Starting from the analysis in Chapter II of the biological and evolutionary relation between action, perception and language, Chapter III has proposed that there must be, at the most elementary level, a systematic relation between the unitary features of vision, action and speech, from which more complex forms of perception, action and language are constructed. This Chapter carries the process a stage further. It is concerned with two distinct but closely related matters, which bear on and explain the functioning of individual words in natural languages. The first contention is that the collection of words initially acquired by a child, its primitive or core vocabulary, is built very obviously and necessarily from those words which refer most directly to the features and speech with which the child is brought most regularly into contact; the child's first world is a simple one and the words it acquires are those with the most direct, experiential reference. A child's first vocabulary assumes the form it does and is structured directly by the fundamental properties of the child's need to perceive, act and speak. Part I of the Chapter is therefore concerned with the formation of this Primitive Vocabulary (which goes to match the child's Primitive Repertoire of Perception and Action); the significance of the primitive vocabulary is not only as a stage in a child's development but, in its content and structuring, as the core or framework on which year by year the child extends its vocabulary to cover a wider range of experience and ultimately arrive at the full vocabulary, the total lexicon used by the adult. The proposition is thus that, in any language, the full lexicon derives its internal structure from the representation in language of the basic human behavioural abilities of action, perception and speech, and that extension and refinement of the vocabulary, to deal with abstract matter, thought processes, the creative use of language, can ultimately be traced back to the basic structure, which will have been extended and deepened by a variety of processes, amongst which perhaps the most vital is that described as metaphorical transformation or extension, the application of structures initially derived from simple situations of experience to similarly structured but much more remote operations of thought and language.

Accordingly, the first part of the Chapter deals with one way in which language is structured by perception and action, in terms of semantic organisation. The second part of the Chapter, which follows directly on from the material in Chapter III, explains how at the micro-level, so to say, the particular forms of words used to refer to particular percepts or actions derive their structuring from the natural basis provided by the equivalences of the elementary units of vision, action and speech, set out in Chapter III. Individual words are thus natural not arbitrary because their structures carry within them either a direct representation of a percept or action or an indirect clue or indication of the percept or action to which the word relates. The natural basis of words in a language is to be looked for in a natural relation between the articulation of the speech-sounds forming the word (essentially a sequence of fine movements of the articulatory organs), the movements of the body or body-parts, particularly the arm (again essentially a sequence of precisely-patterned movement of the muscles determining bodily position and movement) and patterns of activity of the eye in perceiving visual contours and movement (determined once again by precisely controlled movements of the muscles of the eye, the ciliary muscles controlling accommodation of the eye and the muscles regulating movement of the head as part of the process of directing the gaze towards the perceived object).

The two parts of the Chapter, the first dealing with the primitive vocabulary and structure of the lexicon and the second with the natural basis of the structures of individual words, are not unrelated. There are two important links between them: both demonstrate the fundamental importance of the hypothesis of the integration in behaviour of action, perception and speech; secondly, the words identified as plausibly forming part of a child's primitive vocabulary are used for selecting examples of individual words in Part II for which the natural basis of the relation between word-form and word-meaning, in visual contour, action-pattern or body-movement, is presented.

The arbitrariness of words and the natural structure of the lexicon

Chapter I has discussed at length the traditional view that individual words are arbitrary and has argued that no part of language in fact can be wholly arbitrary, neither the forms of words nor the syntactic and grammatical structures by which words are joined together into phrases and sentences. But the difference of words between different languages has always been the root reason why linguists and others have virtually without a moment's hesitation taken it that language must be arbitrary. John Locke (1706 [1964]} in his Essay Concerning Human Understanding wrestled vigorously and extensively with the problems presented by the arbitrary nature of words. For him, words stood for nothing but the ideas in the mind of him that uses them ; but there are no innate ideas. 'To ask at what time a man has first any ideas is to ask when he begins to perceive, having ideas and perception being the same thing.. Perception derives from sensation: When a man begins to have any ideas, is when he first has any sensation, which is such an impression or motion made in some part of the body as produces some perception in the understanding ... When the understanding is once stored with these simple ideas, it has the power to repeat, compare and unite them, even to an almost infinite variety, and so can make at pleasure new complex ideas ... It may also lead us a little towards the original of all our notions and knowledge, if we remark how great a dependence our words have on common sensible ideas, and how those which are made use of have their rise from thence, and from obvious sensible ideas are transferred to more abstruse significations, and made to stand for ideas that come not under the cognisance of our senses e.g. to imagine, apprehend, comprehend, adhere, conceive, instil, disgust, disturbance, tranquillity &c are all words taken from the operations of sensible things and applied to certain modes of thinking".(1)

Whatever view one may take of Locke's dismissal of the possibility of innate ideas or of his assertion that "since sounds have no natural connection with our ideas, but have all their significance from the arbitrary imposition of men ... a man may use what words he pleases to signify his own ideas to himself", there is a great deal in the passage quoted above which seems founded on common sense and is fully compatible with the thesis advanced in this Part of the present Chapter. Locke himself came, somewhat to his own surprise, to a recognition of the significance of words for his philosophical enquiry: "When I first began this discourse of the understanding, and a good while after, I had not the least thought that any consideration of words was at all necessary to it" but he then found himself led into the still interesting and useful discussion of language contained in Book III.

Both in Locke and in the later extensive studies of Bishop Wilkins (1668 [1968]}, one can note some tinge of regret that words should be arbitrary and that there should be no natural relation between the structure of the world and the structure of language. Locke notes that there cannot be any natural connexion between particular articulate sounds and certain ideas, "for then there would be but one language amongst all men" and Wilkins (in his Essay Towards a Real Character and a Philosophical Language) comments rather sadly: "It were exceedingly desirable that the names of things might consist of such sounds as should bear in them some analogy to their natures; and the figure and character of their names should bear some proper resemblance to those sounds that men might easily guess at the sense or meaning of any name or word, upon the first hearing or sight of it. But how this can be done in all the particular species of things I understand not, and therefore shall take it for granted that this character must be by institution".(2)

A more recent, interesting, attempt to express how language and perception might be naturally related is to be found in Uhlan von Slagle's monograph Language, Thought and Perception (1974). He combines an approach derived from Kant and Humboldt with insights from gestalt psychology and proposes a direct relation of the structure of meaning to the structure of experienced reality. Language, he says, is "man's tool for adequately orienting himself in extra-linguistic reality, and one must look for a systematic correlation between the structure of thought and the structure of experience and reality...The drive is to overcome the dichotomy of perception and thought" and this he found in the contention that the immanent organisational factors of sensory fields are the immanent organisational factors of thought; semantic form classes correlate with perceptual forms; concepts are 'rules' of perceptual organisation, or rules of 'categorisation' of perception. The structure of information stored in the mind can be described in terms of the structure of the perception itself. The structure of sensory experience can be correlated with the structure of linguistic meaning and that of thought. The position he takes on the correlation of perception and conceptualisation "parallels the growing recognition by linguists of the interdependence of syntax and semantics in natural language. In each case, there is the underlying assumption that form and content are interdependent. "Linguistic patterning rests ultimately on extralinguistic patterning".(3)

The attempt in the first part of the present chapter, in the presentation and development of the concept of the Primitive Vocabulary, is to go some way towards meeting the objectives of Locke, Wilkins and Slagle, to demonstrate a natural relation between the structures of perception and action and the structure of the vocabulary of any language (on the argument that the full lexicon of a language is a structure built up round words relating to a concretely-based core of experience). In modern terms, the subject-matter of this part of the Chapter is essentially semantics, semantic fields, semantic structuring, the relation between the semantic groupings of words and the experiential grouping of perception and action.

Traditionally, the field of linguistics has been divided into phonetics, syntax and semantics - and semantics has tended to be the most neglected of the three , the section where the problems are greatest and the solutions least convincing. Often, linguists have concentrated on phonetics or syntax and ignored semantics. In any case, it has been very unusual among linguists to attempt to find or demonstrate any real operational relation between these domains of language study, even though, in practical experience, in speech 'performance' (as Chomsky would describe it) the three aspects of language are inextricably linked and mutually definitive and supportive. Linguists have tended to be more interested in syntax (or phonetics) and philosophers more interested (though at a very refined level) in semantics than in syntax. For many years, in this century, linguists were induced to believe (following influential authors such as Bloomfield and Sapir and the behaviourist psychologists) that 'meaning' (semantics) was none of their proper business and indeed hardly a respectable subject for study. This anti-semantic approach was initially at any rate demonstrated by Chomsky and his followers, even though gradually and covertly as time went on, it became necessary to introduce a larger semantic element (in the form of selection restrictions) into syntactic structures, to make more explicit allowance for the mutual interaction of syntax and semantics.

Outside the strict field of linguistics, rather more attention has been paid to semantic issues. Some of the most interesting work is in progress as part of natural language understanding programs in the field of artificial intelligence, the work of Roger Schank, Terry Winograd, Bobrow, Yorick Wilks and others and some reference to this will be made later. Apart from this, perhaps most relevant for this chapter has been activity of the general semanticists, the studies of semantic fields associated with Weisgerber (1962)and Trier (1931)(4), and discussed at some length(5) in Miller and Johnson-Laird's Language and Perception (1976)and John Lyons' treatise Semantics (1977)(6). Nevertheless, even the experts in the field would admit that the state of theoretical discussion of semantics is far from satisfactory and so far there has not been a great deal of practical value for understanding the real functioning of language which can be derived from it. As has been remarked by an expert in the field: "No one has yet presented even the outlines of a satisfactory comprehensive theory of semantics"(7).

Despite this rather discouraging situation, one can perhaps extract from the literature some probably valid and valuable points:

1. the general semanticists put as their chief principle that language and reality are related in the same way as a map is to the terrain to which it refers. Language therefore is not an image of reality; at best it is an image of the structure of reality. The second principle is that of incompleteness; the representations in language are always less than what is represented; the map inevitably ignores details of the terrain. The third principle is that of the self-reflexiveness of language; we use language to speak about language; we make judgments about judgments; we evaluate values. This process of abstraction can be removed in varying degrees from the level of the concrete event; multiordinal words have different meanings on different levels of abstraction.

2. the lexical items of a language form a coherent pattern. They are not as rigidly organised as rules of grammar but sufficiently patterned to enable each item to have a definable relationship with all the other items. Weisgerber and Trier would argue in this way that lexical units should not be viewed in isolation; instead they form a closely knit and articulated lexical sphere, where the significance of each unit is determined by its neighbours, their semantic areas reciprocally limiting one another, the relation of lexical units being crudely similar to that of individual atoms in a complex molecule, where the displacement of one atom will to a greater or less degree affect the nature of the entire complex. It should be noted that there are imprecisions in the formulations of Weisgerber and Trier and what they say, in detail, is open to attack at a number of points but the principal emphasis of their work, the stress on the systematic structuring of the lexicon, and particularly of defined areas of the lexicon, has been stimulating and led to much more precise examination of the mutual interaction of meaning-elements.

3. Semantic structuring may fundamentally be innate with the ultimate terms of a semantic description such presumably biologically given notions as identity, time, space, body, movement, territory, life, &c; there may be a universal set of 'presumably innate' concepts which are expressed in all languages by affixation, suppletion, enclitic particles or word order. There may be an alternative possibility that the lexicon is organised with regard to the representational derivation of its entries: that is, the degree to which their meaning can be specified by sensorimotor images and contextual memories and associations; form-class semantic may leave its traces in the nervous system, facilitating thought in some directions, inhibiting thought in other directions.

This harvest of ideas on semantics drawn from a number of sources is perhaps not a very rich one. The general conclusion remains from this brief survey that as a field of study semantics is still far from well-developed and despite the volume of philosophical discussion of meaning and more recently of the mode by which semantics might be fruitfully incorporated in syntactic models of language, there is still rather little of solid value. One has to agree with Chafe(7) that semantic structure is the crucial component of language, so that modern linguistics has been like trying to describe milk-production while omitting the cow, and that much of the linguistic effort within semantics appears to have little or no relevance for a neurobiological study of language. Linguistics in fact has learned least about semantics.

Given the relative barrenness of academic discussion so far of semantics and semantic structuring of the lexicon, this Chapter experimentally adopts a quite different approach (which can however at some points be related to what has already been said). The approach is centred round the concept of the Primitive Vocabulary. The idea of the Primitive Vocabulary is that all language is built round a limited number of primitive words (used for referring to a limited number of primitive percepts, primitive actions and states, external or internal). The relation between the particular words in the primitive vocabulary and the percepts &c to which they refer can be established only by actual experience. The relation cannot be established by the mediation of other words, so that the Primitive Vocabulary represents the categories into which our 'naive' experience is divided, without reflection, without the discursive use of language and without second-hand communication of knowledge. The Primitive Vocabulary thus consists of the words which no one else can explain to us if we have not ourselves had experience of the percepts &c to which they refer; and this means that we cannot, without the actual experience, recognise a particular instance of a percept as belonging to the category named by a specific word. The initial set of words learned by a child is the prime example of what is described as the primitive vocabulary, that is it consists of words learned from the confrontation of the child with the object, the action or the event (normally accompanied by a direct or indirect naming of the object, action or event). The acquisition of words in the Primitive Vocabulary normally goes in step with, though occasionally follows, the formation of what could, in parallel, be described as the child's primitive perceptual repertoire, that is, those shapes, objects, things and relations that a child identifies first as its visual apparatus and experience develop.

The assumption is that the Primitive Vocabulary (and the associated primitive perceptual repertoire) are structured, that is organised, in terms of the sensory modalities involved, in terms of the frames of experience in which the objects or actions are normally encountered and in terms of the distinct physical frameworks to which particular objects or actions are directly related. This Primitive Vocabulary is set in contrast to the full lexicon acquired by the adult; the primitive perceptual repertoire is similarly contrasted with the much more extensive repertoire of perceptual and other experience acquired by the adult. To adopt one of the ideas of the general semanticists referred to earlier in the Chapter, the suggestion is that the Primitive Vocabulary maps on to the primitive perceptual repertoire in the same way as the full lexicon of a language (or of an individual adult) maps on to the complete perceptual repertoire inherent in the language or actually possessed by the individual adult.

Perhaps the most expeditious way of indicating what is meant by the Primitive Vocabulary is to present an illustrative list, an example of a Primitive Vocabulary. This is done on page 61 [moved to top of document] but before presenting the list, there are a few preliminary points which can usefully be made:

a. The idea of the Primitive Vocabulary is that it contains words whose meaning and understanding can only come from the direct confrontation so to say between the word as a patterned speech-sound and the specific experience (visual or action) to which it relates in the language used in the particular community, that is: we could not understand what in experience a flower might be without seeing a flower and being told that it was a flower; we could not understand what RED as a word means without seeing something red and being told that it is 'red';

b. There is no such thing as the absolute Primitive Vocabulary, in the sense that there is a defined minimum of words matching a defined minimum of items of perceptual experience but there is probably a near enough core primitive vocabulary which will be shared by all normal children and adults - body parts, body movements, colours, visual qualities. For a particular child (or adult), what in fact is contained in the Primitive Vocabulary depends firstly on the core items, secondly on the community environment in early years and thirdly on the particular family or social situation of the individual child in its early years: a Duke's child may well have a considerably different Primitive Vocabulary (and different repertoire of perceptual experience) from a dustman's child; the child of an Indian peasant will have a different Primitive Vocabulary (and primitive repertoire) from the child of a New York stockbroker and so on;

c. There may well also be, at first at any rate, a significant difference between the contents of a child's Primitive Vocabulary and the contents of its primitive perceptual repertoire; an infant will perceive and recognise a straight line, an angle or a circle but it does not follow that its parents will have told the child that these percepts are called LINE ANGLE or CIRCLE. An infant may also have inner perceptual experience, feel, think, know, without learning or being told that these experiences are named as FEEL THINK and KNOW. These gaps in the matching between primitive words and primitive percepts will normally be filled in later on and of course there can equally well be reverse gaps where an infant hears a word used without the matching percept to which it refers being presented to him;

d. Listing of the various words shown in the illustrative Primitive Vocabulary does not of course mean that all words listed are acquired together, though some words may be learned in groups. For each infant, there will be an individual order of acquisition of particular primitive words, though certain uniformities are probable in the order of word acquisition, particularly for functional words (as Roger Brown's observations suggest). For concrete words, the order of acquisition will depend on the order of actual experience of the objects referred to, which may differ between children and may be spread over a shorter or longer period;

e. Beyond what is presented here as an illustrative Primitive Vocabulary, there is a very much larger number (perhaps many times larger) of other primitive words (matching other items of primitive perceptual experience) where the understanding of the meaning of the word depends on conjoined experience of the word and the object or event. There is a mass of other words which cannot be understood mediately, that is by definition in terms of already familiar words;

f. One final obvious point: for individuals belonging to different cultures and climates, there will be words which are never found as part of their primitive vocabularies. ICE will certainly' not be in the Primitive Vocabulary of someone living on the equator and MONSOON will not be in the primitive vocabulary of a child brought up in England. Correspondingly, there may be a range of words in the Primitive Vocabulary of a child in one culture corresponding to a single primitive word in the vocabulary of a child in a remote culture: the familiar examples are the number of different words said to exist for different types of snow found among the Eskimos and the range of different words for animals, e.g. for camels found among the Arabs. The words forming a particular Primitive Vocabulary are not confined to common objects or even common percepts but extend over the whole of an individual's immediate, first-hand experience; they represent the irreducible elements in his linguistic mapping of his world.

The list described as the Primitive Vocabulary is, as has already been noted, only one among many possible primitive vocabularies though the proposition is that some primitive list of this kind is the starting point for every individual's acquisition of language. The presentation of a specific illustrative list of this kind, however, allows one to consider a number of points, to raise a number of questions:
1. Why should the particular words included in the list figure in it?
2. How is this concept of a primitive vocabulary related to other uses of semantic primitives, in a variety of fields of study?
3. What kind of internal structuring does a list of this kind display both in terms of meanings and in terms of the grammatical classifications of the words included in it?
4. How is a list of this kind related to the development of a child's repertoire of perceptual experience (and experience of action)?
5. By what processes can the field of application of words included in the list be extended?
6. More generally, what is the relation between a Primitive Vocabulary ('core' vocabulary) of this kind and the full vocabulary acquired by an adult, the full lexicon of a particular language?
7. What significance should be attached to the scope for the metaphorical use of words included in the Primitive Vocabulary or of other words subsequently added to form the adult lexicon?
8.In what sense can one treat words included in a Primitive Vocabulary of this kind as natural, that is as having a non-arbitrary relation between the sound-structure of the word and its meaning?
Selection of the five hundred or so words which figure in the illustrative Primitive Vocabulary has been guided by a number of principles. First of all, the essential qualification for inclusion of a word in the list is that it should not be possible for the meaning of the word to be understood by definition in terms of other words; though some broad indication of the meaning of one of the words may be given by using more evolved parts of an individual's vocabulary, the way in which a child learns the meaning of 'saucer', 'sit', 'finger' is by actual experience. Locke commented on the impossibility of indicating to a blind man the sensation associated with the word 'scarlet'; the nearest one might come to it, he thought, might be 'something like the sound of a trumpet', but this would still be far from understanding the concrete meaning in experience of 'scarlet'. Not all the words included in the Primitive Vocabulary are ones for which an object can be indicated by pointing, handling, hearing or smelling; some words are understood as associated with a particular facet of experience: 'bad' or a particular use of language: 'What?' No parent explains to a child 'This is what 'what' means'; or 'bad' is defined in the following way. The learning of words such as these is as much a part of the child's lived experience as its learning of the meaning of 'cat', 'chair', 'black' and so on.

There is nothing particularly new, of course, about an attempt to identify a restricted, primitive list of words, for linguistic research or other purposes. In the investigation of unknown or little known languages, it is commonplace for the field investigator to make use of a restricted basic vocabulary. For example, the interesting work done by members of the Summer Institute of Linguistics into the genetic relationship of groups of American Indian languages Comparative Studies in Amerindian Languages (ed. Esther Matteson et al 1972(10)) was based on restricted word-lists which, for the different languages, ranged between 100 and 300 words. Morris Swadesh (1972)(11), for work in support of glottochronological theory, used a basic vocabulary list containing 100 words. The constructors of Basic English(12) confined themselves to 850 words (including many words for commercial applications which would not be suitable for a primitive list of the kind with which the discussion in this chapter is concerned). In the field of artificial intelligence, there has been intense discussion of the extent to which knowledge structures can be built up from extremely limited lists of semantic primitives. Yorick Wilks (1977)(13)proposed that some 100 semantic primitives would suffice; others, e.g. Roger Schank (1972)(14) have made do with very much smaller numbers. An illustrative Primitive Vocabulary containing nearly 500 words is thus of a respectable size and may serve as a basis for useful discussion of some of the issues which have been noted above.

The items included in a primitive vocabulary of this kind fall naturally into a number of distinct categories; the list can be structured in various ways. Most obviously, the words are divided between those referring to shapes and things, movements and actions and qualities and relations. More broadly, one might say that the items in the list divide into those which are specifiable in spatial terms and those which are specifiable in terms of time, and that relational words, as well as words relating to objects or movements, can be referred primarily to space or time relations. Another basis of structure of items in the list is the sensory modality with which a word is primarily associated: things seen, heard, felt, tasted, smelled, felt internally, and so on. Another basis of arrangement of items in the list is in terms of the specific physical structures of which the items normally form part or in which they are experienced, for example, parts of the body are naturally grouped together as are movements of those body-parts, objects found in a room or a house can be grouped together as can be objects in the external landscape, whether the smaller-scale landscape of street and neighbourhood or the larger landscape of field, wood, lake, sea and sky. Objects can be classified in terms of their size or use as instruments, as parts of sequences of activity or types of occupation, furniture, clothing, utensils. Against this wide variety of possibilities of classification of the items to which the words in the Primitive Vocabulary relate one can set the traditional groupings of grammar, the analysis into parts of speech, nouns, verbs, adjectives, prepositions, adverbs, pronouns; the traditional classification has a rather arid and academic air compared with the groupings and structuring of words in the Primitive Vocabulary given by experience. Though MOUTH may be classified as noun, DRINK as verb and SWEET as adjective, the real experiential grouping of the items is as part of one process for the child, for example, drinking a glass of milk or a cup of tea, and this sort of analysis is closer to that suggested by Fillmore(15) in his case grammar, to the concept of selection restrictions used in transformational grammar and to the frame approach suggested by Minsky(16) in artificial intelligence and Roger Schank's (1980) approach in terms of 'scripts', that is extended descriptions of the kinds of setting in which particular words are likely to figure. This issue, of the way in which the lexicon is naturally and operationally structured, as an aspect of the process of speech comprehension, is one to which it will be necessary to return later. It is of particular importance because the suggestion will be that the express or implicit structuring of items in the Primitive Vocabulary (and of the parallel items in the primitive repertoire of perceptual experience) is the foundation for the organic structuring of the total lexicon of any individual adult in the extreme, of the total lexicon of any given language. What is clear however at this stage is that the traditional analysis into parts of speech has only a rather distant relation to the natural structuring which the set of words in the Primitive Vocabulary displays; in English, the same word may readily serve as verb, noun or adjective; adverbs and prepositions are often only with difficulty distinguishable; parts of speech normally associated with distinctions of space are often used, with little complication, to refer to distinctions of time. Consideration of the practically effective structuring either of the primitive vocabulary or of the full lexicon will require a pretty radical reconsideration of the formal or real significance of traditional parts-of-speech categorisation.

Not much perhaps needs to be said about the relation between a child's primitive vocabulary and the development of the child's repertoire of perceptual experience, which has been the subject of such close study by Piaget and his assistants. The effective acquisition of a word by a child must proceed in step with the maturing of its sensorimotor capabilities and the meaning attached to a particular word may broaden or narrow as the child's experience of perception and action extends. Individual words may receive a wider extension: 'mother' (initially one person in a particular relation and to some extent itself a generalisation from the separate occasions on which the phenomena of 'Mother' manifest themselves) may gradually become a word referring' to 'mothers in general'; the mother of that other child as well as the mother of the individual child; there is a quite natural extension of a word from referring to one particular percept to referring to a class of similar percepts. 'Chair' may initially be one chair and later on 'any chair'. Words included in the Primitive Vocabulary are thus not all necessarily at the same level of concreteness or abstractness, particularity or generality. Apart from the implicit generalisation in each individual word, there may be words which are already classes of classes; e.g. 'colour' may refer to any of red, green &c and indeed the child may in fact acquire the word 'colour' as 'a set of colours', crayons, or a colouring book. Nor will the acquisition of words necessarily match that of the theoretical class-order. 'Clothes' as in bed-clothes or the child's own clothes may be a functioning item of experience, a usable word, before particular types of clothes, whether these are 'sheet', 'blanket', 'dress' and so on. The word 'fruit' and the experience 'fruit' may come before the acquisition of words for individual fruits. The schemes proposed in more abstract approaches to semantic features, the hierarchical ordering of words and concepts, may simply be irrelevant for the mode in which in fact a child, or an adult acquires the elements in his Primitive Vocabulary. The reality is that every word is an abstraction from the multiplicity and individuality of perceptual and action experience and acquisition of particular words may proceed, in temporal order, either from the more specific to the more general or from the more general to the more specific (as the entomologist, for example, proceeds from 'ant' and 'bee' in his primitive child's vocabulary to the hundreds of thousands of specifically distinguished ants and bees recognised in his professional career. Theoretical semantics so far has generally disregarded this vitally important aspect of the growth of the individual's semantic structure, his knowledge structure.

Some indication has already been given of how words forming part of an individual's primitive vocabulary can gradually and naturally acquire a wider range of application, as the individual's experience itself develops. Most simply, apart from extending the word 'chair' to apply to any chair rather than just one particular chair (that is to any item sharing some broad structural features with what has first been identified as 'the chair') there can be a further very obvious analogous extension of the application of a primitive word. For example, 'leg', originally perceived as part of the child's own body, is extended first to apply to 'leg of any human', then to 'leg of any animal' (in many respects resembling the human body) and then, more adventurously) to 'leg of any object' which has an even looser structural resemblance to the human or animal body; so one has 'leg of a table' and then to any appendage attached to something broadly conceivable as a unity, a body, 'leg of an electricity pylon' or even 'leg of a race'. These are all processes of generalisation and abstraction, noting once again that even the original percept 'leg' itself was a generalisation from repeated, separate perceptions of a persisting leg, itself a construct from the visual elements forming each individual perception of the leg. At some point, 'leg' receives an even bolder extension, a metaphorical transfer where the word is applied to refer to non-external, perceptual experience, such as the 'leg of an argument'. This final, metaphorical transfer between external perceptual experience and abstract, remote, internal perceptual experience, is one of the major, absolutely fundamental processes in the development of language, both lexically and syntactically. A good deal more will have to be said about its significance later. It is one of the major ways in which a limited initial lexicon can be adapted to map on to a progressively greater and greater area of human experience. But beyond these fairly straightforward ways in which individual words in the Primitive Vocabulary may acquire a wider area of application, there are, of course, important, well-known ways in which words can be produced by composition, derivation, the use of prefixes and suffixes, which can operate very substantially to enlarge the area of experience covered by words initially in the Primitive Vocabulary and beyond that go to the construction of the fuller lexicon in due course acquired by the child and the adult.

The question for consideration is, first, how, from the rather small initial primitive vocabulary of some 500 words does the individual come, over the years, to increase the number of words in his vocabulary many times over, up to 25/30,000 words or so. Beyond that is the total extension of the English lexicon, which even if one allows for the rarities and hapax legomena of the full Oxford English Dictionary, must reach astronomical numbers going well beyond half-a-million words; with the accelerating growth of specialised vocabularies from year to year, it must now be getting to the point where there are between 3/4 million and 1 million English words. No precise estimate can be made. The available words change and increase from year to year; there is room for debate amongst linguists about what exactly should count as one word: the verb with all its inflections, singular and plural forms, adverbial forms and so on, but no one can challenge the proposition that English, as a language nowadays, must have one of the largest and most imposing vocabularies of any language in the world. There is obviously a mammoth task in any attempt to explain how this huge accumulation of words is built, some rationality in its structure is achieved and preserved and a widespread facility to make correct and creative use of the resources offered by the English lexicon is acquired by probably a quite substantial part of the English-speaking population. Though at first sight, the thought of trying to trace a link between the diminutive Primitive Vocabulary and this great mass of English words must seem intimidating, yet in other scientific fields equally intimidating tasks have been tackled and resolved in reasonably satisfactory ways; one has only to think of the millions of species, varieties of animals, insects, fish, plants, bacteria, viruses, of the endless multiplicity of chemical forms, of the ever-growing complexity of research in physics, to feel encouragement that the task of bringing some order to the phenomena of language should not be impossible. In the case of other complex systems, gradually sharpening ideas about the elementary units of the system, about the principles for the combination of those elements and the structuring rules governing the operation of the whole system have made it possible to formulate plausible and useful comprehensive hypotheses.

How then is the gap between the five hundred or so words in the illustrative Primitive Vocabulary and the typical adult lexicon of 25/30,000 words to be bridged (and the even greater gap between the initial 500 words and half-a-million or million total words available in theoretically complete English lexicon ) ? How, as a parallel question, can the elementary vocabulary at first available to map on to a very simple world (of the child) extend to the range of words needed to map on to the human intellectual terrain, the whole range of human knowledge and experience? The following attempts to list systematically the way in which the mustard-seed of the Primitive Vocabulary can grow into the huge tree of language:

1. Without introducing any new word-forms to the illustrative Primitive Vocabulary, as has already been indicated the words already in it can be used in new ways to provide a much more extensive mapping of perceptual experience. Words originally applied to single percepts or actions can be applied to a range or class of percepts and actions; simple words can be applied to refer to structurally similar aspects of experience (a primitive form of metaphorical extension of the use of words). Words referring to percepts first experienced as objects can be used as indications of action, nouns converted into verbs or adjectives; there can, in English, be a rich grammatical extension in the application to experience of a range of words. By these simple processes, the range of perceptual experience covered by the words in the list can without difficulty be extended many times over, possibly by a factor of ten or even a hundred;

2. Still confining oneself to words in the illustrative Primitive Vocabulary, one can extend their range of application by the straightforward process of composition, by putting together two words in the primitive list, some new percept can be identified or some specialised variant of a known percept indicated. From the word LEG, one can form, with other words in the Primitive Vocabulary, a whole range of words: TABLELEG, CHAIRLEG - simply by combining words falling in the same broad category of shapes and objects. Even more extensive combinations can be derived from putting together words referring to objects and words referring to movement, action, quality or relation, e.g. BACKDOOR, BACKSTAIRS, BEDSIDE, CATCALL, OVERSEA, UNDERCOAT and so on; obviously a very large number of combinations and permutations of the words in the Primitive Vocabulary is possible;

3. Then there is the set of minor variations in word-form, part of the process normally described as derivation, which may allow each word in the Primitive Vocabulary to develop into a whole family of words; there is, in English, a large number of bound morphemes (which might well be included in their own right in the Primitive Vocabulary) which are highly productive in yielding new applications of meaning from existing words, for example, prefixes such as RE- or UN-, suffixes such as -ER, -NESS, -ING, -LESS. These bound morphemes may provide particular applications of words in the Primitive Vocabulary, mark changes in grammatical function, or in fact generate words with quite distinct meanings, where the relation to the earlier primitive meaning is more remote.

4. From the three types of process described, it is apparent already that from a Primitive Vocabulary of about 500 words, many thousands more words can be derived and that, for the most part, these new words and new applications will still be closely linked to the perceptual structuring underlying the categories into which the Primitive Vocabulary naturally divides. At a guess, something like 5000 or 10,000 words may be readily derived in these ways from the original 500. Perhaps even more important, the area of perceptual experience covered by the available vocabulary (restricted though it still is) is likely to have increased by a much larger factor still. This is still only a small way from the initial 500 words of the Primitive Vocabulary to a reasonably typical adult vocabulary (though the words available may already be adequate for the greater part of the language-communication needs of the typical member of the community).

5. The next major extension of the vocabulary (which may in fact of course proceed in parallel with the extension derived directly from the words of the illustrative Primitive Vocabulary) comes with the addition of what are essentially other Primitive Vocabulary word-forms, words related to discrete items of experience, which may be old and familiar or novel. All the words for birds, animals, fruit, trees, minerals, fish, insects, flowers come into the vocabulary as primitive elements. A camel, just as much as a cat or a dog, has to be seen before the word can become a functioning part of someone's vocabulary. Some of the names of animals may, in the Saussurean sense, be completely opaque, non-self-explanatory; others nay be apparently combinations of preexisting word forms, such as BLACKBIRD, even though the identification of the particular bird cannot be derived from the combination of the two apparently familiar words. At this stage in the continuing growth of the Primitive Vocabulary, we are still chiefly concerned with words not borrowed or deliberately introduced as learned terms into the English language, words which for the typical English-speaker are thought of as native. Note that in parallel to the expansion of the individual adult's Primitive Vocabulary (in the sense defined), there is also a continuing growth in the so-to-say collective Primitive Vocabulary of English, as new items of experience come to be named; even a word such as 'television' is thought of as primitive, an object of experience rather than a word understood by definition. Insofar as the new items acquired are essentially extensions of the Primitive Vocabulary, they are not isolated but will find their place in the field-structure into which the Primitive Vocabulary divides; they will have a direct perceptual base and will not be learnable by the intermediation of other words i.e. it will still remain the case that experience of word and percept conjoined is essential, though extension of the perceptual repertoire can all the time, and will all the time tend to, proceed in advance of the corresponding extension of the lexicon in fact comprehended and usable by the individual.

6. How much progress then has now been made from the small initial vocabulary to the full adult vocabulary? or to the full English lexicon? Allowing for the application of processes of composition and derivation, the possible variations in grammatical use of individual words, it is now quite possible that the total available vocabulary may have reached something like 25,000 to 40,000 words or even more, certainly reasonable progress towards the typical adult vocabulary though still a long way short of the contents of any substantial English dictionary. The next great step forward in the size of the English vocabulary comes from a feature particularly marked in English as compared with most other languages, its extraordinary readiness as a historical fact to take in words from other languages (French, the Germanic languages, Latin, Greek) and often to naturalise them, to use them in parallel with or to supplement English words not far removed from them in meaning. Sometimes there is no English equivalent - the word TELEVISION has already been mentioned- but sometimes the words introduced seem almost duplicates of already available English words, so that the effect has been more to allow a finer discrimination of perceptual experience by the differentiation of the English and the foreign word. See for example the range of colour-words introduced from other languages. Some of the words borrowed into English in this way are, one may say, elements in the Primitive Vocabulary of the originating language; others, as borrowed into English, in every respect operate as primitive words e.g. one cannot know by definition the perceptual meaning of the word BEIGE. But where the borrowed words are not themselves primitive words in their original language, they will have been indirectly derived from the primitive vocabularies of the originating language. For example, JUSTICE comes from a Latin primitive word JUS, INDECISION from a combination of Latin primitive words equivalent to UN-, OFF and CUT (IN DE CAEDERE). Meanings of derived words such as these in English are generally not primitive; their meaning can be explained by a string of simple English primitive words; the motive for their introduction into English often has been allow the compression into a single more concise form of a complex set of actions, percepts, events, relationships. Such words are for convenience rather than essential but they represent an extremely important new principle in the generation of the English lexicon, as compared with the development directly from the English primitive vocabulary. Such words survive because they are useful; no one, without a rather lengthy use of other words, could by immediate contemplation of a word such as MISDEMEANOUR, rather than a definition, arrive at a precise view of its meaning or use. The process by which the particular word MISDEMEANOUR becomes attached to a particular area of experience is complex, historically and intellectually, but no doubt it could plausibly be attributed to the operation of what the general semanticists would describe as their third principle, the self-reflexive potentialities of language, the creation of new words to refer to groupings or strings of existing words. One final comment on the significance of this type of borrowing for the expansion of the English lexicon: the driving force for the addition of words to the lexicon comes from the ever-growing extent of individual and collective perceptual and intellectual experience, the growth of knowledge in all its forms, science and technology, knowledge of the world and of societies. Perception for the individual and for the community is open-ended and language has to keep growing to match it.

7. Starting from the small initial Primitive Vocabulary of some 500 words, one has proceeded to a vocabulary of some 25,000 to 40,000 words by processes of derivation, composition &c applied to primitive English elements and beyond this one has an immense accession to the vocabulary from the introduction of words originating in the primitive vocabularies of other languages and from the application of composition, derivation, prefixation and suffixation to these words. And note that what has been borrowed from other languages includes not only particular simple or compound words but also word-elements, bound morphemes, which form part of the primitive resources of foreign languages and have been applied to extend the range of variation of meaning of existing English words e.g. the use of suffixes like -ISM, -TION, derived from Latin, Greek or French to provide new adaptations of originally English words. It would be an immense task, only to be undertaken by those who have devoted their professional lives to the compiling of dictionaries and the problems of etymology and philology to present any close statistical analysis of the format of the total English lexicon in terms of the different sources from which and the different processes by which it has been built up but the vital point with which the discussion in the present part of this Chapter began, that is the distinction between words understood only by experience and words understandable through description in terms of other words remains. The greater part of the lexicon, the borrowed words, and particularly those borrowed for specialised uses, fall into the second category, that is they must be explained in terms of words which ultimately are self-explanatory but one should not forget that within the lexical structures of the languages from which English has borrowed its words, exactly the same process has operated, that is the formation of the foreign lexicon from an original primitive core; one can arrive at the meaning of compound foreign-derived words often by analysing them to discover the primitive core from which they have been built. The primitive core in the other foreign languages is, on the contention of this chapter, as firmly based in the necessities of human behaviour, human perception, action and internal experience, as is the core Primitive Vocabulary of which an illustration for English was given earlier in this Chapter; that is, mediately or immediately, the whole lexicon is pervaded by the naturally-derived structuring of human experience.

To sum up, where has the discussion so far reached? That segment of the total lexicon which relates to external perceptual experience, has a strong, simple framework, directly relatable back to the original core Primitive Vocabulary (matching the original primitive repertoire of experience), and this structure is the foundation of order and correctness in the use of words. The problems of semantics really start with the use of words to deal with more complex internal human experience, going beyond the more direct internal perceptual awareness of one's state of feeling and so on, to the content of intellectual experience, verbal rules, laws, judgments of value. It would not be surprising if, in a sense, at this remote point, the lexicon, that is the use of language, tends to get out of control, to 'run amok', as the links tying back the compound and complex words used to the firm central perceptual framework become too tenuous; the endless verbal disputes of philosophers, theologians, lawyers, literary critics, politicians, pedagogues and, not least, linguisticians and grammarians, show that in these higher reaches, the lexicon has in effect gone out of control, reflecting of course the unclear perceptions of the parties to the disputes of the functioning of their own minds, their own languages, their own societies. Language cannot preserve those who use it from their own failures of perception.

Perhaps this is a disappointing result but there is a very large part of the lexicon which is solidly based and relatable to perceptual structures and this is adequate for all normal uses in life. However, at a certain point, the checks on innovation in language, on idiosyncrasy in language and the diversity of idiolects, become too weak because of the distance of the discussion from the reliable normal processes of human perception (particularly visual perception). One might say there is a failure in the satisfactoriness of the structure of the lexicon, though from another point of view, it could be said that much or all of the disputes about the use of terms are failed attempts to arrive at a clearer view of the organic relation of the available words to the strong central structure of language as a matching of perceptual experience, a mapping of human experience. Structures in the lexicon which are perfectly comprehensible and reliable applied to simple material become chancy applied to less clearly-defined material. This brings the argument straight to the most vital feature of language-development, which can be welcomed for its contribution to creativity and deplored for its contribution to confusion, that is the process of the metaphorical extension of words and language-structures. Metaphorical transformation is the key process in the development of language and thought, a subject which must be examined with some care. Metaphor, etymologically, simply means 'transfer'; traditionally it has been treated as a figure of speech, an elegance of language; 'the transfer of a word or phrase to describe or denote something entirely different from the idea, object, action or quality which it primarily and usually expresses, thus suggesting a resemblance or analogy'. But this is to take far too narrow and traditional a view of what is as fundamental a process for the operation of language as the power of abstraction and generalisation is for the functioning of thought. One might quote again, with even greater emphasis this time, John Locke's observation "how great a dependence our words have on common sensible ideas, and how those which are made use of have their rise from thence, and from obvious sensible ideas are transferred to more abstruse significations, and made to stand for ideas that come not under the cognisance of our senses e.g. to imagine, apprehend, comprehend, adhere, conceive, instil, disgust, disturbance, tranquillity &c are all words taken from the operations of sensible things and applied to certain modes of thinking."(17) In the terms which have been used in this Chapter, all the words mentioned by John Locke and a vast host of other abstract words can have their practical significance traced back to elements in the Primitive Vocabulary either of English or of other languages from which primitive elements have been borrowed; we may comprehend or apprehend an idea or equally say that we grasp or hold an idea in our mind; both the words derived from Latin and the simple English words say exactly the same thing, and express, through transfer of the most ordinary movement of our hands and arms, the remote conception of fixing in our mind a particular complex of thought.

Metaphorical transformation is a process which can operate at the word-level, the phrase-level or the sentence-level (as the standard dictionary definition would say) and note, most importantly, that it operates to transform the character of every word in the metaphorical phrase; if one says that 'one has an idea on one's mind', not only is the sentence a strange kind of metaphorical picture where a theoretical object, an idea, is placed on a theoretical surface (like a table, perhaps a tabula rasa) just as a cup might be placed on a table, but the small functional word ON itself is being used in a transferred, metaphorical way; ON gains its meaning as part of the Primitive Vocabulary from the normal experience of seeing something on a table or other object, or placing something on something, a cap on one's head. By the process of metaphor, the whole relational system, both in time and space, expressed in English typically by prepositions is transformed into an abstract intellectual space, the space of one's mind and consciousness. Perhaps much more important than the traditional 'poetical' use of metaphor is the structural role of the metaphorical process in making available for use in discussing abstract matters (the complexities of human thought, language and intention) the simple structures of the Primitive Vocabulary, the patterned relationships of ordinary perceived and experienced objects, actions, events, qualities.

This high estimation of the value of the metaphorical transformation is not original or unusual. Though great reams of commentary have been written about the use of metaphor stylistically, there has also been a substantial literature about the more fundamental role of metaphor. To take a selection of comments from others who have written on the subject:

'All language, by the nature of its 'transferring' relation to reality.. is fundamentally metaphorical. Metaphor is not something special and exceptional..It is the omnipresent principle' of all language...All languages contain deeply embedded metaphorical structures which covertly influence overt meaning'. 'Metaphor in short is the way language works. A metaphor is made out of and it makes those realities (of life and language).. Metaphor is at the centre of language that the process of metaphor is located at the heart of language and indeed defines and refines it, and thus man himself, remains the central stance of most Twentieth Century writers on the subject and their overriding preoccupation" (Brook-Rose 1958, Hawkes 1972, Richards 1936, Black 1954)(18)

Much of what is quoted above is taken from I.A. Richards' The Philosophy of Rhetoric(19). Language, as others have pointed out, teems with fossilised metaphors, with abstract words from one European language to another tending to derive from the same metaphors.

So far, this chapter has discussed semantic structuring in terms of the formation of a Primitive Vocabulary based on simple perception, action and language experience of the child; it has explained how, starting from this structure, by a variety of processes, a much larger lexicon is built up (in English) covering an ever greater area of the total perceptual and other experience of the typical adult, and eventually the whole field of experience and knowledge enshrined in the English language. The growth of the lexicon has taken in and domesticated material from foreign languages, originally based in exactly the same way in those other languages on the primitive elements of experience so that underlying the whole of the experiential lexicon are the simplicities of the original structure, the origin of words in the different sensory modalities, in the different physical frames of experience of the body, the room, the house, the street, the landscape and in the different direct internal experience of the individual, his feelings of hunger, weariness, joy, excitement, expectation and so on. Beyond this great block of the English lexicon is the even larger block of the abstract lexicon, for manipulating thought, constructing scientific theories, evaluating literary forms, describing the subtleties of the human mind and character, elaborating philosophies and religions. These two great blocks of the lexicon, the more immediate perceptual block and the abstract block, are knitted together, organically related to one another through the absolutely basic neurological process of metaphorical transformation. Metaphor is not a linguistic device primarily but a perceptual device; it is the manner in which our perceptual system extracts from the multifarious patterns of reality analogies of structure, uniformities of organisation between things and events which may at first seem quite distinct. The metaphorical process as it operates uniformly to extend language and perceptual experience to match each other is only the last stage in the great hierarchy of physiological and neurological processes by which the eye and the ear, the visual and auditory apparatus and brain, extract from endless apparent diversity of form the definite, patterned persisting objects, sounds, events and actions from which our world is made. At the humblest level, the visual and neurological processes which from the endlessly varying appearances of any single visible object (at different angles, in different lights, in different positions and relations),the brain identifies this structure (and records it) as A CHAIR or A TREE (however many different sorts of chair or tree there may be), is the same category of process as that by which a political theorist identifies a persisting society, a lawyer the structure of a legal system, or a linguist the persisting inherent ordered structure of a language.

Most briefly, the metaphorical process (of language and perception) unites the Primitive Vocabulary and the total lexicon, the primitive perceptual repertoire and the total range of human experience.

The natural origin of individual words in the extended Primitive Vocabulary

The first part of this Chapter has discussed how what is described as the Primitive Vocabulary (that is: words whose meanings can be acquired only from perceptual experience and not by definition in terms of other words) forms the core of the total English lexicon and communicates a specific structure to the total lexicon, reflecting the origin of words as sound-forms matching the familiar features of experience, the different aspects of human behaviour in action, perception and language. So far, the chapter has argued that on the largest scale, the lexicon is structured by the physiologically and neurologically derived features of actin, perception and language. The present, second part of the chapter complements the first part by presenting an account of the way in which the sound-structure of each individual word in the extended Primitive Vocabulary gives either a direct representation of the percept or action referred to or at least a helpful clue or cue to the meaning of the particular word, the contention being (as set out in Chapter III) that there are necessarily direct physiological and neurological relations between the patterning underlying action, perception and articulation and that the combination of elementary sounds to form a particular word derives from and represents the elements in visual processing or in action which go to form a perceptual contour or a contour of action related to the meaning of the word. The final section of this Part of the Chapter demonstrates specifically in relation to a variety of words drawn from the illustrative Primitive Vocabulary on page 61, the exact manner in which the sound-structure of the word and the visual or action contour of the percept, action or event are related.

However, before these specific illustrations of the links between individual words and associated visual or action contours can be presented, there is an important intermediate stage. At the end of Chapter III, a scheme of equivalences between the elementary units of vision, action and speech was set out, that is the isolated elements in each modality. But the words we may have to deal with related to complete percepts, distinct actions, defined shapes and objects. One has to bridge the gap between the presentation of the elementary units (the individual sounds &c) and the formation of groups of elementary sounds into words, groups of visual elements into shapes. and elements of bodily action into complete types of action. At one stage Saussure(20) spoke about the need at some future time for a new science of combinatorics, the principles on which elements can be formed into larger unities, and it is precisely this which has to be tackled in the present part of this Chapter. The matter is one of great difficulty: in vision research, progress in pattern-recognition and the construction of patterns from elementary visual features has made less progress than was at one time hoped for; in speech research, whilst there has been intense research in phonetics (the isolable elements of speech-sound) there has been much less attention devoted to the principles of the combination of phonemes, the merging of phonemes into the integrated structure of the word; in relation to the organisation of bodily action, the position if anything, is even less encouraging. In Chapter III, Karl Lashley's(21) pessimistic assessment of the progress made in the syntactic analysis of action, of voluntary movement, was quoted and much less is known than one would wish about the manner in which the intention to perform a bodily action is translated into the actual physical action.

Nevertheless, an attempt must be made to tackle these difficult matters. The order in which they are treated is the same as that adopted in Chapter III, that is first the construction of visual forms from elementary visual units, secondly the formation of more complex actions from the elementary units of bodily action and thirdly the formation of individual words from the combination of the elementary speech-sounds (the phonemes). The construction of visual forms

This section considers the basis on which in perception the visual system extracts from the light falling on the retina particular shapes, forms and objects, as a second stage building on the initial identification in perception of the isolable elementary visual units discussed in Chapter III. The whole subject is extremely complex and no definitive results have yet been reached in the massive research programme which has been under way for many years. There are some preliminary points which can usefully be made, however:

1. The concern of this section of the Chapter is not with the general organisation of the visual field (the classic analysis by the Gestalt psychologists &c of the division of the perceived scene, by semi-automatic processes, into Figure and Ground, the total-field characteristics which the visual constancies of size, shape, depth, colour &c seem to derive from or the interrelation of distinct items in the visual field. These are matters more relevant for the comparison of 'visual scene' and 'sentence-structure' which is dealt with in the immediately following Chapter V. The emphasis is on the individual visual shape, from the point of view not so much of its recognition in practical experience but of the manner in which the individual shape is formed, derives its structure from more elementary units. This emphasis is chosen because the comparison of critical interest is with the overall structure of the word, the sound structure formed from elementary sound-units (and again not primarily with the recognition of the word in the complex, always abnormal conditions of ordinary speech);

2. So far the problem of shape recognition in ordinary vision has remained unsolved, though there have of course been a number of theories, not least those stimulated by Hubel and Wiesel's discovery of receptive cells in the visual cortex which respond selectively to various stimuli (as more fully described in Chapter III). As has already been noted, it is still a far cry from the elementary ability to respond to specific elementary contour elements (the elementary visual units) to mechanisms that can learn to classify shapes such as tables and chairs, according to complex properties for which one cannot expect innate detectors to have developed during evolution. However, a great deal of work is in progress, not only in the field of the physiological bases of vision but in that of pattern-recognition, machine-vision, in the field of artificial intelligence and useful clues towards the way in which shape-construction and shape-recognition may operate are gradually being gathered;

3. There is no reason to believe that there is any single, unchanging mode of construction and recognition of visual shapes either throughout the life of the individual, from infant to adult, or necessarily between one adult and another. There seems every reason for assuming that the visual apparatus will exploit. all the modes of shape exploration that are open to it, and the overlap and redundancy between different modes will offer real benefits: in terms of the reliability of perception, the accuracy of perception under distorted or difficult conditions and so on. If there are several ways of forming a view of what a particular shape represents, the chances that the shape will be correctly identified will be substantially increased and the effectiveness of perception, as an instrument serving the organisation of action, correspondingly increased; it is possible that in the development of perceptual experience, there can be a move from initial reliance on the identification of specific elementary visual units to form percepts towards use of much more shortcut methods when familiar percepts have become stored and labelled by words in the central nervous system. In this way, a new much more rapid process of recognition is made available to the adult than is available to the infant or child. In much the same way, whilst a child, in reading, may first identify a word by taking each letter separately, the adult reader learns to recognise whole word-shapes, and so makes possible very much greater reading speeds;

4. What at a minimum seems clear, as a result of research into the perception of visual shapes, is that there are two vital elements. The first is that visual perception is an intensely active process, depending on the construction of particular percepts from unceasing movement of the eyes, in saccades and fixations, pursuit and search movements of the eyes, movement of the eyes in accommodation by adjustment of the ciliary muscles, and change in direction of gaze as a result of co-ordinated movements of the head. As Lashley said: "I have come to feel that the problem of scanning underlies many other problems of neurophysiology. Visual perceptions are rarely based upon a momentary stimulation of the fixed retina ... most of our perception of objects is derived from a succession of scanning movements, the succession of retinal images being translated into a single impression of form."(21) Neisser, after quoting these remarks, commented that this act of translation of a succession of 'snapshots' into a single form has hardly ever been studied but is evidently among the most fundamental cognitive processes so that, under normal conditions, visual perception is an eminently constructive act: "The individual snapshots are remembered only in the way that the words of a sentence are remembered when you recollect nothing but its meaning: they have contributed something which endures"(22)). The second vital element in the perception of visual forms is the profound preoccupation of the visual system with the contours of visual shapes which suggests that contours play a special part in shape recognition (and hence in shape construction). Children gradually develop the ability to follow contours, and contours retain their importance in the recognition of novel shapes throughout adult life.

After these rather extensive preliminary comments, the reader may look for a systematic statement of the accepted processes of shape construction and recognition by the visual apparatus. No such generally accepted account exists but it may be right here to quote one of the more interesting, and ambitious, accounts of the process of visual shape perception, that in Sommerhoff's wide-ranging book Logic of the Living Brain(23) from which some of the preliminary material above has already been drawn. The following outline of possible mechanisms for shape recognition (based on a tracking of contours, generally relying on central vision - that is the fixation of the perceived object in relation to the fovea) is drawn from him:

1. At the lowest level, we have the 'primary analyzers', the wired-in detectors; for particular contour features such as for bars in particular orientations, bars stopped at one end, angles &c. We cannot assume of course that the wired-in detectors discovered by Hubel and Wiesel are the only ones; there may be more complex analyzers not yet discovered. At some level of complexity, we must reach a point where a particular feature is no longer 'innately' detectable but requires visual exploration and learning;

2. For these higher levels of complexity, learning processes are required which, initially at any rate, depend on the outcomes of exploratory activities such as contour-following. To pass from knowledge of the salient contour features or elements of a shape to knowledge of the shape as such we need knowledge of the spatial relations between these features or elements. On the reafference principle (that is: the process by which the visual apparatus derives information from its own self-generated movements), the brain can factor out information regarding the spatial relation between particular contour elements by moving the eyes from contour element to contour element while registering at the same time the nature of the eye-movements required. It can do this either by scanning the contour elements in their proper serial order (contour-following) or by jumping haphazardly from salient element to salient element. Either method would yield the required information. To register the outcome of these movements is in effect to compile a list of the characteristic properties of the contours in question, that is of the shape formed. The movement information required could be derived either directly from the oculomotor apparatus (movement of the eyes &c) or from corresponding internal attentional processes.

3. In continuous contour-following, the characteristics of the contour are represented by a family of permitted eye-movements; this representation is invariant in respect to the translation of the object within the visual field. And indeed if only the relative directions and relative magnitudes of the movements were registered by the brain, the resulting representation would also be invariant with respect to the size and orientation of the image i.e. a process by which one would arrive at the record of a constant visual percept (somewhat analogous to the lines of constant orientation (though varying retinal position) identified by Hubel and Wiesel's complex cells in the visual cortex). It perhaps ought to be noted here that 'contour-following' itself is not a smooth process; it proceeds by a series of jumps, saccades, from one point of fixation to the next, and the record of a shape derived from contour-following would thus take the form of a succession of these irregular eye-movements in straight lines from one point to another.

4. Given a sufficiently comprehensive exploration of the shape by the eye, the aggregate of the registrations of the eye-movements forms an adequate neural basis for categorising the shape according to the mutual relations between the continually recognisable elements (lines, curves, angles, terminations etc.). It is difficult to see why the brain should not avail itself of such records. The attentional responses with which we inspect one portion of a stimulus after another must take a form which corresponds to the shape and other attributes of the object. Internalised versions of these responses could therefore serve as bases for the internal representation of the objects.

5. If, as perceptual learning progresses, a sufficient sample of these classifying outputs can become conditioned to any particular projections of the shape, then the mere occurrence of that projection in due course permits instant recognition of the shape i.e. recognition without prior exploration (in the same way as wired-in detectors for elementary units such as straight lines allow their instant recognition by the young animal). The system as it stands eventually permits the shape in any projection experienced during learning to trigger a characteristic set of expectations about what stimulus the fovea of the eye will receive following any one of a number of possible movements of the eye. The complete set of these shape-expectations constitutes the 'model' that is activated in the separation of the figure from the ground and the act of recognition (and in the associated act of naming of the perceived object).

The above account, drawn from Sommerhoff, is unavoidably somewhat complex and it may be helpful to summarise at this point what seem to be some of the points of more direct relevance for the hypothesis being presented in this Chapter, that is, the direct structural relation between visual shapes, word-forms and action-contours. Contours are very important for recognising shapes and initially shapes appear to be recognised by the elementary features they contain and the eye-movements involved in transition from one elementary feature to another;

b. recognition of a shape (in visual search) seems, as Lashley says, to involve readiness to recognise a motor sequence, starting from some line or angle in the picture and then scanning adjacent lines. Eye-movement and shape-perception seem to be interlocked, mutually dependent processes; one might say that we perceive visual shapes through movements of the oculomotor apparatus, the eyes, the turning of the head, the adjustment of the ciliary muscles for focusing; the interaction between movement and perception seems to remain reciprocal;

c. the key question is how the succession of retinal images resulting from scanning is converted into a single stable impression of form. The structure of percepts at first appears to be formed from the elementary visual units but these units may be assembled into higher-level sub-groups. What seems to be involved is both an elementary feature-detection system for shape-recognition (as used in some computer programs for vision) and, at a later stage in human development, with perceptual learning, a 'template' or 'schema' matching system, with which can be associated the permanent labelling of visual shapes and objects provided by word-forms.

Looking ahead to the section of this part of the Chapter dealing with the formation of words from elementary speech-sounds, one can relate the discussion to the vital importance of the contour of the word-shape, formed from elementary units of speech-sounds, parallel to the elementary units of vision, and chained together into a more complex structure by the succession of movements of the articulatory organs necessary for the production of the word, the characteristic pattern of movements identifying the stable word as the characteristic pattern of eye-movements identifies the stable visual shape. In parallel to the visual template or 'schema' developed for familiar shapes and objects, one has auditory and articulatory 'templates' and 'schemas' for familiar words. For both vision and speech, there is a parallel progression, by hierarchical stages, from the simplest to the most complex forms.

The construction of action-forms

Chapter III attempted to identify the elementary units from which more complex forms of visual shape, action and speech-forms are constructed. The immediately preceding section of this Chapter has attempted to give some indication how, physiologically and neurologically, the elementary units of vision may in fact be combined to form more complex visual shapes. The purpose of the present section is to make a corresponding attempt for the formation from the elementary units of more complex types of bodily action. The discussion of the natural units of action-organisation in Chapter III started from Karl Lashley's stimulating speculations about the 'syntax of action', and the extent to which it shared properties with the syntactical organisation of language and vision. At the same time, Chapter III made it clear that relatively little directly useful research work has been done on the voluntary organisation of bodily action, apart from that of Nicolai Bernstein and that reported on by Evarts. The elementary units of bodily action identified in Chapter III were relatively straightforward, that is simply the varying directions, magnitudes and types of movement possible for the different bones and joints of the human body, movements of the hand, arm, leg, trunk, neck, head. If one continues to follow, as has been done so far in this book, the concept that behaviour is organised hierarchically, progressing from the most elementary units of vision, speech and action through intermediate stages to the most complex forms of vision, speech and action, then in this Chapter one is concerned with the stage of organisation of bodily action equivalent to what for vision is the stage of the formation of the visual shape, the recognition of the individual object and for speech is the stage of the formation (from the elementary speech-sounds) of the word-form, the combination of speech-sounds into the multitude of discrete words which go to form the English lexicon. The question is: what, for the organisation of bodily action hierarchically, is the equivalent stage to the visual shape or the word-form? Clearly, it is not continuous and complicated sequences of action, or rather of interaction with the external environment, of the kind represented by playing a musical instrument, using a typewriter, painting a picture, driving a car, or dancing a Scottish reel. These complex continuous actions are much more obviously analogous either to the complex visual scene or the compound and complex sentence which are the main subjects for consideration in the next Chapter, indeed to the idea, for vision and speech, of the continuous stream of perception, the continuous stream of discourse, which individual scenes and individual sentences go to form. What one is concerned with in this section of the present Chapter is the intermediate sub-routines of action, the well-defined but limited organisation of a small number of elementary units of action into a characteristic general type of bodily action, such actions as taking hold of an object, stretching out to touch an object, throwing something, sitting down or standing up, possibly walking or running, pushing or pulling at something biting, eating, yawning, turning or climbing - and so on. Thus, the kind of sub-routine action this section is concerned with is similar to those which have already been listed amongst the section for "movements and actions' set out in the illustrative Primitive Vocabulary on page 61. Combinations of actions of these kinds, arranged in sequence, carefully timed and co-ordinated, related specifically to objects in the external environment, go to form the syntactic expression of action. A stroke in tennis, for example, is made of many of those action sub-routines, the grasping of the racket, the various movements of the arm, the striking of the ball with the racket, the adjustment of the movement and positions of the foot, legs and trunk, and so on. For the construction of the complex sequence of action represented by a stroke in tennis, the player must have instantly available the patterns of activity necessary for each component part of the total activity and since, at any moment, the player's bodily position will be different, the position and direction of travel of the ball will vary, the direction in which the stroke must be played must be adjusted to the needs of the exact point reached in the rally in progress, no rigid sub-routine of hitting lifting the arm, changing the position of the legs, will be of any use; the player must have available, as a prepared pattern of muscular co-ordination, a sequence of elementary units of bodily action which can be implemented through the use of any of a very large number of different muscles. As Karl Pribram(24) has described it, the player must have an 'image of action', the actual realisation of which may take any one of many different forms, dependent upon the necessities of the real-life situation.

In considering then, in this section, what sort of combination of elementary unite of action may go to form a simple action, one is concerned, as in the case of the construction of the image of a visual shape, or indeed of the construction of a word-form, with a structure which is to an extent a generalised representation of a sequence of actions directed towards a specific goal, a particular type of action. One is concerned with the representation of the inner-structure of the action of say, hitting or throwing, the simplest, most general specification of the action of throwing which can be mobilised as an element in a total sequence of action and integrated smoothly and flexibly with other elements in the sequence of action. If one asked someone, as a perfectly general matter, to demonstrate or illustrate what is the nature of hitting or throwing, they would respond with a particular movement of the arm, the arm bent and then extended with some force in the case of hitting; the direction of movement of the hand and arm would probably generally be downward and in the case of throwing, there might be some closing of the hand, with the direction of movement of the hand and arm being forward rather than downward.

For each of the primitive words relating to bodily action, those in the illustrative Primitive Vocabulary and many others, we all knew from experience, not by being told, what component units of action are required to go to form a typical example of any particular type of action-routine. The way in which elementary units of action are formed into intermediate forms of action and then into complex sequences of action is much more familiar to us, much more immediately understood, than the ways in which elementary units of vision go to form visual shapes or complex visual scenes, or the ways in which elementary units of speech-sound (phonemes) go to form words and combinations of words in sentences and discourse.

The points which were identified as of vital importance in the discussion of the formation of visual shapes in the immediately preceding section of this chapter, that is, the significance of the contour of the action-form, the formation from elementary units (of action) and the chaining together of the elementary units into a more complex structure by the succession of movements, apply with equal force to the analysis of the formation of the intermediate action-types just discussed. One can expect that just as for vision and speech, there are 'templates' or 'schemas' developed for familiar, frequently used shapes, objects and words, so there are centrally-recorded templates or schemas, at a high and rather abstract level, for the production of the frequently used and absolutely essential component action-forms, which are used to construct complex sequences of skilled activity.

One final comment: what has been said in this section may seem either unduly speculative or unduly obvious, but this is inevitable because of the difficulties in neurophysiological investigation of the control of movement. In his recent book The Purposive Brain, where he deals to a good extent with questions of motor control, Ragnar Granit comments that twenty years of experimentation in neurology and related disciplines has provided a mass of very important facts but these represent but a minute fraction of the knowledge we need to understand how one makes a gesture with the hand or interprets a complex visual image... [Given the incredibly complex information and command systems in the brain for motor control] ... "My personal reply to this is that scientists may have to abandon the hope of ever understanding the whole miraculous performance of our sensorimotor brain, but they can never give up searching for leading principles of organisation ...It does not seem probable that neurophysiology will ever succeed in tracing all the relevant connectIons leading up to a purposive movement".(25)

Ragnar Granit in effect is saying that, because we may never understand everything, the complete functioning of our own brain as it models and controls our behaviour, there is no reason to abandon the scientific endeavour. We have to aim to understand as much as we can understand, and the analysis in this Chapter of the functioning in parallel of the processes of vision, action and speech is intended as some contribution to that growth in understanding.

The construction of word-forms

The two previous sections were concerned with the manner in which elementary visual units and elementary units of action were formed into, respectively, visual shapes and action-patterns. The present section considers how, similarly, the elementary units of speech-sound identified in Chapter III, essentially the familiar set of phonemes, are formed into word-structures, into the simple and complex words which go to form both the content of the Primitive Vocabulary and the total English lexicon. The interest is in how different phonemes can be combined in English, what constraints there are on combination and on the positions in any word taken by any particular phoneme and what the implications are for the production or recognition of a word as being formed from certain phonemes, of the phonetic context and, in particular, of the initial or final position taken by the individual phoneme. The kind of work represented by Bloomfield's chapter on phonetic structure shows that one can, justifiably, reach conclusions about the frequency or the difficulty of particular combinations of phonemes in English words, regardless of what view one takes on the acoustic or neurophysiological status of the phoneme as such. But this is not to say that, for the purposes of the thesis in this book (that is that there are physiological and neurological parallelisms between vision, action and speech) one can lightly assume that the phoneme is an artificial or purely formal and conventional construct. Chapter III has in fact argued that, following Liberman, the discrimination of phonemes is a reality, not so much in terms of the acoustic trace but in terms of patterning at the level of the neuromuscular system controlling the distinct articulation of different phonemes, and providing for the integration into smooth sequences of (permitted) combinations of phonemes. The argument will be that the constraints on the formation of word-structures from particular combinations of phonemes which are as a practical matter observed in English are not accidental or conventional but reflect underlying biases and constraints on the combination of articulatory movements, and at a higher level on the patterning of neural commands for words. Whether or not speech is comprehended by a kind of reverse process to that for speech-production as Liberman suggests with a good deal of plausibility, the form of words actually available in a language must depend on the permissible combinations and permutations of the required adjustments of the articulatory organs (and there may be neurological as well as anatomical factors which bias any particular language towards the use of one set of phonemes rather than another or towards the use or avoidance of particular combinations of phonemes in forming the words of the language.

In a sense the problem of speech recognition, speech comprehension, is a quite different one from and much less fundamental than that of speech production, the formation of phonemes into word-structures. Speech becomes audible because of the impact of the variation in the air-stream from the lungs (as modIfied by adjustments of the vocal tract) on the pressure-patterns in the external air, the medium of distance-communication, but the acoustic patterns in the air so produced are, to speak in a rather extreme way, simply a by-product of the "controlled gestures of the vocal organs"(27) (Ladefoged's phrase) by which speech is produced, physiologically and neurologically; the events taking place in the individual are the sequencing of the neural impulses to the muscles which control the shape and movement of the various elements which go to form the bodily speech-apparatus, There is an essential identity of central nervous control of speech-movements and other types of voluntary movement. It is a reasonable hypothesis that the production of speech involves sequences of movements similar to those of rapid typing, or playing the piano. If one takes this essential identity to its logical extreme, one could in fact analyse the production of speech, the control and patterning of movements of the articulatory organs, simply as one sub-species of the general problem of the neural and muscular control of voluntary movement of all kinds, a sub-division of what I have earlier described as 'action-organisation' as a part of total human behaviour.

Approached in this way, the analysis of the formation of word-structures from the elementary units of speech-sound (that is the phonemes as representing unitary organisations of positions and movements of the articulatory organs) can readily proceed in parallel with much of what has already been said about the formation of visual shapes from elementary units and action-routines from elementary units of bodily action (the unit movements of the mobile parts of the body). The concept of contour is just as important and informative for considering word-formation as it is for the formation of visual shapes or action-routines; others have noted the profound preoccupation of the auditory system with articulatory contours - and one might link this with what Brain(28), in a general discussion of the relation of physiological speech mechanisms and the aphasias referred to as the tendency for the system to operate with what (following Head and Holmes) he termed 'schemas' not only auditory phoneme-schemas but also central word-schemas, word-meaning schemas, sentence-schemas and motor phoneme-schemas. What is in vision constituted by the formation of contour-elements into unique discriminable patterns is in speech formed into unique discriminable articulatory patterns (with speech comprehension as a process of 'breaking the code" which the translation of successive phonemes into acoustic variations represents).

When phonemes are linked together to constitute a word, this process is more than a simple addition: a word constitutes a unity or pattern at a higher level of organisation, with Head's word-schema playing the same part in the recognition of the word as the phoneme-schema does in the recognition of the phoneme. A word is perceived as a whole which is something different from the sum of its parts. The relation of invariance for words and phonemes, derived from the unique patterning of articulatory movement, the ideal form which may be translated in varying ways depending on the immediate circumstances of the position of the articulatory organs and general anatomical factors varying from individual to individual, is no different in character from the invariance for visual shapes derived in some such way as that suggested by Sommerhoff, as described in the earlier section of this chapter, from the elementary information provided about the sequences of movements of the eye-muscles in tracing contours and identifying elementary features.

The recognition of the word (or phrase) in continuous speech-sound and of the simple or complex object in the stream of visual stimuli seems to involve the same problems and require some of the same solutions; in each case, some version of the Figure/Ground mechanisms seems to operate, with a known form being brought sharply out by enhanced contrast from the background, remembering that in the case of heard speech-sound, the word has to be detached, isolated, not only from other words, other patterning of speech-sound, but also from ancillary sound (dialect, emphatic and attitudinal variations) and more generally, from the whole flow of non- speech sound in progress at any moment. The word-object has to be identified in the 'auditory scene' (that is the total play of sound of all kinds on the basilar membrane of the ear at any one time) in the same way as the visual shape-object has to be identified in the visual stream playing on the retina. The fact that instrumental acoustic research shows that neither the phoneme nor the word is necessarily identifiable by segmentation or distinct markers in the acoustic record (in terms of frequency or intensity) is really no more surprising than if one recorded similarly the equivalent components of the light-stream, the intensity and frequency of the dots of light, and found that there was no mechanical operation of segmentation which allowed one directly to identify the diversity of objects represented in the stream. Acoustic research has shown simultaneous cueing (multiplexing) for phonemic and syllabic features and clearly the phenomena of figure and ground, the visual constancies, indicate that there can be simultaneous cueing for distinct aspects of the visual scene.

In the case of both speech recognition and visual perception, there clearly must be elaborate decoding processes, to recover from the crude acoustic and visual stimuli the structures from which they were originally derived, in the speaker or in the objects from which light has been reflected. The subtleties of the process of decoding, certainly in the case of speech, appear to be very considerable, though substantial progress has been made in identifying them. A great deal of this work has been done by Liberman(29) and his associates at the Haskins Laboratories. Given the inexplicitness of the acoustic record, the listener appears to use the inconstant sound as a basis for finding his way back to the articulatory gestures that produced it, and thence, as it were, to the speaker's intent. Perception depends on inferences unconsciously drawn about the originating patterns of articulation. Indeed, it has been suggested that in the same way as the eyes are used to scan a visual scene, to extract information from the periodic fixations on salient points, so the hearer's articulatory apparatus in some way is used as a basis for scanning the information provided in the sound-stream. In the process of decoding; as Liberman(30) pointed out at the 1975 New York Conference, silence may be as indicative as sound; the gap in the acoustic record for the stop consonant because it provides information that a total closure of the vocal tract was made necessary for the production of the stop consonant, and the hearer has in this way been given his clue to the articulatory patterning of the speaker. Variations in acoustic shape may not simply indicate presence or absence of phonemes or syllabic fragments but encode the order of phonemes to allow one to distinguish e.g. the word DAB from the word BAD.

There has been some direct evidence of the invariance of the phoneme (and beyond that of the invariance of the word) in terms of articulatory patterning from electromyographical research techniques. Liberman recorded that when two adjacent phonemes were produced by spatially separate groups of muscles, there were essentially invariant tracings in neural signals to the muscles (the EMG record) from the characteristic articulatory gestures for each phoneme, regardless of the identity of the other. When the temporarily overlapping articulatory gestures for successive phonemes involved more or less adjacent muscles that controlled the same structures, they found essentially identical EMG signals for the initial consonant but the EMG signal for the following phoneme involving the same feature of the articulatory structure might show substantial changes from its characteristic form. In the onsets and offsets of EMG activity in various muscles, the researchers found a segmentation like that of the several dimensions that constitute the phoneme i.e. one could find in the neuromuscular record where the phoneme boundaries must be, a matching of the contour of a word from a succession of phonemes, in a way quite unlike that found in the acoustic record. This seems clear evidence of the fundamental relation between word-structure form and the sequencing of articulatory movements, that is a relation between the word as a unit in speech and the action-routine in bodily movement as a sequencing of changes in the muscles involved in any particular bodily action. But the fact that in each case the combination of individual elementary patterns of muscle-movement is bound to alter the specific expression of the pattern in the sequence so that B at the end of a word is formed differently from B at the beginning of a word, or a grasping movement with the arm stretched out may be realised differently from a grasping movement with the arm contracted, does not in any way run counter to the basic hypothesis that complex actions or complex speech forms are built up from a flexible and smoothly integrated set of elementary units, initial patterns of articulation or bodily action, in much the same way as, in the Hubel and Wiesel approach, complex visual forms may be built up hierarchically from elementary visual units, invariant elementary visual patterns.

This Part of the present Chapter has set out to explain how it is possible and plausible that, quite apart from unity of structuring given to the Primitive Vocabulary, and beyond that to the total lexicon by the origin of the word/meaning relation in the fundamental characteristics of human behaviour, human perception and human action, the individual words in the extended Primitive Vocabulary, the structures formed from the elementary speech-sounds, can have a natural origin, a natural relation between the sound and the meaning of the word because of the underlying physiological and neurological parallelisms between the processes of visual perception, bodily action and speech. The chapter will conclude with a detailed presentation of the manner in which the sound-structure of a variety of words selected from the Primitive Vocabulary indicates in one or another way real features of the percepts, actions or qualities to which the words relate. Before this detailed presentation, it will be useful to sum up exactly what seems to emerge about the parallelisms and relationships of the physiological and neurological processes of vision, action and speech from the three sections of the chapter just completed dealing respectively with the construction of visual forms, the construction of action-forms and the construction of word- forms. The important points (which contribute both to establishing the plausibility of the equivalences between word-form, visual contour and action-contour presented in the concluding section of the Chapter and to explaining how such equivalence can be possible) can be listed as:

I The mode of construction of a form, whether a visual shape, an action-routine or a word-form, is a quite separate matter from the research issue of how such forms are recognised when they are, typically, presented as part of a continuous stream of visual information, an unbroken sequence of activity or a continuous stream of speech-sound (or more generally of sound of all types). In a rather special sense (different from that used by Chomsky) one can distinguish between 'competence' in vision, action or speech, the elements and structures which make these faculties possible, and 'performance', that is the functioning of vision, action and speech in all the special circumstances of the real world, in conditions of obscurity, distortion or confusion.

II Because one can postulate elementary units of vision, action and speech and the manner in which these elements might be formed into larger structures, visual shapes, action-routines, word-structures, this does not mean that throughout the individual's life, from infant to adult, the shapes perceived, the action-routines used or the words formulated must on each occasion be formed de novo from the elementary units. The growth of experience of vision, of action and of speech, clearly provides the individual with more efficient processes, with 'templates' and 'schemas" representing more complex visual, action or speech-structures (no doubt represented in the cortex at some level higher than that so far reached by Hubel and Wiesel's researches), which can be used directly both to increase the speed and effectiveness of perception, to allow the more rapid construction of complex chains of actions and to form whole words directly into complex sentence-forms. For vision, action and speech, there is not a choice between a feature-detection (or a feature-construction) process from elementary units and a 'template-matching' scheme, for production or comprehension; vision, action and speech will use whatever methods are most efficient and best adapted to the circumstances of actual experience. A word, a visual shape or an action-routine, can in their turns be treated as whole units, to form even more complex structures;

III What emerges from the detailed consideration of the processes underlying vision, action and speech is the overriding importance in behaviour of the organisation of movement, of the central co-ordination neurally of complex sequences of movements which may relate in a parallel way to movements of the eyes (the oculomotor apparatus) in seeing, to movements of the parts of the body and the body as a whole in the organisation of action or to movements of the vocal organs, the articulatory apparatus in speech. The highest level of control of bodily skill is now thought to be mediated by a set of motor schemas or motor programs that can be executed with a wide variety of initial positions of the muscles and organs involved and of the local environments. Vision, action and speech are all examples of highly skilled human activities, and the probability is very high that they are organised and executed in very similar ways, by neural structures which in their formation and manner of operation are closely similar. Everything points to a central role for the motor cortex, not only in relation to bodily action but also to visual perception and speech; and this coming together in the brain of the structures for planning complex action in these different forms provides a possible, and a probable, basis for the detailed interrelation between visual forms, action-forms and word-forms, which is the central hypothesis of this Chapter. Such a relation between patterning at the cortical level would allow one to understand how it is that the structure of visual shapes and action-routines, which everyone would accept as natural and not arbitrary in any sense, can he related to the structures of words, which traditionally have been thought to be wholly arbitrary, despite their inexplicably close relation in consciousness with the percepts and the actions to which they refer. Penfield has remarked that "the image of how to speak a word is really a pattern of the motor complex required to produce the word"(31) and in this precise fact one should look for the naturalness of language and for the appropriateness of the structure of the individual word for the meaning it has (the visual shape or object, the action-contour).

ANNEX TO CHAPTER IV

The structure of the illustrative Primitive Vocabulary in terms of the types of visual or action contours to which the word-structures are related; and examples for a selection of words from the Vocabulary of the specific relation between the particular word-structure and the particular associated visual or action contour

The earlier parts of this Chapter have described how the total English lexicon can plausibly be thought of as built from a Primitive Vocabulary of words known directly from the developing experience of the child or adult and not by way of explanation or definition in terms of other words. On page 61 a list was provided as an illustrative Primitive Vocabulary showing the kinds of words which might be acquired in this way. The second part of the chapter has attempted to describe how from elementary visual units, elementary units of action and elementary units of speech-sound, more complex forms are built up, visual shapes, action-routines and individual words, and has emphasised the physiological and neurological common elements underlying these processes of construction, for vision, action and speech so that it is plausible that there should be direct physiological and neurological relationships between the motor-patterns, the schemas, in the central nervous system, primarily the motor cortex, governing the production of the complex patternings of bodily movement (eye- movement, bodily action, articulation) which go to form the processes of visual perception, bodily action and speech. Chapter III had earlier proposed that this relationship, one of close equivalence, could be found at the most elementary level and at the end of Chapter III a table was presented, a scheme of detailed equivalences, between the visual units identified by Hubel and Wiesel, the elementary units of action of mobile parts of the body and the elementary speech-sounds (phonemes).

The purpose of this Annex is to bring together this general argument and evidence in a practical demonstration of the manner in which the appropriateness of individual words for their meaning, the naturalness of individual words, can be derived from the relation between their structure and the visual contours, the contours of action and other features of perception and action necessarily associated with the particular speech-sounds used and the order in which those sounds are arranged in the individual word.

This Annex does not set out to explain or justify in detail the particular way in which the sound-structures of the individual words give rise to or are associated with the particular visual or action contours which indicate their meaning. That is a lengthy matter which has been dealt with elsewhere. A careful explanation of the process by which elementary units of speech-sound are combined and matched with combinations of elementary units of bodily action or vision is contained in The Physical Foundation of Language (1973).

The Annex deals with two matters: first, it explains how the words in the Primitive Vocabulary (on page 61) can be classified in terms of visual or action contours associated with each of them, as a consequence of the particular sound-structure of the individual words, Secondly it sets out specifically for a substantial sample of words drawn from the Primitive Vocabulary the character of the visual or action contour associated with the particular word which constitutes the natural bond between the sound of the word and its meaning. Practical verification of the validity of the relationships for the individual words chosen is something that can only be taken a limited way at this point. The reader will have to rely on his own efforts to decide whether or not the relationships proposed for the individual words seem plausible and helpful.

Classification of the Primitive Vocabulary in terms of associated visual or action contours

The list on page 61 contains rather less than 500 words. The broad classes into which these are divided is already indicated in the list; beyond that, against each word there is a letter or letters set which refer to the particular type of action or visual contour associated with the sound-structure of the individual word. The following paragraphs describe more fully the meaning of these classifications and the general characteristics of the structure of the Primitive Vocabulary when one analyses it in terms of the number of words found in each class.

The first broad classification is into:

Shapes and Things
Primary Visual
Body Parts
Animate
Food etc.
Near Objects
Larger Scale
Auditory
Time aspects
Movements and actions
Primary
Movement of body parts
Arm actions
Whole body actions
Complex actions
Internal actions and states
Qualities and Relations
Visual Qualities
Other Qualities
Relational

Some notes on the interpretation of these categories may be helpful:

1. Primary visual includes words which seem to have a simple visual contour
2. Animate extends to include a few parts of animate creatures (the category body parts relates to parts of the human body)
3. Food &c includes a few parts of edible things
4. Near Objects consists largely of objects found within a room or within a house (though some items are classed here purely for convenience e.g. ship, boat, car)
5. Larger scale could be divided into two sub-categories, near landscape (things outside of but near to and on the scale of the house) and larger-scale landscape features
6. Auditory contains a few items where the mode of perception seems clearly to be auditory rather than visual
7. Time aspects this is a list which could be considerably extended but it is a difficult group which obviously needs further consideration
8. Movements and actions Most of the categories are self-explanatory. The list does (somewhat inconsistently) include a few stative verbs; where IS (AM or ARE) and similar verbs should be placed is a rather arbitrary decision
9.Complex actions These refer to actions involving instruments or specific external circumstances or goals
10. Internal actions and states It is arguable in several cases whether words in the list should be treated as referring to internal actions or internal states. Some could equally well be either (and a good deal of philosophical discussion has been devoted to the attempt to decide whether they are state or action)
11. Spatial qualities and relations These divide into positional and directional (mostly positional) but the distinction is not a clearcut one.

In terms of the numbers of words in each of the main categories, the analysis straightforwardly is as follows:

Primary Visual                          33 words
Body Parts                              34
Animate &c                              32
Food &c                                 17
Near Objects                            66
Larger scale                            40
SHAPES AND THINGS                                       222 words

AUDITORY
7 words

TIME ASPECTS
8 words

Primary movement                     17
Movement of body parts                  24
Arm actions                             34
Whole body actions                      30
Complex actions                         12
Internal actions states                 14

MOVEMENTS AND ACTIONS                                131 words

Visual qualities                     32
Other qualities                         32

All Qualities                                        64 words

Spatial relations                            39
Other relations          22

All relations                                        61 words

QUALITIES AND RELATIONS
125 words

TOTAL PRIMITIVE VOCABULARY                           486 words

The above gives a broad first analysis of the composition of the illustrative Primitive Vocabulary in terms of the common-sense classifications into which the different words in the list on page 61 have been grouped. However, the reader will have noted that against each word in the list a letter or letters has been set, and these are intended to indicate more precisely the kind of visual or action contour (or other perceptual characteristic) associated with the sound-structure of each word. Taking a couple of obvious cases, the two words SUCK and LICK are put in the category of 'movement of body parts' and have the letter A marked against them; the A simply means that the kind of contour or action associated with the sound-structure of these words is straightforward bodily action. It seems self-evident that the words SUCK and LICK, by virtue of the sound-elements of which they are formed: and the way these sound-elements are in fact produced, are very closely associated with the actual actions of sucking and licking; if one says either of the words carefully and emphatically, one finds that in the very action of saying them, one comes very close to performing the specific actions which constitute the meanings of the words. Of course, for many of the words in the list, the relation between sound and action or visual contour is less immediately apparent and the letters set against each of the words are intended as a guide in each case to the associated effect.

The following shows the significance of the different markings placed against each word in the illustrative Primitive Vocabulary:

- V this means that there is a visual contour associated with the sound-element. going to form the word, which resembles or at least gives a cue or clue to the percept to which the word refers

- A Similarly this means that there is an action or action-contour associated with the sound-structure of the word. So, for SUCK, as already referred to, the marking is A because the action of sucking is clearly associated with the saying of the word

-P This means that there is a Pointing action directly associated with the saying of the word. This is, not surprisingly, the kind of action particularly associated with words referring to parts of the body: EYE EAR NOSE

- SH :This means that the action associated with the sound of the word is an actual showing of the part referred to. For example, the simplest indication of HAND is to hold up the hand, and the sound-structure of the word is associated with that action

- VP This indicates that the contour or action associated with the sound of the word combines visual and pointing elements, with the form first indicated V for visual, appearing to be the more important

- PV :This is the reverse of the previous mark, that is the Pointing element appears to be more important than the visual element in the associated contour or action

- S :This means that associated with the sound-elements going to form the particular word is a sound typical of the thing to which the word is related e.g. most obviously the word THUNDER has associated with it a sound imitating thunder

- SR this moans that associated with the word-structure is a sound referring to the percept or thing to which the word relates e.g the word BEE has associated with it the sound of a bee buzzing (see for example BUMBLE-BEE and HUMBLE-BEE, clearly onomatopoeic word-formations in the traditional sense)

- AR Similarly this means the associated action refers to the percept or thing to which the word related. For example, associated with the sound-structure NEEDLE is not a contour indicating the shape of a needle but the action of a needle used in sewing

- PA This means that the associated contour or action for a word has both a primary Pointing element as well as more a general action element. In the case of HAT, for example, there is a combination of a pointing action with an action of putting on or taking off a hat

- VA This means that there is a combination of a visual contour associated with the word and an element of action. For example, the word FLOWER appears to have associated with it both an indication of the general shape of a flower and of the action of plucking a flower

- SV This means that there is a combination of a sound and a visual element associated with the word-structure. For example, SAND has a sound associated with it of the sea-shore (no doubt part of the typical effect of the fricative S) as well as an indication by a visual contour of an extent or surface

- CVS this means that there appears to be a complex contour or action associated with the sound-structure of the particular word and that this complex includes visual and sound elements. This seems to be the case for DOG and BIRD, and it would not be surprising that this should be so since both the visual contour of a dog or bird and the sound that each of them makes are obvious clues to the nature of the thing perceived.

- CSA Similarly this means that the complex contour or gesture includes sound and action elements. An example, again an unsurprising one, is the word HEN, where the sound element seems to be primary

- ASR This means that the contour or action has both action and sound elements which refer to the percept or thing to which the word-structure relates. So the word SNAKE has both an action element related to the movement of a snake and a sound element (apparently deriving from what Firth described as the 'sibilant' phonetic quality of the initial S)

- I This means that the action or contour associated with the word refers to am internal state or activity. For example, the word BITTER has associated with it a mouth-action which relates to the reaction to a bitter taste

- PI This means that the action or contour associated with a word has both a pointing element and an internal state or activity which is referred to. For example, the word THINK has associated with it a Pointing to the head (a very familiar gesture) as well as an awareness of an internal state or activity

- IA This means that there is a combination of an internal state or activity and an external action, both represented in the sound-structure of the word. So LOVE has associated with it an internal state or activity and an external action of clasping or holding

- AI This means that the external action component seems to be more important than the internal state or activity. For example, in the case of the word CHOOSE, the external action seems to be one of picking between objects but there is also an internal awareness of the state or activity of choosing

- ? This means that so far applications of the equivalences and rules for the combination of elementary units described in the previous book has not made it possible to arrive at any plausible identification of the contour or action &c associated with the structure of the particular word.

Following this explanation of the significance of the markings placed against each word in the illustrative Primitive Vocabulary, it is possible to make a number of supplementary points:

1.The classifications in terms of the associated visual contour or action apply in the same way to all the words in the list, regardless of whether or not they would normally be treated in grammatical terms as nouns, adjectives, verbs, prepositions, adverbs or conjunctions, &c. So, the adjective HEAVY is marked AR - that is, the word has associated with it an action reference to the meaning of the word; similarly the noun PENCIL is also marked AR, again because it has an action associated with it referring to the use of a pencil (not a visual contour of a pencil). Again, the preposition TO is marked PA, that is it combines a pointing element and an action element referring to movement in a direction, just as, already indicated, the noun HAT is marked PA because the word-structure contains both a Pointing and an action element.

2. The main categories of contour or action associated with words in the list are clearly Visual, Pointing and Action with, not surprisingly, most of the V markings relating to Qualities and Shapes and Things, most of the P markings relating to body parts, prepositions and pronouns, and most of the A markings relating to movements and actions. But the divisions are by no means clear-cut. Some words with which one might, at first sight, expect to find associated visual contours. are, on examination, found to be indicated in other ways. For example, the sound-structure of FROG does not relate to a visual contour but to a combination of an action element and a sound element. The word HEAR has associated with it a pointing action, very similar to that associated with EAR. MEET as a word is marked VA, that is the word refers to a visual indication of meeting, not to an action.

3.The small selection of words relating to Time are of particular interest. Some time-words have associated with them a pointing element, pointing forward or backward, an obvious transfer from spatial reference to temporal reference. Others seem to involve action referring to the meaning of the time-word e.g. NOW is marked AR and seems to involve an action with the arm moving sharply to point straight upward, as distinct from THEN, also marked AR, where the arm moves similarly but to point forward from the vertical. It is difficult to decide whether time-words should be marked P or AR; they seem to refer to a temporal frame of reference modelled on a one-dimensional spatial frame of reference, so that points in time are distributed along an imaginary line

4. A small class of words, but an important one, is included in the list referring to internal states and activities. In these cases, it is not just an assumption that the sound-structures of the words make this kind of reference but it is proposed that the sound-structure of the words has a specific felt relation to the internal meaning of the words. Certainly this is likely to be the case for words referring to emotions which have associated with them changes of physical state e.g. SAD, GLAD, TIRED.

5. The class of interrogative words is somewhat difficult to categorise and they have been marked I?. Associated with words such as HOW WHY WHO seems to be a specific attitude, some particular internal state of organisation.

6. Also of interest, of course, are those words where so far it has not been possible to identify the contour, action, &c associated with the sound-structure of the word with any plausibility or confidence. Such words in the list include TIME SPACE COLOUR SHADOW SHAPE FORM SOON EARTH BY WITH AND OF AMONG IS, together with a miscellaneous collection of other more specific words DRY WET CLEAN DIRTY BEGIN LOSE WAIT BIT PIECE THING STOOL CHAIR CAR BOAT SOCK COAL ASH PAPER BUTTER SUGAR SALT SHEEP - with some more significant words such as MAN WOMAN BOY GIRL MOTHER FATHER BABY BROTHER SISTER ANIMAL. Perhaps most important of these are those words referring to family relationships &c, some of the functional words and some very general words TIME SPACE THING FORM SHAPE COLOUR. Further study is obviously needed of the possible implications of the sound-structures of these words.

7. Finally, in analysing the Primitive Vocabulary, it may be informative to sort out precisely how the words in the list divide up between the different categories of associated contour, action &c which have been identified above. Out of the 486 words in the list (including the doubtful cases), 128 refer to action as the sole element or an element in the natural character of the word, 132 to visual contour and 79 to pointing or showing the percept or thing referred to. If one adds together all the words where the sound-structure seems to indicate some form of action reference i.e. including items of pointing and showing, the total is 249, that is, more than half the words in the illustrative Primitive Vocabulary derive the naturalness of their meaning from some associated action, some reference to action. The corresponding figure for words where there is some kind of reference to visual contour is 144. Comparatively, words with some association with sound are rather few, 55 in total, and there are 26 words with some internal reference. From this analysis emerges the dominant role is played by action-patterns in determining the sound-structure of words found in the Primitive Vocabulary, and of course it must be remembered that even where a word is marked as associated with a visual contour, that contour in fact is formed either by a movement of the eyes in scanning the percept or by a representation of the visual shape in terms of movement of the arm and hand. Examples drawn from the Primitive Vocabulary of the natural relation between the sound-structures of individual words and their meanings

Although the 500 or so words in the illustrative Primitive Vocabulary have been classified above in terms of the type of visual or action contour &c associated with the sound-structure þf each individual word, very little has so far been said about the specific character of the visual or action contour associated with the sound structure of individual words (though a few incidental examples have been given). This concluding section of the present Annex explains more specifically for a substantial sample of words drawn from the Primitive Vocabulary the form taken by the visual or action contour derived from the sound-structure of individual words and serving to indicate or give a clue or cue to the meaning of the particular word. The sample of words selected comes from the different main categories which have already been described, that is, words where the associated visual or action contour relates to action directly pointing to or showing, to visual shape, to internal activity or state, to sound directly, to time aspects or to complex combinations of visual, sound and action contours.

The sample of words selected in these different categories are:

Action words

Of the 15 words in the list, 9 have as their meanings movements or actions (including 2 which relate to movement of body-parts, 4 which refer to arm-actions, 2 to whole-body actions and 1 to a complex action; of the remaining 6, 4 have as their meanings shapes and things (including one animate thing) and the remaining two include 1 preposition and 1 adjective. The actions associated with the words (and derived by the processes described in Chapter III from the combination of elementary speech-sounds in the structure of the word) are as follows:

YAWN The sounds in the word combine to form in fact the action of a yawn. The reader can test for himself that saying the word slowly and vigorously produces a movement of the mouth closely resembling a naturally-occurring yawn

LICK The combination of sounds in the word again is closely matched to the action of licking; as one says the word, the tongue in effect performs a licking movement

TOUCH Chapter III shows that associated with the sound T is a stretching of the arm; the combination of T with the other two sounds simply produces a stretching forward of the as in act of touching.

PUT Similarly the action resulting from combining the in the word on the basis given in Chapter III is simply that involved in extending one's arm to put something down, to place something

PUSH Again a very straightforward representation through the sounds in the word of the action of pushing something

HIT Simply an act of hitting, the arm-action involved in HIT partially resembling that involved in IT or HE (both of which involve forward pointing movements)

GO The word does not directly model the act of going but a forward and outward gesture of the hand clearly and usually meaning GO.

SEW The associated action is simply the normal act of drawing a needle up through material and outward in a curved line

NEEDLE An action very like sewing or threading an needle, not a representation of the shape of a needle

WORM An action which refers to the wriggling movement of a worm, not a direct representation of the visual contour

POINT This directly represents the action of pointing with the index finger (the action associated with POINT resembles that associated with PUT - as the sounds also are similar in the two words)

UP This has associated a movement of the arm, so that the forearm and hand point vertically upwards

MEET The significance of the VR associated with this is seen clearly in the movement of the two hands to come together, with palms facing each other in front of the body

SAME The movement for SAME is very similar to that for MEET, except that the hands, with palms facing each other, are a little further apart

HAMMER This straightforwardly represents the action of hammering, similar to the action of hitting

For all these words, the relation between the meaning of the word (the percept, thing, action) and the associated action, derived from the sounds going to form the word, is extremely plain and straightforward and convincing. As the explanation of the formation of the action-forms from the equivalent speech-elements shows, the correctness of the equivalences between action-elements and speech-elements given in Chapter III seems to be clearly established.

The same applies to the Pointing and Showing words which are next dealt with:

Pointing and Showing Words

EAR This is an extremely simple pointing gesture, with the hand held at the level of and quite close to the ear

HEAR Almost identical (as the sound is almost identical to that for EAR). There is some aspect of bringing the hand to the position near the ear which no doubt reflects the active quality of hearing.

EYE This is a straightforward action of pointing to the eye (note that the sounds of the words EYE And EAR are close together -something seen in other languages as well)

I Exactly the same as for EYE

HAND The action is simply the holding up of the hand with palm forward, to show the hand

HAIR This is a pointing gesture, with the hand held over the hair

THAT A simple pointing gesture towards a somewhat distant object

THE A less emphatic pointing gesture

A Very similar gesture to that for THE but slightly less marked

YOU A strong forward pointing with the forefinger

IT A less emphatic pointing gesture similar to YOU

There seems no difficulty with the equivalence between these words and the pointing or showing gestures associated with them. Perhaps the only thing that strikes one is how similar the pointing gestures associated with the different words are. In the case of YOU however as compared with IT, there seems a stronger and more emphatic aspect which perhaps almost amounts to an internal component so that YOU might be marked PI, rather than just P.

Both action and pointing are more clearly seen than the visual ones or words in the next list:

Visual words

EDGE A simple word, if one represents it as EJ. The movement of the arm simply follows the contour of an edge, a short horizontal movement turning down over the edge

CIRCLE The arm moves in a circle

HEAP The arm follows the contour of a heap of something, a convex curve

CUP The action associated with the arm held forward is a cupped hand

HOOK The arm is drawn down from above the head to follow the contour of a hook

LAMP The arm is held up above the head with the hand facing upward as though holding a lamp

ARCH The movement of the arm is very similar to that for HEAP, a convex curve, traced above the head

FALL The associated contour is similar to that for EDGE, to trace the course of an object thrown and falling

WIDE The arm movement indicates width by the arms being stretched out from the sides

NARROW The arm-movement is very similar to that for MEET and SAME , the position of the hands, brought together with palms facing in front of the body, being somewhat closer together than for SAME but less than for MEET. The meaning of the action associated with NARROW clearly contrasts with that associated with WIDE

THICK Again a similar arm-movement but the palms facing each other nearly the full body width apart

THIN Almost identical with the position for NARROW

HIGH Similar to the position for UP but with the arm stretched further upwards

CLAW The associated action results in the hands being held in a claw-shape

BOUGH The associated action is the arm stretched out to the side and somewhat forward and upward, resembling the outline of a limb of a tree (a tree of course also has a 'trunk' and a 'foot')

The illustrations for the next two classes, Internal and Sound, are less obvious than the preceding ones:

Internal Pointing Words

THINK The external aspect of the action associated with the word, is the hand being drawn in so as nearly to touch the upper forehead with the fingers placed together (a very usual gesture for thinking). Besides this, there seems to be an internal state or activity associated with the sounds forming the word. The action associated with THING seems to be very similar - so perhaps a THING is something thought (see Latin REOR - think and RES - thing)

KNOW Very similar position of the hand, in relation to the head, to that for THINK, with a somewhat different Internal component

ME The hand points towards the mouth or face, with an accompanying internal feeling (which is not apparent for I)

Sound words

THUNDER This has associated with it the making of a sound clearly meant to imitate the sound of thunder. See other straightforwardly onomatopoeic words such as DING DONG, BANG, WHIZZ (and onomatopoeic words for THUNDER in other languages eg. Malay GUNTOR)

CALL This involves stretching the arm forward and outward with the sound of a call (some relation to a clearly onomatopoeic word like CAW)

SPEAK Again the stretching of the hands forward is accompanied by voice-sound

CRACK This is a combination of an action (of breaking something) with a flat sound, like SMASH

Time aspects

NOW This is associated with a short upward movement of the hand indicating spatially a time-concept

THEN A movement similar to that associated with THERE but a shorter, forward indicating movement

WHEN A more emphatic movement than for THEN, apparently with some characteristic internal aspect, aspect of feeling, similar to that for other interrogative words

BEFORE A forward indicatory action, again as a result of transfer of reference from spatial position to temporal position

Complex words

BIRD Largely a sound-word but there is also an element of visual indication, perhaps of movement of the wings

HEN This seems to combine an indication of the characteristic sound made by a hen with action reflecting the equally characteristic movement of a hen

FIRE This combines an indication of the rushing sound of a fire, with an indication of flames rising

LAKE This combines an indication of an extent or stretch of space (of water) with some sound of wind, over water

RAIN This combines an indication of the sound of raindrops, with some visual indication of the lines of rain falling.

NB: It is much harder to be sure about the actions and shapes associated with the complex words in the Primitive Vocabulary, to describe them satisfactorily or to indicate clearly how the actions, sounds &c are formed as a result of the structure. of the words - but the broad indications given above seem , as an empirical matter, reasonably reliable. There is a whole field of study to be explored of how speech-sounds indicate or model but do not imitate obviously and simply types of sound other than speech-sounds. Many names of birds appear to indicate, in the sound-structure of the words, the characteristic cry or call of the bird - but not through direct imitation which would involve using types of sound outside the accepted range of human speech-sounds, that is the available set of phonemes.