Empiricism, the philosophical theory that all our ideas and knowledge are derived from experience, has in recent years been the target of radical and persuasive objections. In the seventeenth century, and for long after, rationalism seemed the only alternative to empiricism, but, like Kant, many contemporary philosophers have been convinced that empiricism and rationalism are equally unacceptable, and that both positions, and the conflict between them, are the result of trying to answer confused, misleading, and perhaps senseless questions. Of all the traditional theories in particular, none, I suppose, has looked to modern philosophers less capable of being revived than the rationalist doctrine of innate ideas: the doctrine that some at least of our ideas and knowledge of things are not derived from outside the mind, through experience, but are present from birth in the mind itself and thus represent the mind's own contribution to our understanding of reality. What Strawson has called ‘those old and picturesque debates regarding the origin of our ideas’ seemed to have become museum pieces, charming, as antiques can be, but with little relevance to current issues. These complacent attitudes, if that is what they are, have been rudely disturbed by Chomsky and his colleagues. Chomsky has suggested ‘that contemporary research supports a theory of psychological a priori principles that bears a striking resemblance to the classical doctrine of innate ideas’. In the prevailing situation, the novelty of this suggestion is twofold. It is not only that the theory of innate ideas may be true after all, but also that the support alleged for the ancient doctrine has a distinctively modern look. The research that Chomsky claims supports the doctrine is research in the empirical science of linguistics. The implication is that empiricism is refutable empirically, and more generally that a philosophical dispute such as the dispute between empiricism and rationalism, once it is made clear, can be settled by science.

In his contribution to Human Senses and Perception, R. J. Hirst has made a number of important suggestions about perceptual consciousness, (i) He has emphasised the need to describe ‘what the percipient is or may be conscious of’ from the percipient's own point of view (p. 294). This mode of description is contrasted with stimulus or neurological description. Perceptual consciousness of one object is distinguished from perceptual consciousness of another object ‘only by or on the evidence of, the person concerned’ (p. 295). The method of obtaining descriptions of perceptual consciousness is either to question a percipient or to reflect on our own experience, (ii) The second important point stressed by Hirst is that the end product of perceiving is ‘the conscious experience of external objects’ (p. 303). Such an obvious point is often lost sight of in behavioural, dispositional, or neurophysiological analyses, (iii) The third and final suggestion made by Hirst to which I want to call attention is the usefulness of a genetic hypothesis to explain and account for perceptual consciousness. Hirst feels that perceptual consciousness is ‘unanalysable at the conscious level’, meaning (a) that it is ‘a unitary awareness of objects or scenes’ and (b) that the ‘various interacting unconscious activities’ which coexist with awareness ‘cannot be brought forward into consciousness’ (p. 305). The various analytical theories of perception (e.g. the traditional empiricist, the sense-datum) have been designed to break perceptual consciousness into its components. This has resulted in their being false to perceptual consciousness. A genetic explanation of awareness has several advantages, prime among them being its ability to explain the complexity and development of awareness.

I shall attempt something rash in this paper. I shall draw your attention to some past and current work on perception by psychologists and others. I shall concentrate on work in vision and hearing. This outline will occupy the first part of my lecture. I shall then go on, in the second part, to suggest that this scientific work has certain philosophical implications. This whole attempt is a bit rash for obvious reasons. It is not easy to outline fairly and accurately past and current work in any branch of science. I am very liable, therefore, to do an injustice to the efforts of psychologists and others in this field. What makes matters more difficult for me is that I also have to show that this work is of philosophical interest. What has led me to embark on this perilous enterprise is a hunch I have developed in recent years. I have the hunch that philosophers who are interested in perception would do well to pay rather more attention than (perhaps) they have been wont to do in the past to the work and discourse coming out of the scientific laboratory and similar places.

In his book Attention, Professor Alan White says ‘When you see X, it follows that if X is Y, you see Y whether you realise it or not.’ If, in passing through Paris, I saw a tall complex iron structure and that structure is the Eiffel Tower, then I saw the Eiffel Tower whether I realised it or not. I accept this, but because recent philosophical writings and discussions have cast doubt on the validity of the inference-pattern

I saw x; x is y; so I saw y

and certain related patterns, it is clear that we cannot be content with this unvarnished statement. Various entertaining examples are produced to show that some instances of this pattern are invalid and therefore that the pattern itself is invalid. If I saw Jones at noon and at noon Jones was bribing Smith then, it is alleged, I cannot conclude that I saw Jones bribing Smith. Similarly, it is said, from the facts that I saw a man in the far distance and that that man was my father, I cannot conclude that I saw my father in the far distance; from the facts that I saw a foot and that that foot was Lloyd George's I cannot conclude that I saw Lloyd George.

There is an ancient and ambiguous philosophical doctrine that perception is passive. This can mean that the mind contributes nothing to the content of our sensory experience: its power of perception is a mere receptivity. In this sense the principle has often been questioned, and is indeed doubtful on empirical grounds, given one reasonable interpretation of what it would be for the mind to make such a contribution.

The explanation of change or movement has always been a central concern of philosophers. Some, like Aristotle, have taken the movement of living things as their paradigm, and tried to explain all movement or change in that way. Others, after the fashion of Descartes, concentrate on the movement of inanimate things and generalise explanations of this to encompass all movement or change. For Aristotle, things have a principle of growth, organisation and movement in their own right. The movement or change of a natural thing is explained by its tendency to move in that way. The line he draws is not, as the line which we would perhaps like to draw is, between organic and inorganic things, but between these grouped together as subject to the same kind of explanation and, on the other hand, artificial things. A problem that results from this division is that while it might seem plausible to explain changes which occur in a baby when it grows into a man by saying that babies naturally tend to grow into men, and if they do not then something has interfered with their natural development, it seems odd to treat inorganic things in this way. Restricted to the contrast between the natural and the artificial, the explanation of stones falling when unsupported is clearly going to provide some difficulty. Although it is true that Aristotle does not think that because in the case of man the form with which matter is formed to make that substance is called a soul, that therefore any kind of form joined with matter to make a substance is called a soul, nevertheless the explanation of things which are a combination, a natural combination, of form and matter is the same for both man and other substances. Confronted with the explanation of falling bodies in this way, it first of all seems implausible and then suggests that things should happen which in fact do not. If it is assumed that the principle of movement is in the stone, ought it not to be assumed that the principle of stopping is in it too? Babies grow into men because it is in their nature to do so perhaps, but that stones fall downwards because it is in their nature to do so has, as Molière noticed in the case of a similar explanation of why opium puts you to sleep, a hollow ring.

Professor C. A. Mace, the psychologist, once wrote: ‘It is difficult … to present and defend any sort of behaviourism whatever without committing oneself to nonsense.’ I shall illustrate this thesis. I shall comment on the writings of some psychologists. This is relevant to my topic; for psychologists' expositions of behaviourism contain much more philosophy than science, and the inconsistencies which permeate their versions of behaviourism reappear in the works of eminent philosophers. My quotation from Mace comes from a paper defending what he calls ‘analytical behaviourism’; which he distinguishes from ‘methodological behaviourism’ and ‘metaphysical behaviourism’. According to Mace, analytical behaviourism does not question the truth of our everyday statements about a person's mind or states of consciousness; what it claims is that such statements ‘turn out to be, on analysis, statements about the behaviour of material things’, that is, about a person's ‘bodily acts, bodily states, bodily dispositions, bodily “states of readiness” to act in various ways’. The father of behaviourism, J. B. Watson, rarely says anything suggesting this doctrine. As he presents it, behaviourism is both a methodological principle and a metaphysical theory.

When we try to think about the causal nexus and the physical nature of the world as a whole we may be struck by two quite different difficulties in finding room in it to accommodate together (a) knowledge or reasoned belief and (b) causal determinism. (a) may seem to us to exclude (b) and (b) may seem to us to exclude (a). Taking it as a fact that there is knowledge and that knowledge seems to be indefinitely extensible, it has been felt by some philosophers that we can disprove total determinism by showing that if there were laws of nature which purported to govern all movements of matter in the universe there would still be something which even an ‘all-knowing’ predicter could not predict, viz. his own predictions or his own actions; and that given enough knowledge any agent could refute anybody else's predictions of his actions. So it has been thought that the phenomenon of knowledge somehow shows there cannot be laws to govern all movements of matter in the universe. This comfortably anodyne reflection is examined in the second part of the lecture. It elevates human minds and even confers a sort of cosmic importance on them. The other difficulty in making room for both (a) and (b) is in some loose sense the dual of this. Instead of taking knowledge for granted and questioning total determinism, it merely takes causality for granted but then deduces the total impossibility of knowledge. It simply asks: ‘How can we take a belief seriously, or consider it seriously as a candidate to be knowledge, if it is no better than a simple physical effect?’ This is a more pessimistic reflection and I shall begin with it.

The object of this paper is to examine what evidence we can have for or against the truth of determinism, a doctrine often set forward by the proposition ‘every event has a cause’. I understand in this context by the cause of an event a set of prior conditions jointly sufficient for the occurrence of the event. Since the determinist is concerned with all physical states and not merely with changes of states, which are most naturally termed events, we may phrase this claim more precisely as follows: There is for every physical state at some earlier instant a set of conditions jointly sufficient for its occurrence.

While philosophers feel relatively comfortable about talking of the present and the past, some of them feel uncomfortable about talking in just the same way of future events. They feel that, in general, discourse about the future differs significantly from discourse about the past and present, and that these differences reflect a logical asymmetry between the past and future beyond the merely defining fact that the future succeeds, and the past precedes, the present time. The problem is: how can we talk about events which have not yet happened, or at any rate are not yet bound to happen, or whose participants do not yet exist? The effect of these worries has led them to claim to recognise restrictions on our talk about the future which do not govern talk about the past and present. The most famous of these views is Aristotle's. According to one familiar interpretation, he holds that a statement about a future event which is not yet settled, a contingent event in the future, is neither true nor false, even though the statement that the event either will or will not happen is necessarily true. Proponents of this view felt that if a future-tensed statement were already true then the fact that it stated would already be settled. I do not propose to discuss this well-known and muchdiscussed doctrine of Aristotle's, but I do want to consider some allied views which have been aired recently, and to look at their philosophical significance. Before I look at these, however, it will be convenient to recall three of the main reasons why the Aristotelian doctrine is unpopular. In the first place it is paradoxical to accept that a statement of the form p v ∼ p is (necessarily) true while claiming that neither of its disjuncts is true. Then there are misgivings about the notion of truth involved: many feel that truth is essentially an attribute of timeless propositions and that it is nonsense to talk of a statement's becoming true as you would of Aristotle's views if the event described became inevitable. There is also the difficulty of accounting for the meaning of a future-tensed sentence which may express a statement that is neither true nor false simply because what it states is not yet settled. It could not be said of the sentence expressing such a statement that you know what it means if you know what it is for the sentence to express a true statement. I know the meaning of the present-tensed sentence ‘A sea-battle is now being waged’ if I know that it can normally be used to make a true statement precisely in the event of there being a sea-battle being waged at present. But I do not know the meaning of the future-tensed sentence ‘A sea-battle will be waged tomorrow’ simply by knowing that the sentence expresses a true statement if it is already settled that there is going to be a battle: the statement doesn't mean that the battle is already settled, otherwise it would not lack a truth-value when the matter was still open – it would be false.

As in the case of great books in all branches of philosophy, Pierre Duhem's Le Théorie Physique, first published in 1906, can be looked to as the progenitor of many different and even conflicting currents in subsequent philosophy of science. On a superficial reading, it seems to be an expression of what later came to be called deductivist and instrumentalist analyses of scientific theory. Duhem's very definition of physical theory, put forward early in the book, is the quintessence of instrumentalism:

A physical theory is not an explanation. It is a system of mathematical propositions, deduced from a small number of principles, which aim to represent as simply, as completely, and as exactly as possible a set of experimental laws [p. 19].

In this paper I want to discuss two separate problems about memory, connected in that they both have to do with memory as a source or ground of knowledge.

If my argument here is successful, it will follow that this paper should have been included in the first volume of this series, The Human Agent (Royal Institute of Philosophy Lectures, 1966–7); for I want to contend that the topic of dreams belongs to philosophy of mind rather than to theory of knowledge.

What is the function of moral principles within the body of moral knowledge? And what must be the nature of moral principles in order for them to carry out this function? A specific set of answers to these questions is widely accepted among moral philosophers – so widely accepted as almost to constitute a sort of orthodoxy. The answers embody a view of the place of principles within the body of morality which crosses the lines between cognitivism and non-cognitivism. Though I have put the question in cognitivist terms and shall discuss it in those terms, I think a similar question and a more or less parallel discussion could be given in non-cognitivist terms. Perhaps the time-honoured debate between the two positions can be suspended, at least temporarily, while we examine, not the nature of morality, but its structure.

Those who despair of the possibility of proving the existence of God tend, naturally, to hold that knowledge of God's existence and of those religious claims that depend upon it can only be had, if it can be had at all, through some direct religious awareness or insight. On this view appeals to authority or to revelation rest on appeals to such insight, if it is agreed that the credentials of the revealing authority cannot be established by the methods of natural theology. It is common for debate between believers and sceptics who share this despair about the possibility of proof to take on an air of hopelessness and unreality because of a fundamental epistemological cleavage: on the one hand the believer has an allegedly cognitive experience and on the other the sceptic lacks and suspects it. I want in this paper to scrutinise some aspects of this division. I shall not do much to mitigate the pessimism of my earlier statements, since I think the division really is, in certain critical ways, an unbridgeable one. But it is worth while to come to a clearer understanding of its nature than I think some philosophers have. What follows has been influenced by reflection on recent controversies about the meaningfulness of religious discourse, but is not intended to be a contribution to them. Some of the best-known contributions, however, seem to me to have made the epistemological cleavage I have referred to seem even worse than it is.