ‘neural negator’, for example, is a particularly unlikely bit of machinery. We cannot project grammatical transformations into the brain and hope to find transformers there. What similarities should we suppose to exist between the stimulus sources for ‘there is danger to the left’ and ‘there is no danger to the left’? Clearly, although we can see the different sorts of effects these signals should have, they need share no structural similarities, nor must they both be ‘derived’ from similar stimulus conditions.

  4 J. Zeman, ‘Information and the Brain’ in N. Wiener and J. P. Schadé, eds., Nerve, Brain and Memory Models, New York, 1963, p. 71.

  5 The ‘brain-writing’ view is plagued by a host of disanalogies in any case. As D. M. MacKay et al., point out, ‘The loose coupling between language and the world which distinguishes statements from symptoms is notably absent in something that has often been called a language – the representation of information in afferent nerve fibres. This is systematic but purely symptomatic, and the relations between “sender”, “user” and “referent” are considerably different. The retina cannot exercise an opinion whether to tell the brain what energy has reached it and where
…’, ‘Computers and Comprehension’, RAND Memo RM4065PR, Apr. 1964, p. 11. (This article contains many valuable observations on the relation between understanding sentences and the analogue of understanding in computers.) See also MacKay, ‘Linguistic and non-Linguistic “Understanding” of Linguistic Tokens’, RAND Memo RM3892, Mar. 1964.

  6 Cf. L. Wittgenstein: ‘ “And yet you again and again reach the conclusion that the sensation itself is a nothing.” Not at all. It is not a something, but not a nothing either! The conclusion was only that a nothing would serve just as well as a something about which nothing could be said.’ Philosophical Investigations, trans. G. E. M. Anscombe, Oxford, 1953, i. 304.

  5

  INTROSPECTIVE CERTAINTY

  1 Wittgenstein, op. cit., i. 244. See also i. 367, i. 370.

  2 G. Ryle, op. cit., p. 102. Ryle has since described to me a view of ‘degrees’, with ‘Ouch!’ at the avowal end of the spectrum and ‘the pain is in the third tooth, upper left’ at the other, reportorial end. This is plausible, but of course demands either an explanation of how the true reports at one end are infallible or the implausible view that only the avowals are immune to error.