CHAPTER VII

LANGUAGE, PERCEPTION AND ACTION: PHILOSOPHICAL ISSUES

The earlier part of this book has been concerned with very specific questions arising in the field of linguistics (phonetics, semantics and syntax), with the results of research into visual perception (physiological and neurological) and with rather wider speculation about the organisation of bodily action and the relation between the bodily processes underlying action, vision and speech. The hypotheses, arguments, evidence and conclusions reached have not depended to any significant extent on philosophical doctrine or concepts and the question may be asked why should a book essentially concerned with linguistics conclude with a chapter devoted to philosophy. To this question there is a broad answer and a more specific one; the broad answer is that there has been prolonged and difficult discussion between philosophers over many centuries of the subjects dealt with earlier in this book, the origin and nature of language, the relation of language to reality, perception as based on sense-experience and providing the main basis for veridical knowledge, and voluntary human action (the notions of free will and determinism, of reasons and causes of action). The narrower answer, as an occasion and justification for having a philosophical chapter, is that in some respects totally new broad and specific hypotheses are presented about the functioning of language, perception and action, and particularly about their interrelation in human behaviour, and it is worth considering what implications these hypotheses, if true, may have for traditional or current philosophical views. It may be that they ought to involve some radical review of current theory but. in any case, it would be unsatisfactory simply to present a whole range of ideas bearing on language, perception and action without having regard to what relevant to these subjects has been said by philosophers (as in the same way it would be unsatisfactory not to have regard to work that has been done on these subjects by experts in the field of Artificial Intelligence).

Philosophical discussion. in ancient and modern times, has tended to deal separately with language, perception (including knowledge based on perception) and voluntary action, and at different periods the focus of attention in each of these fields has tended to be significantly different. The pattern followed in this chapter is to treat first the philosophical concern with language, and then philosophical discussion of perception (and knowledge derived from perception) and voluntary action. It seems appropriate to deal first with language not only because this is a book about linguistics but also because in the present century philosophy has taken what some have described as 'the linguistic turn'; far more than in any preceding period of philosophy, philosophers have concentrated their thought on language in one way or another. (see Rorty(1)). Indeed, in so many respects have philosophers dealt with language that it becomes something of a problem to classify the separate approaches. Conventionally, one can distinguish between: 1. the philosophy of language 2.linguistic philosophy (in the sense of various uses of linguistic methods in investigating philosophical questions 3. the philosophical aspects of modern linguistics 4. the unavoidable dependence (regardless of any theories about language) of philosophy on language as the medium for philosophic exposition and debate 5. the specialised use of language as a philosophical device in modern logic and the construction of formal languages ( rather different in approach from what is described above as linguistic philosophy). Besides these wide classifications, there are other topics which do not fit neatly into the categories: language universals, innateness of ideas, the interaction of language and culture or of language and thought, the dispute about private language, language as a form of action (Austin), Obviously, together these categories and topics constitute an immense field which it would be presumptuous to attempt even to describe, quite apart from venturing on any detailed discussion or assessment. What makes the position more complicated is that the distinctions between the different approaches to language in philosophy are ones which cannot be made sharply; there would be disagreement amongst philosophers about whether what they have done or are doing should be described as philosophy of language or linguistic philosophy or simply as just philosophising. One view is that (partly because of the rapid growth of analytical philosophy relying on linguistic methods) the philosophy of language (as part of the subject-matter of philosophy) has occupied the central place in the entire enterprise of philosophy in the twentieth century; another is that linguistic philosophy. that is the general use of linguistic methods, has succeeded in putting the whole previous tradition of philosophy (from the Greeks to the modern metaphysicians) on the defensive. The list of philosophers who have in fact concerned themselves heavily with language (even if some of them would reject any description of their work as philosophy of language or linguistic philosophy) includes virtually all the most familiar names in this century in English or American philosophy (Russell, Moore, Wittgenstein, Strawson, Quine, Austin, Ryle, Ayer) and, though Continental philosophy for the most part has found the Anglo_Saxon preoccupation with language odd or incomprehensible, there has also been a substantial contingent of German, Austrian, French and other European philosophers occupied with language in one way or another (Frege, Carnap, Schlick, Derrida, Merleau-Ponty and, earlier of course, Humboldt and Herder).

Whilst detailed treatment of linguistic philosophy (in the broadest possible sense) is, as has been said, impracticable, some brief description is needed to make it possible to establish points of contact between themes and debates in current philosophy involving language and the hypotheses and discussion about the functioning of language expounded in the earlier chapters. Though the exact nature of philosophy as an activity is itself still one of the disputed issues - to reveal the nature of the world (as traditional philosophers believed) or to bring order into one's conceptual structure as some modern philosophers would say - linguistic philosophers would nevertheless see themselves as a continuation of the philosophical tradition and part of the perennial philosophical enterprise. The progression in philosophy in this century has, in general terms, been away from the grand metaphysical structures of the 19th century, to a search for reliable knowledge of the world through language and logic, then to the attempt to construct criteria for the truth of propositions by linguistic methods, then to a more modest attempt to clarify intellectual concepts through the study of ordinary language and finally perhaps to a disenchantment with language and linguistic methods, with signs of a reversion to a larger view of the objectives of philosophy (including the revival of what some might describe as more traditional metaphysical approaches). In terms of particular philosophers, the trend has been from the idealism of Bradley to the attempt to reform language (coherently with mathematics) of Russell and Whitehead(2) and the early Wittgenstein(3) (later developed by Carnap), to the Vienna school and the logical positivism of the early Ayer (using analysis to banish metaphysics), to the ordinary language approach to philosophy of Austin and Ryle and finally to the disenchantment with philosophy of the later Wittgenstein. The swing back to a larger view of philosophy can be seen most clearly in Continental philosophy (an interest in Hegel, the strength of phenomenology) but also (prematurely so to say) in Whitehead and most recently in some of the work of Strawson.(4) The immediately following paragraphs contain not an exposition but some representative statements indicating the different approaches there have been to language and its significance for philosophy.

Russell and Whitehead (and the early Wittgenstein) were, in the use and function of language, concerned to construct a new logic of propositions (of sentences asserting truths or falsities), as distinct from Aristotle's logic of classes; they were in a sense concerned with syntax rather than with lexicon. They saw the proper mode of philosophical analysis as the rewriting of natural sentences to exhibit the correct logical form so that their meaning would become clear and philosophical perplexities would be eliminated; their objectives were those of traditional philosophy but the linguistic method was new. Russell's aim was, by this kind of analysis, to arrive at the universal syntax of knowledge, the universal structure of knowledge underlying the grammar of ordinary language. The tool Russell and Whitehead found, or constructed, for the task was the application of mathematical logic to language; with this tool, and the assumption that the structure of the world, and thus the underlying structure of language, was of the same nature as the theoretical structures of physics (elementary facts forming atomic propositions which then combined to form 'molecular' complex propositions), their belief was that, from the reform or reconstruction of ordinary language, one could arrive at reliable knowledge of the structure of the world. As Wittgenstein put it, in his early philosophy, a perfect language is like a map; it pictures the structure of reality, mirrors the world as a map mirrors it. Philosophy tells us about the structure of the world. Though Russell firmly rejected any classification of his work as linguistic philosophy, it was in fact very closely based on current ideas about the structure and grammar of language; for example, atomic propositions were said to be always in the 'Subject- Predicate' form and the elements to which the atomic propositions related were the 'facts' of which the world is said to be formed.

Technical difficulties (Russell's problems with the theory of descriptions and the paradoxes) as well as the trend away from traditional metaphysics of which logical atomism in Russell and Wittgenstein's formulation represented the last wave, led to a loss of interest in this kind of linguistic analysis and its supersession by a new but still basically linguistic approach, that of logical positivism. The new analysis shared with logical atomism the objective of clarifying language (scientific and colloquial) but with a much more limited end-product, that of showing whether or not propositions or sentences were of a kind which could be meaningful or not; sentences, on this new approach, could be analytic (true because of their structure), empirical (typically scientific propositions) or non-empirical. Mathematical and logical propositions were classed as analytic and there are potentially true and meaningful; the meaningfulness of empirical sentences depended on whether they could be verified by observation or at least a theoretical method of verification by observation could be proposed. Whilst attaching a central importance in philosophy to the use of language and logic, neo-positivism in effect involved the destruction of most traditional philosophy; analytic propositions (of mathematics and logic) survived but could not refer directly to the world; empirical statements could be meaningful, if verifiable, but the methods of verification were scientific rather than philosophical or linguistic; all non-empirical statements, those for which no method of verification could be proposed, were, on the neo-positivist view, meaningless, incapable of being true or false and in fact nonsensical. This last category covered in effect the whole of traditional metaphysics (and indeed the metaphysical implications of Russell and Wittgenstein's logical atomism).

However, as a use of language for philosophical purposes, logical positivism proved to be as vulnerable to attack as had been logical atomism or traditional metaphysics. It was pointed out that until we know the meaning of a sentence, we cannot start to say what kind of sentence it is and therefore whether it is analytic or empirical, or, if it is potentially empirical, what kind of verifiability should be looked for, that is, we cannot make the meaningfulness of a sentence depend upon its truth or falsity or even upon our ability in principle to demonstrate that it is true or false. The meaning of a sentence must be prior to the question of its truth. Apart from this, neo-positivism concentrated its attention on a peculiarly narrow section of language, that is descriptive or assertive sentences; in fact, language is used in many other ways, and used effectively, but for sentences of these other kinds, the issue of truth or falsity need not arise. Indeed, even though a sentence is of a kind which may be false or true, it can be unsuccessful in conveying anything true or false but still not meaningless. Finally, and perhaps most decisively, neopositivism is open to the fundamental attack that, on its tenets, no generalisation can be shown to be true because we cannot test an infinite number of instances; not only is it awkward to have a philosophical or logical theory which rules out as nonsensical the greater part of intellectual and scientific theory and discussion, but, more devastatingly, the tenets of logical positivism are, of course, themselves generalisations ('All propositions which are not analytic or empirically verifiable are meaningless') and there is no empirical method which conceivably could be specified for demonstrating the truth of these non-analytic generalisations. As Karl Popper(10) later pointed out, generalisations can be falsified, though they may not be verifiable - and the generalisations of logical positivism are open to falsifiability by, for instance, any simple proposition of which we understand the meaning but for which we cannot, for the present at any rate, indicate how, empirically, the truth or falsity of the statement can be established.

If logical atomism and logical positivism had proved unconvincing as methods of using language for philosophic purposes, the next step was overtly less ambitious but potentially more fruitful, that is, the development of the 'ordinary language' school (chiefly in Oxford). This was based on two main ideas: first, that ordinary language is more subtle and less confused than the earlier linguistic philosophers had supposed and something could be learnt from the study of use in ordinary language of key terms in philosophy. Austin was the principal exponent of this method and indeed went so far as to believe that philosophy was on the point of giving birth to a new science of language, through a systematic study of the functioning of colloquial language. The second main idea was that many of the problems of philosophers derive from their own misuse of ordinary terms in language; in particular, Ryle's position was that if one corrected the misuse of words like 'know' and 'mind', charted the logical geography of 'believe', 'doubt', 'infer' and so on, misconstructions and absurd theories would be revealed and many so-called philosophical problems would disappear, taking with them the 'ghost in the machine' which philosophers had constructed for themselves from the use of the word 'mind'(5). Beside the optimistic view of the 'ordinary language' approach to philosophy, there was also a pessimistic view, espoused by the later Wittgenstein, after he had rejected the earlier certainties of the Tractatus Logico-Philosophicus; he saw the main problem in philosophy as being to escape from the 'bewitchment' of philosophy by language and at the same time the need to disabuse philosophers of their belief that philosophy could in reality achieve any substantial results. The insignificance of language for philosophy he explained through developing his view of 'language games', that the meaning of a word is no more and no less than the way it is used (in effect a symbolic convention within a particular group or community). Philosophy, according to the later Wittgenstein, is not a theory (or a system of theories) but just one kind of activity; its object is simply to help to clarify one's thoughts; it has no application to reality and it is meaningless to enquire whether one's conceptual scheme mirrors reality correctly or not(6). At best, on this pessimistic view, linguistic analysis is a therapy to correct disordered thought and to recover from the intellectual malaise which an addiction to philosophy constitutes. But in the end Wittgenstein was uncertain whether even this pessimistic conclusion was correct; in his last work, published posthumously under the title 'On Certainty' (and largely concerned with the nature of 'knowing'), amongst a variety of scattered thoughts he included: "If you tried to doubt everything you would not get as far as doubting anything. The game of doubting presupposes certainty" and "Our knowledge forms an enormous system. And only within this system has a particular bit the value we give it"(7) - which seems to imply a final reconversion to the significance of philosophy as the enterprise in the sense of seeking to understand the world.

The most recent turn in the relation between language and philosophy has come partly from Austin's development of ordinary language philosophy into a view of language as action, partly from a new interest in the problem of meaning and partly from the change of direction in linguistics (Chomsky's ideas on the 'deep structure' of language and the innateness of grammar). Austin's views on language as action precede but fit well with views developed in linguistics of the functional basis of language. Whilst the categories of action which a sentence performs were at different times analysed in somewhat different ways by Austin, his key point was that "the basic unit of study is not the word or the sentence but the act which a person performs with the aid of words or sentences; the theory of language is to become, as it were, a branch of the theory of action... the total act performed by the speaker with the aid of the sentence"(8). Eventually, his view was that all sentences are performative, whether they convey information (which he previously classified as 'constative'), induce action (by warning or persuading) or constitute an action in their very utterance (as in promising). The impact of linguistics on philosophy has come both from the discussion of meaning and from the investigation of syntax; Quine at one point commented that pending a satisfactory explanation of the notion of meaning, linguists in semantic fields were in the position of not knowing what they were talking about - but the problem of meaning is now being tackled from a number of directions by philosophers, logicians, linguists and others. It is becoming apparent that the problems of meaning and the problems of syntax are not separate but interlocked, and that the problem of meaning is (as already noted in a previous paragraph) separable from the problem of truth and knowledge. It is again becoming possible to be at the same time a linguistic philosopher and a metaphysician; Strawson has pointed out that the actual use of linguistic expressions remains the philosopher's sole and essential point of contact with the reality which he wishes to understand.. language proceeds directly from the concrete object somehow engendered by life itself; the objective is to understand the foundation of our concepts in natural facts and for this purpose "ordinary language is not a crude, rough and ready instrument but an enormously sophisticated instrument for thinking". In the attempt to show the natural foundations of our logical, conceptual apparatus in the way things happen in the world, and in our own natures, and how the fundamental categories of our thought hang together, "study of the relation between our whole articulated conceptual scheme and language, in other words linguistics, would be a source of the sort of questions and problems that philosophers are characteristically interested in"(9).

Of course, this revival of interest in the structure and functioning of language as an aid to philosophical advance is by no means universal or unchallenged. The contrary view is still held that philosophical problems are only muddles into which we are enticed by our misuse of language and that it is only with the use of language that truth and error, certainty and uncertainty, come fully upon the scene; only things expressible in language are capable of being true or false. Philosophers like Popper still reject any assumption that the subject-matter of philosophy is linguistic and, in Popper's phrase, consider that problems of language are secondhand problems, that philosophers who are not interested in science, rather than language, are no true philosophers(10). Or they agree with Max Black that the refinements of latter-day linguistics are impressive without being philosophically useful. "the conception of language as a mirror of reality is radically mistaken. Language must conform to the discovered regularities and irregularities of experience.. No roads lead from grammar to metaphysics"(11). Certainly, European philosophers for the most part would not concede the central role of language or that language is the sole point of contact with the reality which philosophy wishes to understand; their view would be that the behaviour of things is the foundation of our conceptual structure and at most language expresses hypotheses and inferences about the foundation of things (see the discussion of Strawson's view at the Royaumont colloquium between English and Continental philosophers)(12).

The philosophical views of language and its significance outlined above bear, of course, quite directly on many of the hypotheses and proposals advanced in the earlier part of this book and it would be possible to comment specifically on them at this stage in that light. However, since the fundamental thesis in the book is the integral relation between language, perception and action, a better course is first to review, again briefly, philosophical views on perception, and then to consider how the philosophies of language and perception taken together relate to the hypotheses about language, perception and action presented earlier in this book.

In philosophy, theories of perception have normally been derived directly from epistemology, from theories of knowledge. Epistemology has for centuries been a central interest of philosophers; the means by which knowledge is acquired, the extent of our knowledge and of our possible knowledge, and the criteria for judging reliability, truth or falsity. In examining knowledge, philosophers traditionally have been divided into two camps, the rationalists and the empiricists. The rationalists' view is that the only reliable source of knowledge is human reason; mathematics is the paradigm example of reliable knowledge arrived at purely by reason; all sense experience can be delusory and is therefore not a reliable source of knowledge; true knowledge is the result of the inherent properties of the human understanding (or, in the case or Plato and some other philosophers, of our gradually coming to grasp the true ideas with which we are innately provided). Much of the discussion has been in terms of what it means to say that we 'know' something; there has been much rather fruitless argument about the extent to which to say one 'knows' something must mean that the thing known is objectively so, that knowledge is a guarantee of certainty, a confusion between the subjective and objective aspects of 'knowledge'.

The more influential school, certainly in English and American philosophy, has been that of the empiricists, those who follow Locke in denying the existence of any innate knowledge and derive all knowledge from sense-experience. Apart from analytic truths, such as those of mathematics or logic, whatever a person claims to know must, however he may explain his acquisition of the knowledge, ultimately depend upon direct experience by some human being; empiricists admit that sense-experience can be misleading but absolute rational certainty is unachievable and though we may be deceived on occasion, normally we are not; the search for regularities and uniformities in the world of appearance is a worthwhile endeavour and has provided the basis for the great and rapid progress of science. Adherence to a philosophy of empiricism leads directly to a belief in the central philosophical importance of perception; Locke's prime assertion is that all our ideas (our notions, all the materials of reason and knowledge) come from experience only: "our senses .. convey into the mind several distinct perceptions of things; the second source of our ideas is the perception of the operations of our own minds" on the ideas it has got from sensation, "perception ... being the first step and degree towards knowledge and the inlet of all the materials of it".(13)

The discussion of perception, as a philosophical topic (in the empiricist framework) has for the most part proceeded completely independently of physiological or psychological research into perception. It has been in terms of sense-impressions as the basic particulars, of objects which have spatio-temporal position and those which, like qualities, do not. For Hume, the only items exempt from his universal scepticism were sense-impressions and mathematical concepts. The philosophy of perception in modern times was largely developed as a theory of sense-data (such as a patch of colour, a rap of sound, a felt surface or a smell); on this view, every external object is the source of a set of sense-data which fit together in a family, to form the perceived object, and any description of a material object inevitably involves the description of the sense-data which compose it. This account of perception, as an operation for combining and organising sense-data (such as colour, shape and so on) has been attacked from many points of view. How, it has been asked, does one distinguish what is experienced as a sense-datum from what is supplied by the mind in the perception of the object? Gradually, the sense-datum theory has been amended in an attempt to meet criticism; the amended sense-data are not independent realities but factors in a process of interaction between each person and other things in the world around him; or sense-data are pseudo-entities which stimulate us to perceive the presence of the colour of the object without themselves being the colour. Austin and Ryle both rejected the view that sense-data are the basis for perception and the foundation of knowledge (a view largely developed by their Oxford contemporaries, Pritchard and Price). Ryle(14) argued that the whole theory rested on a logical error, the identification of the concept of sensation with that of observation: "impressions (sense-data) are ghostly impulses, postulated for the ends of a para-mechanical theory, an invention of philosophers"; the notion of a sense-datum is not precise and cannot be explained in familiar terminology. Statements about physical objects are not formally translatable into statements about sense-data and the phenomenalist programme cannot be carried through.

But beyond this rather unsatisfactory philosophical treatment of perception as a source of knowledge, there are the profounder issues raised by the scepticism of Hume and the philosophies of Berkeley and Kant; can perception lead us to the knowledge of anything 'outside us'? Berkeley denied the existence of anything other than our ideas; Hume denied that there was any way in which we could know anything about the outside world and external objects and Kant, whilst postulating a priori knowledge consisting in the forms of understanding, by which external objects become known to us, denied that we could ever make any real contact with the transcendental objects of which the world is composed. Each of these philosophies leaves the nature of perception, as a relation between the human subject and the external world, unresolved - or resolved in ways which eliminate the notion of an external world or the possibility of knowing an external world if it in fact exists. The conundrums for the philosophy of perception that these 'idealist' doctrines present have, of course, been subjects for unending and inconclusive speculation, attempted refutation and, finally, indifference. One popular philosophical reaction has been to move towards a belief in 'naive realism', that is, regardless of the sceptical arguments of Hume, and the idealist positions of Berkeley and Kant, the common sense view that there are real physical objects outside us and we do establish contact with them through perception and action. Naive realism is a reaction shared not only by some philosophers (including the pragmatic philosophers) but also by most scientists and the common thinking man. The position is, no doubt, strictly philosophically indefensible but, in practice, the only possible one - the idealist philosopher may deny the existence of an external world but, in his everyday behaviour, act as if he were fully convinced of its existence. At the same time, the 'naive realist' may be no more interested in the scientific basis of perception than is the idealist; when Ryle's first book was published, he said that he knew nothing about psychology (though he was writing about the mind) and when he was criticised for having no knowledge either of neurology said: "It doesn't seem to me that one is not at home in the field of perception ... I do not want to know anything about the rods and cones in the eye".(16)

As Sloman has pointed out in connection with machine-vision, philosophers have normally ignored the complexities which come to light if one begins to design a working visual system, and even what are taken to be simple sense-data (patches of colour, lines, shapes) are found to be the result of complex processes of analysis and interpretation. Konrad Lorenz(15) has similarly commented that the barrier between philosophy and science has obstructed human knowledge in the very direction where it was most necessary i.e. the objective investigation of the interaction between the perceiving mind and the object of perception. Hume and Kant between them might both qualify for the description of 'destroyers of metaphysics', Hume because he treated our beliefs about the external world as 'a peculiar set of mental habits or customs' and denied the reality of external causation: 'necessity exists only in the human mind' and Kant because of the sharpness of the distinction he drew between the phenomenal (what appears to be the case to our perception and understanding) and the noumenal (the world of things-in-themselves). Kant's view was that we can know nothing about a real spatial or temporal world, the objects in it or the causal relation between events nor, even more constricting, can we discover within ourselves the cause of our a priori categories and system of perception, the source or cause of the organising principles of our mental machinery. As in the case of language, so in the case of perception and knowledge of the external world, philosophy has reached an impasse. Whilst, as someone has commented, it is not excessively uncomfortable to hold to a belief in 'naive realism', it is not a position defensible in philosophical terms. Meanwhile, the central problem of perception, the justification of our belief in the existence of perceivable physical objects, remains; the common sense view of ordinary men and of professional philosophers is that both the perceiver and the perceived exist and are related to one another through perception and action. Science, as in practice a supremely effective intellectual process, bases itself on the reliability of perception and the regularity of relations between events (that is, necessary causation), despite Kant, Hume and Berkeley.

There are ways out of the philosophical impasse. There is the way of pragmatism, to say that truth is determined by experience; the pragmatic test of meaning is whether ideas in fact serve as effective tools for coping with experience, for survival and successful action. By the pragmatic test of truth, universal scepticism and the unsurmountable separation between the perceiver and the perceived world must at some point involve fallacious ideas; Kant's view that the nature of things must always necessarily be unknown to us is 'a kind of nonsense'. Popper's way of dealing with the impasse in the philosophy of language is rather similar; language is no more than a kind of spectacles through which one looks at the world: "one shouldn't waste one's life in spectacle-cleaning or in talking about language ... To be only interested in language is a philosophical mistake ... leading to scolasticism"(10)

But the more convincing roads out of the philosophical impasse are those suggested by three authors who belong to quite different philosophical schools but resemble each other in having interests, in science or mathematics, going wider than academic philosophy. The first of these is Whitehead: the starting point for him is perception considered as a grasping, 'prehension', of its environment by an organism. For him, perception does not present us with isolated sensations; perception is experience from within nature of the system of events which make up nature, the self-knowledge enjoyed by an element of nature respecting its relations with the whole of nature in its various aspects; his theory of perception is a biological one, the perceiver as a natural organism reacting to the world around him. The views of the second author, Merleau-Ponty, a leading figure in French structuralism and phenomenology, are strikingly similar, though reached by a quite different route. His overriding theme is 'back to perception, restore the primacy of perception in philosophy'. His philosophy offers a middle way between rationalism and empiricism, between realism and idealism, in his concept of the 'body-subject', which encounters a world which already has 'meanings' incorporated in it, that is, a world in which this object is recognised as food, that as inedible and so on. At the same time as his concept of the 'body-subject' resolves the impasse in the philosophy of perception, it also, according to Merleau-Ponty, extricates us from the impasse in the conception of language: "language is itself an illustration of the dialectical relation between ourselves and the world and originates from the mixture of the world and ourselves" which precedes all reflection; for the child the name is the essence of the object and the child does not name the object but recognises it. Language is not composed of conventional signs but to Merleau-Ponty is seen as a form of 'psychic gesticulation', with meaning rooted not in abstract intellect but in real bodily gesture and behaviour. The understanding of a speaker's message (the psychic gesticulation which his words constitute) takes place in the same way as we understand his bodily gestures, which 'intermingle with the structure of the world that the gesture outlines',

Though Merleau-Ponty's terminology has some of the cloudiness associated with Continental phenomenology, the essence of what he says is not far removed from Whitehead's views, Nor again are these far removed from the much more straightforward and convincing exit from the impasse proposed by Konrad Lorenz, as an ethologist interested in human as well as in animal behaviour, For him, both the perceiving subject and the object of perception are equally real and all knowledge derives from the interaction between them. Leibniz at one stage had suggested, but not pursued, as a possible philosophical reconciliation between idealism and realism 'preestablished harmony' between knower and known. Essentially, this is what Lorenz as a biologist familiar with the facts of evolution suggests as the obvious answer to the Kantian dilemma: "the cognitive apparatus is itself an objective reality which has acquired its present form through contact with and adaptation to equally real things in the outer world .. The 'spectacles' of our modes of thought and perception, such as causality, substance, quality, time and place (the Kantian categories) are functions of a neurosensory organisation that has evolved in the service of survival ... What we experience is a real image of reality, albeit an extremely simple one, only just sufficing for our practical purposes"(15). For the individual human being, the cognitive apparatus, the modes of perception, are given a priori and determine the manner in which the real world is perceived or experienced. Though Kant might not have been able to see it as such, the 'perceiving apparatus' can be unconditionally equated with the Kantian synthetic a priori. But for the human race as a whole, the mode of perception is not a priori but the result of physiological and neurological evolution of the system of sense organs and nerves. The philosophical impasse of perception is resolved by the evolutionary and historical convergence of the organisation of the perceiving subject and the real perceived world.

The original plan of this chapter was first to give a necessarily summary account of philosophical theory and discussion bearing on language and perception as primary topics dealt with in this book and secondly to consider whether the broad and specific hypotheses advanced on the integration of language, perception and action in human behaviour have any implications for traditional or current philosophical views or whether they leave philosophical theory completely unaffected. Before the chapter tackles this second issue, it may be convenient to recall, briefly, what the hypotheses and proposals have been in the various earlier chapters:

- words, syntax and speech-sounds are not arbitrary. They are determined by anatomical, physiological and neurological structures and for each language, the structures of the language are derived from and directly related to other major segments of human behaviour (perception and action)

- the selection made by a language-community of its specific syntax, words and speech-sounds is not arbitrary or conventional but a selection from a range of possible sounds, words and syntaxes, with the community's preferences being determined by the pooling of genetic features of the population over time. Stability of a language is a result of stability in the genetic composition of the population coupled with the acquisition of a child's particular language by a process analogous to imprinting in animals and depending on the special character of the language as human behaviour

- language was not invented but emerged as part of evolutionary development, in parallel with the physiological, behavioural and social evolution of the human being and human communities

- languages are isomorphic because they are intertranslatable and their isomorphism derives from the integration of language with the general underlying structures and processes of perception and action

- the key to the evolutionary survival of the animal and of the human is the effective organisation of action (action appropriate to the environment). Perception developed from action and to serve action

- language, as a further stage, developed in groups and communities of human beings as a system for precise and flexible communication at a distance of the contents of perception of one individual to another i.e. language developed to serve perception and thus at one remove to serve effective action

- for perception to be an effective adjunct to action it must be integrated as closely as possible with the organisation of action (seeing and grasping must be physiologically and neurologically precisely aligned with each other), In the same way for language to be precise and effective in conveying the content of perception or in referring to action, it must, physiologically and neurologically, be fully integrated with the structures underlying perception and action (looking and speaking must be aligned with each other)

- each language forms a coherent system at each level (that of speech sounds, words and sentences). At each level, the forms and processes of language derive from and are integrated with the structures and processes of the corresponding levels of the systems of perception and complex voluntary action

- besides being integrated with the organisation of action, both perception and language are themselves forms of action (muscular movement in articulation and eye-movements) and form part of the total organisation of action in the human being

- the unity of the organisation of speech, perception and action rests in the neurological patterning of motor-control programmes in the cortex. One can conceive of the relation as three intersecting circles of speech-control, vision-control and action-control, with a shared central area where each intersects with the other and which, together with the self-reflexive character of language and perception and the fundamental unity of total action-organisation, provides the basis for the essentially unified character of human mental processes and behaviour

- research in vision and action-organisation (both in physiology and neurology) can throw light on research in language - and vice versa. Speech, vision and action all appear to be organised hierarchically in much the same way, starting from elementary units which have a specific relation to the construction in parallel of words, visual shapes and action-routines and then to the parallel syntaxes of the sentence, the visual scene and the complex action

- neurological research has provided a detailed basis for understanding the process of visual perception (at the more elementary levels) by identifying the relation between cells in the visual cortex and typical units which go to form a visual shape or the visual scene (lines, angles, curves etc.). Visual shapes appear to be constructed from a combination of the elementary units (in the first instance) with the formation of progressively more complex visual schemas (patternings in the cortex related to the sequence of eye-movements and fixations in scanning a visual shape or scene). The innate characteristics of the visual apparatus are the result of genetic influences (the product of evolutionary pressures); the predisposition to perceive the elementary units and patternings of the actual world is the result of 'imprinting', neurological pretuning to respond to a range of possible types of visual patterning. In the course of visual experience, the animal or the human being builds up more complex persisting neurological structures, schemas, which serve as short-cuts to the recognition of elements in a visual scene. The source of all the structuring found in a visual scene is in the physical visual stimuli, the stream of light energy falling on the eye, but the visual apparatus has evolved, both phylogenetically and in the individual animal or human development, so as to be predisposed to recognise the patterns most probably to be found in the environment

- analogous processes to those postulated for visual perception seem to operate in the production and comprehension of speech and the planning and execution of complex action. The meaning of a sentence can be treated as an equivalent simultaneously present neural patterning to the patterning which constitutes the interpretation, the meaning, of a visual scene, or to the plan of a proposed action. In each case, the translation of the meaning into a sentence, of a plan into an executed action, the comprehension of a heard sentence, the understanding of a perceived pattern of action or of a visual scene, can be analysed as a process of translating the simultaneous (present neural pattern) into a serial form (the spoken sentence or the executed action) or translating the serial (the heard sentence, the scanned visual scene or the perceived external action) into the simultaneous (a return to a cotemporaneous neural pattern)

- at the word level and at the sentence level, the integration between speech, vision and action operates equivalently. Each primitive word (that is, each word whose meaning is learnt from experience and not solely from explanation in terms of other words) has directly associated with it visual or action contours which indicate or give a natural clue to the meaning of the word. There is evidence for all languages of awareness of a natural relation between word-sound and word-meaning (sound symbolism) and considerable evidence that, for all except tonal languages, some kind of sound symbolism operates. "The word falls, one is tempted to explain", according to the early Wittgenstein, "into a mould of my mind prepared for it".

- at the sentence-level, the process of constructing a sentence or planning a complex action and of understanding (decoding or disambiguating) a heard sentence or perceived scene are analogous. They can be analysed in parallel. The key operation is disambiguation, the progressive narrowing down of the possible significance and role of the elements in a sentence or a visual scene. In this process, syntax and semantics operate together (both in vision and in language) to identify the word or visual schemas present and their interrelationship in the sentence or scene. In the operation (which is in effect one of problem solving), the interpretation is arrived at via a series of provisional hypotheses on the meaning and the relation of the elements and depends for its final certainty on the adequacy of the linguistic, perceptual or experiential context in which a particular sentence is uttered or a particular scene is perceived

- the total set of words in a language (the lexicon) and the total vocabulary of an individual are built up from a central core of primitive words (words learnt from experience only). The underlying structure of the lexicon or vocabulary reflects the structuring of the perceptual and other experience of the child or, primitively, of the language-community. The total lexicon is developed by regular processes of composition and extension of the meaning of words, by the borrowing of words from foreign languages (originally founded on the primitive elements in the vocabularies of the other languages) and, most importantly of all, by the vital process of metaphor, that is, the transfer of the structure of the percept to which a word or sentence refers to apply to more abstract, intellectual material. Metaphor is the source of virtually all abstract words in philosophy and science and is the main link between the concrete vocabulary (founded on the primitive words of the child or the community) and the total vocabulary of a language. The structuring of the primitive vocabulary is the basis for the structuring of the total vocabulary of concrete words and, at one remove via metaphor, for the structuring of the abstract lexicon. Disputes about the use of abstract terms are the expression of an attempt to relate them to the strong, natural structure of the concrete lexicon (based ultimately on the categories of human perception and action)

- the 'deep structure' of language is not linguistic but is in fact the structure of reality (as experienced in perception and action). Language can, however, only offer a mapping of human experience (perception or action), not a detailed picture - in effect a 'skeleton' of the particular perceived scene, proposed action or intellectual construct (in the same way as perception presents only a selection, a mapping, of the fullness of the content in any visual scene). 'Competence' and 'performance' are reinterpreted as the ability to comprehend and the ability to express - we can understand more than we can explain in language. The 'open-endedness' of language is not in any way a special property but an inevitable consequence of language's function in matching the open-endedness of human experience, of human perception and of human action.

- other languages, with different syntaxes and different lexicons (derived from originally genetic differences in the 'preferences' between possible forms of the structures of articulation, perception and action) arrive at equally valid representations of the world in words by different, equally natural routes. The extension of these hypotheses on the relation of language, perception and action to other languages is the starting-point for assuming that the hypotheses may have wider philosophical applications or implications.

The summary of the hypotheses and proposals advanced in the earlier chapters of the work is inevitably greatly compressed and may do no greater justice to the important details of the argument than the summary in the first part of this chapter does to conventional philosophical discussion of language and perception. However, in the light of the summaries furnished, one is now in a position to comment on the acceptability or relevance of the philosophical notions and approaches described for the hypotheses in this book and, equivalently, on the implications, such as they may be, of the hypotheses presented for the philosophical treatment of language, perception and knowledge. The discussion cannot possibly be a full one and the comments are framed as briefly as is consistent with their intelligibility. The comments on the philosophical summary are separated into those on the philosophy of language (or linguistic philosophy) on the one hand and those on the philosophy of perception (and knowledge) on the other.

Significance of language for philosophy

Philosophical discussion can only be conducted in language and this is the prime source of the importance of language for philosophy. Different schools of philosophy, if one accepts the hypothesis in this book that language derives its structure from the structures of perception and action and has as its function the representation, in outline, of the content of perception and action, have attached either too much or too little importance to language as an instrument to assist or obstruct the progress of philosophy. Language is validated by perception and action, not the other way round. Language is not 'the sole and essential point of contact for the philosopher with reality' - and the most pressing contacts with reality impose themselves without any intermediation of language. Language is not, on the view in this book, an abstract rational structure but one built on the most mundane of foundations, in human neurophysiological structure; it is subsidiary rather than primary in human behaviour and thought. Accordingly, because language is a secondary coding of experience, one should not expect much certainty or philosophical illumination from the analysis of language in isolation.

On the other hand, philosophical comment on the confusion and deception arising from language has been much overstated. Language originated and was designed, by evolution, for the accurate representation of the ordering of the concrete world of perception and action, and for this purpose it functions well, to the extent that it mirrors reliably the external world in which the human being has to act Between men with a shared approach to an area of knowledge or science, a shared practical objective or a shared skill, language has proved an indispensable and reliable instrument. Difficulties with language, as an instrument, have arisen rather from those who have attempted to make use of it for purposes for which it was not designed, primarily philosophers themselves and linguists. Amongst philosophers, verbal confusions and debates have proceeded from the use or invention of terms with no precise meaning and no clear relation to the central reliable core of the lexicon (founded in the original relation between words, perception and action). If philosophers choose to disregard the essentially social and natural foundations of words and organisation of words into sentences, to formulate their own idiosyncratic terms and syntactic procedures, then agreement and a shared clear view of what is true or at least probable becomes impossible or highly unlikely. Wittgenstein might have been more successful in struggling against the 'bewitchment of our intelligence by language' if he had not fabricated, with Russell, new unexplained terms (atomic facts, molecular propositions, logical space and so on) and if he had turned his attention away from language as a structure in its own right to human behaviour, the normal functioning of perception and action.

Even the terminology of 'ordinary language' philosophers was in its way obfuscating. Austin's perlocutionary, illocutionary and constative classification of utterances - and, of course, much earlier Kantian and Hegelian terminology- created a cloud of linguistic obscurity which no one can be said to have penetrated with confidence. Another source of confusion, due to philosophers rather than to language, has been the assumption, by Anglo-Saxons, French and Germans, that the special features of the syntax or structure of the lexicons of their native languages offer a direct insight into underlying universal truths. Whole philosophies in German and in English have been erected on distinctions of meaning or on differences in grammatical structure which are idiosyncratic to the particular language. Even on the ordinary language approach, the range of meaning of common words such as 'know' and 'believe' and their mode of use vary significantly from language to language. No absolute philosophical understanding can be arrived at by confining oneself to the particular structures of one's native language. What the hypotheses in this book contend about language is that every language certainly can achieve the same practical effect, in terms of the perception conveyed or the action indicated, but the combination of syntactic and semantic devices used to achieve the result may vary from language to language. To concentrate on one feature of syntax or vocabulary in a language is bound to result in distortion if one attempts to draw from it some general philosophical truth. A parallel kind of misdirection and confusion, but this time due to linguists rather than to philosophers, has been the over-estimation of the importance of differences between languages in syntax or lexicon, the theory associated with Sapir and Whorf that language determines perception, that people speaking languages largely differing in structure or in lexicon experience differently ordered worlds. Because in Hopi, for instance, time aspects are conveyed not by verbs but by other devices, Whorf sees the American Indian as living in a different spatio-temporal world from the European - but in reality the American Indian has to live in the world as it is, day after day, just as does the European, regardless of any differences in syntax. The real world imposes its structure on language, and language does not, cannot, impose some different, arbitrary structure on the real world.

At the other extreme, under-estimation of the usefulness of language in philosophy is as unhelpful as over-estimation of its significance. Max Black's assertion that "no road leads from grammar to metaphysics" and Popper's, that language is no more than a set of spectacles through which one looks at the world and one shouldn't waste one's time in cleaning the spectacles or talking about language, seem to go too far in the opposite direction. Language is part of human behaviour and it is a fact as much founded in neurology and physiology as any other part of the human organism. Whilst one must agree with Popper that philosophers who, on principle, disregard the content and progress of science can hardly claim to be philosophers in any true sense, the study of the functioning of language is itself a scientific study of great interest and importance. Certainly this is the case if one accepts the hypothesis that it is derived from, integrated with and structured by the neurological and physiological bases of perception and action. Even Popper needs language as a reliable instrument for expounding his own views on the methods of science and philosophy and, though he may be right to dismiss as fruitless some forms which linguistic philosophy has taken, this does not justify him in rejecting any attempt at a combined philosophical and scientific discussion of the issues which language presents. Whether, even in a narrow sense, Max Black's view on the philosophical insignificance of grammar is sustainable is doubtful. Whilst the syntax and lexicon of any single language may not point unequivocally to reliable philosophical views about the real world, insofar as languages are essentially isomorphic (that is they achieve by different coding methods the representation of the same content of perception, action or thought) and this isomorphism rests on the underlying relation of language, perception and action, a comparative examination of the different ways in which languages achieve the same results, in representing the real 'parts-of-vision' or 'parts-of-action' which correspond to traditional parts-of-speech, or achieve successful disambiguation by the combination of syntax and semantic structuring, could lead to useful general truths, or at least useful general probabilities in one's view of the real world.

Meaning as a central topic in the philosophy of language

Both in the summary account of the philosophy of language given earlier in this chapter and in the range of hypotheses presented in this book about the functioning and basis of language there has inevitably been reference in many different contexts to the concept of 'meaning'. Of all linguistic concepts the question 'What is meaning?' is the one which has been most studied and argued over by philosophers of language. 'Semantics' was originally conceived as a philosophical enquiry into meaning, though the range of application of semantics has now been greatly widened. The discussion about 'meaning' is a peculiarly apt demonstration of the self-reflexive property of language. If, it might be said, philosophers cannot agree to tell us what 'meaning' means, how can they give us reliable guidance on the meaning of any other philosophical concepts? The philosophical discussion of meaning has demonstrated perhaps more clearly than any other segment of philosophical activity both the difficulty and the confusion of the philosophical enterprise. 'Meaning' as a word has been analysed as having very many diverse meanings (one estimate was that there are over 50 distinct meanings of 'meaning') so that to the question 'What is meaning?', an almost unlimited range of answers has been possible and has been offered. Quine's comment already quoted that, until they have a satisfactory explanation of the notion of meaning, linguists in semantic fields are in the position of not knowing what they are talking about, could equally readily be applied to many philosophers. At the same time, if a satisfactory account can be constructed, by philosophers or by linguists, of the nature of 'meaning', it would carry with it the possibility of important advance in many other areas of linguistics and philosophy in so far as the techniques employed to explicate meaning could be used to tackle the relatively less complicated problems involved in understanding other philosophical terms and concepts.

Before considering how a variety of philosophers have tackled 'meaning' and how their views bear on or are affected by the references made to meaning earlier in this book, there are two preliminary clarifications. First of all, we have, at the moment, no basis for picking out any particular use of 'meaning' as the primary one; 'meaning' is not necessarily just a linguistic concept, the meaning of a word or of a sentence, but also we can, and do, speak equally readily about the meaning of an action, the meaning of something we see or perceive (e.g. the meaning of a traffic sign) or the meaning of a situation or an argument. If we assume that 'meaning' is necessarily and primitively linguistic, then we have already prejudged, and so limited, the range of theories of meaning which it is open to us to consider. The second preliminary clarification is one which has already emerged from the summary account given of linguistic approaches to philosophy in this chapter, that meaning comes before truth; the problem of meaning is separable and must, initially, be separated from the problems of truth and knowledge. Apart from 'analytic' truth (that is simply the explanation of one word or phrase in terms of other words or phrases), we must understand the meaning of a word or sentence before we can determine whether the sentence is true or false, whether the word applies or does not apply to a real object, event or situation; it is the world, not the words used or the structure of the sentence, which establishes the truth or falsity of the meaning which the sentence conveys. Truth and falsity are properly discussed, not in the context of meaning, but in that of the philosophy of perception and the philosophy of knowledge.

With these preliminary clarifications, it is possible to look briefly at and assess the relevance of discussion by philosophers of 'meaning'. Different philosophers have approached meaning from many different directions. Some have been concerned with distinguishing the range of uses of the word 'meaning'; of these, some confine themselves to linguistic meaning, others deal with wider uses of meaning. Of those concerned with linguistic meaning, some concentrate on the meaning of words (or concepts) in isolation and others believe that what philosophers should be concerned with is rather the meaning of sentences or propositions as the proper mode of conceptual analysis. Other philosophers approach the topic by seeking to categorise the proper questions to investigate about meaning: are meanings abstract entities or not? what conditions have to be satisfied for an expression to be meaningful? what is the nature of 'sameness' of meaning? why do word meanings alter? what are the smallest linguistic units which possess meaning? Or they seek to categorise the levels and diverse theories of meaning. One philosopher, for example, distinguishes three levels: the meaning of thoughts, the meaning of messages and the meaning of speech acts. Another philosopher usefully categorises theories of meaning as:

- mentalist: to be meaningful involves association with a mental item, an image, a thought or an idea

- behaviourist: meaning is what produces certain behavioural responses to utterances

- use theory: meaning is constituted by use

- verificationist: meaning is determined by how one decides the truth of an expression

- emotivist: meaning is the tendency of a word or phrase to produce a particular feeling.

This is by no means a complete list of theories of meaning and every theory has many variants. However, some of the comments already made apply immediately to the theories listed. If meaning is prior to truth, then one cannot accept a verificationist theory of meaning. The theory that meaning is constituted by use explains (almost certainly wrongly) how words acquire meaning but does not give a philosophical account of 'meaning' as such; ordinary language philosophy explores the meanings of words in use but does not necessarily require that the essence of 'meaning' should be constituted by use. Indeed Austin's concept of the speech act involved something much closer to a behaviourist view of meaning. The objections to a behaviourist theory of meaning are the familiar, and compelling, objections to behaviourist theories generally; behaviourism confuses essences with effects, structure with the expression of structure. An emotivist theory of meaning is open to very much the same kind of objection as a behaviourist one; meaning exists for the speaker as much as it does for the person spoken to, and to identify meaning with feeling is to confuse thought and feeling, information with the reaction to information. One variant of the behaviourist theory of meaning on which more specific comment may be useful is that put forward by Grice(19) (which has received a good deal of attention): his ideas have been presented sketchily but involve a distinction between 'utterer's meaning' and 'timeless meaning' of an expression (for an individual or for the language generally); the timeless meaning amounts to no more than the definition of the meaning of a word or group of words by the use of other words and takes one no further forward in determining what the nature of the 'meaning' of the explained or of the explaining words may be. Ziff's(20) criticism of Grice's approach, that it seems to be concerned rather with the use of an expression than with 'meaning' as such, seems justified. To talk about how meanings are used whether of a word or of a sentence, the effects they cause or the actions they perform, is not to provide a satisfactory philosophical account of the concept of 'meaning' as such, any more than to describe what a machine does, where one can travel to in a car, for example, explains how the machine or the car functions, what its essential structural properties are. Finally, in this categorisation of theories of meaning, a brief reference may be made to the original views of Carnap and Neurath(21) that 'meaning' should be defined purely in syntactical terms, as referring to the properties of sentences in a formal system. If one accepts that words in isolation can have meaning, as well as sentences, then such a theory is unacceptable. Carnap himself later abandoned it and his revised view was that the problem of meaning is a semantic and not a syntactic issue.

If, as the comments in the previous paragraph have done, one rejects verificationist, behaviourist, emotivist and use theories of meaning and the other variant theories described, the only theory left on the list and not so far explained or criticised is the mentalist theory of meaning, that to be meaningful involves association with a mental item, an image, a thought or an idea. This is really not one theory but a category or collection of theories. Quite clearly, there is an intimate relation between the meaning of words and mental processes; some would say that mental processes are the manipulation of words (though there are reasons for thinking such a view mistaken) but at least words emerge from mental processes and express mental processes. But to accept the close link between word- or sentence-meanings and mental processes does not necessarily take one much further in giving a philosophical account of meaning. The concepts of 'idea', 'thought', 'mental image', are themselves as philosophically obscure and unexplained as the concept of 'meaning'. To treat 'meaning' as in some way an operation or relation of 'mind' leads one into the swamp of philosophical debate about mind and leaves one as puzzled as before about the most suitable philosophical approach to 'meaning'.

The outcome of this survey of philosophical consideration of 'meaning' is at the same time disappointing and paradoxical. Disappointing because none of the various theories offers a helpful account of meaning or is immune from straightforward and powerful destructive criticism. Paradoxical because, despite the failure of philosophical discussion, everyone, philosophers, linguists and ordinary people, uses the concept of 'meaning' without the slightest difficulty; we all know intuitively what the words we use mean (except perhaps words which are the artificial constructs of philosophers), we all understand what it is to grasp the 'meaning' of a sentence, to see what the development of some events or situation 'mean', and we all readily say: "I see what you mean" "You have understood my meaning" or "That means nothing to me". The words 'mean' and 'meaning' are some of the most familiar and frequently used by ordinary people, by philosophers and linguists. Surely there must be some way out of the labyrinth, some way of convincing ourselves that there is nothing mysterious or insoluble about 'meaning' as a central concept in philosophy, and in everyday life.

One possibility is that the problem of meaning has been approached from the wrong direction; philosophy, as a discipline inevitably founded on language and on the manipulation of the meaning of words and sentences and which can only offer explanations in terms of the meanings of words and sentences, is in no better position to explain one of its primitive concepts, 'meaning', than it is to explain, in philosophical terms, the nature of the colour 'red'. 'Meaning' was a property of perception and action long before it became a property of words or sentences; a theory of meaning is much more probably to be found through methods used to understand the functioning of perception and action than through the verbal manipulations of philosophy itself. If we can come to understand the nature of the 'meaning' of what we perceive, in a visual scene, or the 'meaning' of an action, then we may be able to extend this to understand the meaning of a sentence or word. On the hypotheses presented in this book, the meaning of a word (for a percept forming part of the primitive visual repertoire of a child) is constituted by a direct neurological and physiological link between the sound-structure of the word and the shape or identifying physical characteristics of the perceived object to which the word refers; in parallel to the neural patterning constituting the schema of a visual percept is the neural patterning constituting the sound-structure of the word. The coincidence of patterning between word and percept is the essence of the property 'meaning'. Similarly the meaning of a sentence, the meaning of a visual scene and the meaning of a plan of action, are constituted by the simultaneous neural patterning from which the expression of the integral meaning in the serial form of a sentence and the serial execution of the complex action are derived.

This account of 'meaning' is not a philosophical or a linguistic one but essentially a neurological, physiological one. It coincides, somewhat surprisingly, quite closely with the account of the origin and nature of the meaning relation given by Sayre in his book Cybernetics and the Philosophy of Mind. For him, words acquire meaning by being intentionally associated (by the individual or as a result of the shared intention of a community) with objects; an intention is a purpose operating consciously in control of an organism's behaviour, a neuronal configuration regulating behaviour through sequential stages and terminating in a particular goal. Language arises as a means of sharing intentions: "The meaning of a term is the shared intention with which the term is associated in its common use. On the one hand, a meaning is an intention, and hence a purpose, and hence a neural configuration ... On the other hand, a meaning is a formal structure that can be instantiated in the cortexes of different individuals, and transmitted to others within the group".(22)

Thus, a meaning is what makes a difference to the neurophysiological organisation of the individual acquiring the meaning or expressing the meaning. Before expression in a particular sentence uttered in particular circumstances, a 'meaning' is a generalised pattern relating a sound-structure and a generalised visual or other schema; all words are universals (as all perceptual schemas are universals) until they are given a particular application, a particular reference by having their range of possible meanings narrowed down by the combined action of the syntactic and semantic structures in the sentence. At the same time, as a neural pattern, a 'meaning' forms part of the total complex of meanings (the total set of neural patterns) which go to form our conceptual system: in its use in a particular sentence, a word and its meaning bring with them, initially, their relationship to other words, and other meanings, the structures within which the individual word, the individual meaning and the individual percept are integrated. The process of determining the 'meaning' of a sentence is, operationally, the use of semantic and syntactic clues, derived from the serial string of words, to determine the particular meaning and role of the word, in that sentence on that occasion.

Perception as the foundation of knowledge of the real world

The summary account of the philosophy of knowledge earlier in the chapter describes the stages by which traditional philosophy had reached an impasse, being unable to give any satisfactory account of how perception (in the view of the empiricists the only sure foundation for knowledge) in fact enables us to know anything with certainty about the existence of objects in the real world (or even whether a real external world exists) or about the reality of the sequences of causes and effects which, in common understanding and in scientific theory, we rely upon to explain our experience of the world and the processes of physical nature. The idealism of Berkeley and Kant and the scepticism of Hume, not so far successfully challenged or refuted by orthodox philosophy, left philosophy in the position that we could know with certainty no more than our own ideas and the forms of our understanding (the Kantian categories of space, time, etc.) whilst reality and true knowledge would for ever remain beyond our reach (if indeed anything existed beyond the circle of our own ideas) including, according to Kant, any knowledge of the sources of our modes of understanding or of the functioning of our mental mechanisms. Attempts to develop a separate philosophical treatment of perception in terms of sense-data have not been convincing or provided a way out of the dilemma; naive realism and pragmatism react simply by disregarding the philosophical arguments of idealism and scepticism which they cannot refute. The practical common-sense philosopher, like Ryle, opines that theory about perception in philosophy can proceed perfectly well without any reference to the investigation of perception as a physiological or psychological process.

It now remains to relate the hypotheses about the functioning of perception (in relation to language and action) presented in this book to the situation in the philosophy of knowledge and perception and to the suggested routes from the philosophical impasse suggested by Whitehead, Merleau-Ponty and Lorenz. The theory advanced in this book has been that perception developed from action and to serve action; for perception to be effective, it has to be closely integrated, physiologically and neurologically, with the organisation of action. Visual perception is organised hierarchically, built up from elementary visual units (on the genetic basis provided by 'imprinting', that is, the predisposition to adapt the neural structures to the visual patterning found by the infant in the environment in which it finds itself at birth; the elementary units are then formed into visual shapes and objects (which are stored, in parallel with the words referring to them, as motor patterns, schemata, in the cortex). The comprehension of the visual scene, the unity of the act of perception, derives from the extraction from the visual stimuli affecting the retina of the familiar schemata and their relationship to form the coherent visual scene. On this view, all the structuring of perception of the visual scene is derived or extracted from the stream of light energy, the real physical patterning of the stimuli (accepting as sound the views of the well-known authority on visual perception, James Gibson(23)) but the visual apparatus has been adapted, partly genetically through its anatomical structure and neurological basis, partly through the operation of the 'imprinting' process described and partly through its ability to store visual schemata for recurrent patterns of visual stimulation, so that the functioning of the apparatus reliably and naturally is correlated with the visual information derivable from the real world of external objects.

This view of the way in which perception provides us with reliable knowledge of the world seems fully compatible with the accounts given by Whitehead, Merleau-Ponty and Lorenz: namely that perception is reliable because it constitutes experience from within the system of nature by an element in that system (the human being) of his relations with the whole of nature of which he is part (Whitehead); that perception should be given back its primary importance in philosophy; that the reliability of perception derives from the human being as the 'body-subject', from the mixture of the world and ourselves which precedes all reflection (Merleau-Ponty); that the perceiving subject and the perceived object are equally real, all knowledge derives from the interaction between them and the reliability of our perception is the result of the evolution of the human cognitive apparatus to adapt to the equally real things in the external world (Lorenz). For the human race, the Kantian forms of understanding have, as a result of evolutionary development, become embodied in our neurological and physiological structures. For Kant, human understanding 'prescribes its laws to nature' - but evolution, developing nature, had much earlier in biological history 'prescribed' to the human mind its categories and functioning, to accord with the real world.

The integration: of perception, action and language, resulting in the underlying isomorphism of all languages, combined with the true knowledge of the external world which the evolution of the cognitive and visual apparatus has made possible, opens the way to a new pursuit of philosophical truth through language. We need no longer distrust our own reasoning or our belief in the reality of causation in the external world. The intellectual development of mankind can proceed, as it is doing, but on a philosophically more secure basis and in the knowledge that language, as a flexible instrument designed to match the open-endedness of human experience (perception and action), can be a reliable medium for exploring, recording and developing man's knowledge of the external world and of his own nature.