The Mind/Brain Identity Theory:
A Critical Appraisal

3. The Problem of Phenomenal Qualities

I have so far argued that sensations do not have phenomenal properties. Nonetheless, there is something about sensations that provides a major stumbling block to the acceptance of a materialist identity theory of mind. This stumbling block is illustrated in the now well-known 'inverted spectrum' argument. Let us briefly consider J. J. C. Smart's and D. M. Armstrong's versions of the materialist identity theory. Smart's account of colour perception goes something like this. He first defines a 'normal percipient' as, roughly, a member of the class of people that can make the most colour discriminations. Smart [1977: 59f] then claims:

that 'This is red' means something roughly like 'A normal percipient would not easily pick this out of a clump of geranium petals though he would pick it out of a clump of lettuce leaves'. Of course, it does not exactly mean this: a person might know the meaning of 'red' without knowing anything about geraniums, or even about normal percipients. But the point is that a person can be trained to say 'This is red' of objects which would not easily be picked out of geranium petals by a normal percipient.

Let us ignore the obvious point that it is difficult to see how an ordinary person can mean by 'This is red' something about the discriminatory powers of normal percipients when it is not necessary for the person to know anything about normal percipients. This is so irrespective of what Smart says about what the person is 'trained' to say in particular stimulus situations.

Book cover: The Emperor's New Mind by Roger Penrose

Here is a version of the 'inverted spectrum' argument in which a scenario is drawn such that the definiendum, in this case, 'This is red', is rendered false while the purported definiens is rendered simultaneously true. The purpose of this example is to demonstrate that the two sentences cannot be synonymous. The scenario is this. Everybody wakes up tomorrow morning to find that their colour perception has undergone a reversal. The objects that were previously yellow are now seen as blue, and vice versa; those previously red are now seen as green, and vice versa, and so on. Furthermore, upon investigation, no change is found in our neurophysiology or in the optical properties of objects. Nor has our ability to discriminate between colours diminished.

In this new situation, what we had previously called a red rose petal, we would now say of it that it is green. It would remain true, though, that we would not be able to pick it easily out of a clump of geranium petals (because they are now green also). So, in this situation, the sentence, 'A normal percipient would not easily pick this [rose petal] out of a clump of geranium petals though he would pick it out of a clump of lettuce leaves', would be true, but we would deny emphatically that 'This [rose petal] is red'. (A person who held Smart's semantic theory would be in the absurd position of denying that the rose had changed colour.) The two sentences, therefore, could not even roughly be equivalent in meaning.

The solution, of course, is not to choose other paradigm cases of red and green objects, nor to include in one's definition a lengthy list of such objects, for the problem lies in Smart's tying of the meaning of colour words to an operationally defined 'normal percipient'. However, as the above counterexample demonstrates, an operational definition of colour terms misses the qualitative character of colours. This same 'inverted spectrum' objection applies equally to Armstrong's [1968: 248] account of colour perception as 'the coming to be of a state of the person apt for the bringing about of certain sorts of discriminative behaviour'.[10]

We can now see why Smart's topic-neutral account of avowals is so unsatisfactory. His view was that:

when a person says, 'I see a yellowish-orange after-image', he is saying something like this: 'There is something going on which is like what is going on when I have my eyes open, am awake, and there is an orange illuminated in good light in front of me, that is, when I really see an orange.'

[Smart 1977: 60]

The problem with this rendering is that it is logically possible that oranges will change their colour overnight, with no corresponding structural change to anything else in the universe. And the same could possibly occur, logically, to any or all other yellowish-orange coloured things. This argument shows that oranges are only contingently paradigmatic instances of yellowish-orange. There is no logical connection here, for if our perceptual experience of oranges were to change radically tomorrow, the last thing that we would do is to continue to call them 'yellowish-orange'.[11] Armstrong's [1968: 254] account of 'sensation' as 'the stimulation of the mind by the world in complete abstraction from the further effects of this stimulation on behaviour or impulses towards behaviour' is an extrapolation from his defective theory of perception, so I shall not consider it further.[12]

Smart and Armstrong, in their versions of the materialist identity theory, sought to define sensations in terms of (but not logically reducible to) physical stimuli and bodily responses. With these definitions, they had the specific intention of leaving a logical gap concerning the nature of the effects of stimuli and causes of behaviour that could be filled definitively at a later date by statements about brain states and processes. I have argued that the 'inverted spectrum' case shows that any such attempt is doomed to failure. This failure arises not because mental states are not identical with brain states, but because any such account ignores the phenomenal qualities of many mental states. The meaning of many 'mental' terms, especially 'sensations', such as 'sensation of redness' and 'sensation of burning pain', can only be indicated ostensively. This traditional view is only all too familiar.[13] It is this brushing aside of the phenomenal character of mental states in Smart's and Armstrong's formulations of their 'logical gap' semantic theories that make me think that these materialistic philosophers have not taken the mind-body problem seriously.[14]

An incisive objection to the 'ostensive definition' view of mental terms is Wittgenstein's criticism of the idea of a private mental language. How then do I handle these two specific Wittgensteinian [1953: part 1] objections:

  1. if some mental terms refer to something essentially private, then it is not possible to learn their meaning from public demonstrations, for all that we can observe from such demonstrations is outward behaviour, and
  2. if the referents of some mental terms are not open to public examination, then we cannot know whether or not other people are ever in these mental states.
Book cover: The Self and Its Brain: An Argument for Interactionism by Karl Popper and John C. Eccles

Here, I will not attempt a comprehensive response. I will content myself with the less ambitious task of indicating how such a reply would look. The most promising place within which to bed a convincing reply is within an expanded Lakatosian evolutionary epistemology. I think his framework throws light on the nature of language acquisition. From this perspective, we see that the child learns 'mental' terms through a process of unconsciously generating hypotheses about the meaning of a word and testing such hypotheses in novel circumstances. However, as with our scientific theories, the child's word-meaning hypotheses do not meet the test of experience in isolation. They are always tested in conjunction with certain background assumptions (the auxiliary hypotheses).

In the case of learning 'mental' terms, a critical background assumption for the child is that his private experiences form certain causal connections with publicly observable events. For example, that the child's private experience of a sensation of redness is causally correlated with the publicly observable rose petal. It is this background assumption, when conjoined with a particular hypothesis about the meaning of a new word (such as 'red'), that proves the most fruitful for the child in predicting the outcomes of new situations. One such successful anticipation of new events could be the approval that the child receives the first time he describes a postbox as 'red'.

An inextricable part of the background assumptions also being tested by the child as they acquire their 'mental' language is the theory that other people have similar private experiences to their own. He posits that these experiences also bear similar causal connections with publicly observable events as his own. The child's learning that he has a private mental life happens concurrently with his learning that others similarly have private mental lives. In this way, his learning of a 'mental' language is inextricably linked with learning about the world and other minds—a world of interactions between private experiences and public events.


  1. [10] Even though Smart held a Lockean 'powers' view of colours, while Armstrong adhered to a direct realist view, we cannot help but note the essential similarity of their views on colour perception in that both define colours in terms of discriminative behaviour. Armstrong tried to meet the colour reversal objection, albeit in a different form, in his [1968: 256–60]. I think that his reply is unsatisfactory, for he seems to have misconstrued the nature of the objection. For that part of the objection that Armstrong [1968: 256] correctly stated as, 'it seems logically possible that there might be no evidence to show that my perceptions were reversed in this way', he appears to have misinterpreted it as meaning 'no behavioural evidence'. Armstrong then, expectedly, argued that the differences in colour perception could be revealed by the discovery of a systematic difference in the nature of the causes that enable the same behaviour in the two groups. But this is beside the point, for the objection is that it seems logically possible that there might be no evidence simpliciter for a reversal of colour perceptions.
  2. [11] There are other, what I consider to be, decisive objections to Smart's topic-neutral analysis that are not directly relevant to the point that I am making here. See, for example, Cornman [1977: 126] and Shaffer [1977: 135f].
  3. [12] For an effective criticism of Armstrong's theory of perception and sensation, see Ellis [1967: 148–57].
  4. [13] The case of the ascription of colours to physical objects is a little more complicated, for here, what is being ascribed is a dispositional relational property of the object. So, to say, for example, that 'The post box is red' is to say that a standard observer perceiving the post box under standard viewing conditions will have a sensation of redness (where what constitutes a 'standard observer' and 'standard viewing conditions' is spelled out).
  5. [14] Smart and Armstrong have since gone over to the heir of the identity theory, functionalism. But even here, the 'inverted spectrum' argument tells equally against this new theory.

Copyright © 2016

You will be interested in

Share This

  • twitter
  • facebook
  • linkedin
  • googleplus
  • gmail
  • delicious
  • reddit
  • digg
  • newsvine
  • posterous
  • friendfeed
  • googlebookmarks
  • yahoobookmarks
  • yahoobuzz
  • orkut
  • stumbleupon
  • diigo
  • mixx
  • technorati
  • netvibes
  • myspace
  • slashdot
  • blogger
  • tumblr
  • email
Short URL:


PDF Download The Mind/Brain Identity Theory: A Critical Appraisal

Download this essay