It is true, of course, that when we see we do not simply see that there is a table in front of us, but a table of a particular colour and shape in a particular position and so forth. All this need mean is that the information we receive is vivid and rich in detail. This is not true of the vision of many lower animals. The frog, for example, can see that there is a small moving object before him, but he cannot see that it is a fly or a bit of paper on a string. If the small object is not moving, he cannot see it at all, because motion signals are required for the production of the higher-level signals that will initiate a behavioural response. A frog left in a cage with freshly killed (unmoving) flies will starve to death, because it has no equipment for sending the signal: there is a fly (moving or still). Dangle a dead fly on a string and the frog will eat it.9 The difference in degree of complexity and vividness between frog and human perception does not warrant the assumption that there is a difference in kind – however much we may feel that a picture is worth a thousand words.10
XVIII COLOURS
Getting rid of images as things present to consciousness is also getting rid of the qualities that these images would have to have, were there any. For some of these qualities it is a clear case of good riddance. We can all do without the dimensionlessness of mental images (that strange quality that prevents us from putting any kind of a ruler, physical or mental, along the boundaries of mental images), and their penchant for inhabiting a special space of their own, distinct from physical space. With colour, however, there are sure to be misgivings. If anything is a quality, one is inclined to say, colour is a quality, and physics tells us it is not a quality of the ultimate particles that make up the physical universe. Yet colour as a quality is eminently spatial (‘Everything coloured is extended’), so it must exist in the phenomenal space of mental images. The Lockean distinction between primary and secondary qualities has remained compelling in spite of several centuries of rebuttal, and it tends to lead to the view that the primary qualities (the real, physical properties of particles) somehow work to produce in our consciousness the secondary qualities (seen colours, heard tones, the feeling of heat and cold) and that these qualities are on an ontological footing with the primary qualities but somehow insusceptible to analysis and explanation within the physical sciences. They are real, but ‘emergent’, the essentially novel and unpredictable product of a fantastically complex collaboration of the primary qualities of particles.
Locke did not, of course, say that primary qualities produced secondary qualities in us, but produced the ideas of secondary qualities in us. For Locke, whose view was clearly imagistic, this amounted to much the same thing; an idea of a certain shade of red was that shade of red, had that shade of red for a quality. But if, in the interests of a truth more important here than any biographical truth about Locke’s beliefs, we misread Locke, he can come out saying something that involves no ontological problems about the status of colours as existing qualities. If ‘idea’ is stripped of its imagistic connotations, having an idea of a colour need no more involve the existence of anything mental that has the colour as a quality than having the idea of a unicorn involves the existence of anything that is a unicorn.
The notion that colour words mean what they do in virtue of their being used by people as names for inner, private qualities succumbs to an argument that has become familiar in many forms: if, say, ‘red’ were the name of a private, inner quality, then one person could not teach the word (its use or meaning) to another, for he would never know whether or not his pupil was associating the word with the right private quality. If the teacher used a collection of red objects as props to help his pupil, he would either have to assume that every time he held up a red object his pupil experienced the right private quality, or admit that he could never know whether his pupil had caught on, in spite of the fact that his pupil called all and only those things red his teacher did. The latter alternative is absurd, and the former amounts to a disguised admission that ‘red’ has a public reference and criterion; if the pupil’s agreement with his teacher’s public use of ‘red’ is satisfactory to show that the pupil has learned the word, we can, as Wittgenstein said, ‘divide through’ by the private quality, which is superfluous to the analysis. In other words, supposing colour words to refer to inner qualities is another case of the error, familiar from Chapters 4 and 5, of taking the process to be analysed and using it as an unanalysed part of one’s analysis. In this case, in an effort to understand how a person ascertains the colour of some external object we launch our analysis with the assumption that a person does this by ascertaining the colour of some private, internal object.
If colours are not private qualities, and if they are also not among the primary qualities of physics, we seem almost to be left with the intolerable position that nothing at all is coloured. But this is absurd. When a person says he is looking at something red he is not describing any internal event, he is describing something external, and he can be right or wrong, so redness must be a property of external objects. But what property? The facile supposition is that colour properties are reflective capacities of surfaces that can ultimately be characterized in terms of physical structures, probably at the sub-atomic level, but unfortunately this is already known to be false. The relation between colour experiences (being aware1 that something is red, mauve, green) and light waves striking the retina is not at all the one-to-one correspondence one might expect. A variety of different combinations of wavelengths can produce the same ‘experienced colour’, and even monochromatic light can produce, under certain conditions, the experience of a wide range of colours.11 Then, although the sub-atomic characteristics of surfaces that reflect light predominantly of one wavelength can now be described in some detail, these different types of surface do not correspond neatly to the colours we observe things to be. It is possible, in fact, that two reflective structures both producing the experience of a particular colour might have no characteristics in common to distinguish them from reflective structures producing different colour experiences except the mere fact that they do produce the same colour experience in people. If this were the case, what would someone be saying when he said something was red? He would be saying that it had reflective property x or y or z or …, and the disjunction of properties associated with one colour might be very long.12 If the surfaces seen as one colour did in fact have nothing in common with one another except that they were seen as one colour, should we say that the phenomenon of colour is an illusion, that red and green and blue are not real properties in the world?13 This can be made plausible by considering the case of colour-blindness.
A man might be colour-blind only to red and green and if he did not know he was colour-blind, he would probably suppose that grass and ripe apples, fire engines and wine bottles had something in common – after all they were all the same colour (call it gred). Since, he might speculate, they are all the same colour, they must share some objective, intrinsic property of their surfaces. This would be an unwarranted leap, however, for the fact that these things were all one colour and not two would be due to an idiosyncrasy in the man’s visual system, and not any common features of the objects. We might say that the man’s experience of grass and ripe apples as both gred was an illusion. Then what are we to say of normal colour discrimination? If it happened to be the case that fire engines and ripe apples were not both red ‘for the same reason’, by parity of reasoning with the case of the colour-blind man should we not say that our seeing them the same colour is an illusion? Suppose there were another race of creatures relative to which human beings were colour-blind. We can suppose, for example, that they see fire engines as one colour and apples as another entirely different one. If all the surfaces they saw as one colour did in fact have some structural property in common, would we say that they had veridical colour vision, in contrast to human illusory colour vision? Or consider a race with an even more sensitive discrimination system, one which varied directly with length of light wave entering
the eye. Wouldn’t such a creature have the truest colour perception of all, since it would see red if and only if light from the red band of the spectrum entered the eye? To suppose that either of these races had a truer perception of colour than human beings is to lose sight of what we mean by the colour of a thing. The latter race would report that things were constantly changing colour, that you couldn’t count on a thing remaining the same colour over even a short period of time, but given what we mean by colour, this just is not true and any being that saw things that way would be suffering from colour illusions. The very meaning of our colour words is anchored in such facts as that red things are the ones that look red, and stay looking red, under most conditions. The former race, which saw fire engines as one colour and ripe apples as another, would be similarly viewed as having faulty colour vision, not truer colour vision – for they would be unable to tell when an apple was just the same shade as the fire engine, or when a lady’s handbag perfectly matched her coat. Paintings that we found to have a proper colour balance might seem garish and confused to these creatures; they might see the delicate shadings of a Botticelli as a series of contrasting bands of colours. Where the argument that our colour vision might be illusory goes wrong is in supposing that what we mean when we say something is red or green or blue is that the thing has some sub-atomic surface structure. In fact what we mean when we say something is red is just that it is the same colour as ripe apples and glowing embers. This is the true subjectivity of colour qualities: not that they are private, internal qualities, but that red things are all and only those things taken by normal human beings to be red, regardless of their surface structures or reflective capacities. This subjectivity does not, of course, prevent information about colours from being useful. We can rely on our colour vision to tell us that iron is rusting, bananas are ripe, solutions contain copper, snakes are of a poisonous variety. We might know, for example, that the only snakes in a certain area which were poisonous had spots of a certain shade of green on their backs, and might hand out samples of this colour on little cards to a group of hikers. What would then be important would not be that the snakes’ spots and the paint on the card share some structural property, but only that they be the same colour.
Colour, then, is not a primary physical property like mass, nor is it a complex of primary properties, a structural feature of surfaces. Nor again is it a private ‘phenomenal’ quality or an ‘emergent’ quality of certain internal states. Colours are what might be called functional properties. A thing is red if and only if when it is viewed under normal conditions by normal human observers it looks red to them, which only means: they are demonstrably non-eccentric users of colour words and they say, sincerely, that the thing looks red. Their saying this does not hinge on their perusal of an internal quality, but on their perception of the object, their becoming aware1 that the thing is red.
8
THINKING AND REASONING
XIX PEOPLE AND PROCESSES
The developing picture of consciousness as merely awareness1 of the contents of certain states or events leaves no room for the common and unreflective view of consciousness as the place where thought processes occur. Thinking and reasoning are things that we do, not merely experiences of which we are aware, but if thinking is in fact an ‘activity of mind’ or reasoning a ‘process of thought’, we seem to have a dilemma. To repeat Lashley’s dictum, no activity of mind is ever conscious. We have, of course, no introspective access to the actual processes of the brain, but neither do we have introspective access to what may be called the activities of the mind; we can think a thought in dactylic hexameter, but we have no inkling of how this is done – it just ‘comes to us’ from we know not where. Yet what is thinking and reasoning if not a ‘conscious activity of the mind’, an activity of which we are aware?
There are several separable senses of ‘think’. There is a sense related to belief or opinion (‘what do you think of that?’, ‘I think that …’); a sense alluding merely to our ‘stream of consciousness’ (‘I can’t stop thinking about her’); and then there is the sense of interest to us here, connoting purposeful and diligent reasoning, as in the sign on the office wall ‘Think!’. The sign does not exhort the workers to have opinions, nor is it the unnecessary directive to have a stream of consciousness. In some way or other thinking in this sense, or reasoning, is a process, for it takes time, can leave us exhausted, go astray, be difficult, bog down.1 Yet we are aware of this apparently internal process in some rather special way, for whereas if I am aware at all of my digestive processes it is only by observation and inference, my access to my reasoning is more direct. There are two questions before us, then. What is going on (what process is it) when someone reasons or thinks, and what access does he have to whatever is going on? Answering these questions will hinge, as before, on drawing the distinction between the personal and sub-personal levels of explanation.
The absurd view of thinking and reasoning, but one which occasionally infiltrates current thinking about the nature of mind, is that consciousness is an arena into which are led propositions, thoughts, logical operators and universal rules. The logical operators, like drill majors, direct the propositions into proper marching order, subsuming particulars and classifying concepts according to the behests of the universal rules, and then produce out of thin air a conclusion to bring up the rear. The audience is the introspective eye which reports this ‘marshalling of thoughts’ to the world at large. Akin to this view is what might be called the ‘hammer and tongs’ view of thinking or reasoning. One supposes that there are conscious acts of reasoning, acts of judgment and acts using concepts, and on the model of public acts we expect some organ, arm or tool to be acting on some object or some raw material – all this within the arena of consciousness. An act, one is tempted to say, cannot be a blank nothing acting on nothing, so there must be both agent and objects in consciousness. Ridiculing these views is not, however, finding an alternative that adequately accounts for what is remotely plausible in them. Something is going on when we reason, and an account of this must be found.
One account that has been offered tries to solve the puzzle by denying that reasoning is an internal process at all. Reasoning, according to Ryle, is a social activity – nothing more than the expounding of already formulated argument, usually to other people: students, colleagues, juries.2 This may be one sense of the term, but it is not what one is talking about when one wonders whether animals can reason, or says someone is not very sharp at reasoning. Were we to adhere to this sense of the term, the question whether animals can reason would receive an abrupt answer: since they cannot expound, they cannot reason. Ryle’s attempt to make the public, external presentation of arguments the fundamental and underived activity of reasoning is unconvincing, but it has a germ of truth in it. Reasoning is not, as Ryle urges, fundamentally a social activity, but it is fundamentally a personal activity. As Ryle points out, such quasi-logical verbs as ‘conclude’, ‘deduce’, ‘judge’ and ‘subsume’ do not refer to processes at all, but are used in the presentation of results already arrived at.3 One could not design – let alone construct – a concluding device – for the only time one concludes, e.g., that Smith is the culprit, is when one says or writes or thinks to oneself: ‘and so I conclude from this evidence that Smith is the culprit’, or ‘Aha! So it was Smith all along!’ One could make a device to utter these words, but just saying the words is not reasoning. If reasoning is a process, it is not a concluding process, and if it is made up of operations (e.g., ‘logical operations’), none of them will be concluding operations.
The fact that in using such verbs as ‘conclude’, ‘deduce’ and so forth we are not describing or naming operations or processes which we somehow observe does not license the conclusion that there are no temporal operations or processes going on behind our announcements of conclusions. People arrive at conclusions, and, as the bland verb suggests, this is not a process that people go through or an activity in which they eng
age, so we cannot ask the question ‘How do you arrive at a conclusion?’ and expect an answer in the form ‘First I do this, and then I do that’; people do not do anything in order to arrive at conclusions, but their brains must. The distinction between the personal and sub-personal levels of explanation is nowhere more important than in the area of thinking and reasoning. People can reason, but brains cannot, any more than feet (or whole bodies) can flee or a hand can sign a contract. People can use their feet in fleeing or their hands in signing a contract, but it would not be correct to say in the same sense that people use their brains in thinking and reasoning. We say ‘Use your head!’, but if this were understood in the same sense as ‘Use your index finger!’ said to a violinist, we would be at a loss to know what to do.
Were we to take what goes on in the brain and analyse it into parts, we should not expect those parts to be, say, concluding or deducing operations, for that is to confuse levels,4 and yet some operations of a different sort must occur. When computers are made to perform logical operations, the abstract, timeless transformations and operations of logic are realized in physical, temporal operations, and the production of results or conclusions takes time and energy. That there must be analogous processes in the brain can be seen by considering what Bennett calls ‘mental trial-and-error’, the sort of pondering one does when one imagines various outcomes to one’s behaviour before acting. This is a sort of thinking or reasoning, and ‘sometimes we do this with words, and sometimes by a kind of imaginative and experimental picturing of the outcome of various possible courses of action.’5 Saying that reasoning of this sort is a matter of imagining different scenarios cannot be the whole story, however, for how is it that the outcome of a course of action follows in the imagination once we have imagined the course of action?
Content and Consciousness Page 18