Presentations: New College Oxford
Christ's College Cambridge

[Language Origins Society. University of Pecs, Hungary. 1995]

LANGUAGE INSTINCT?
Gradualistic Natural Selection is not a good enough explanation

Steven Pinker's The Language Instinct was a comprehensive and ambitious attempts to account for the origin of language. It approached the topic from within the Chomskyan framework (Chomskyan linguists have generally remained silent about language evolution). Language, he said, was not a cultural artifact but a distinct piece of the biological make up of the brain. We would all agree that a biological and essentially evolutionary approach is desirable, though 'language instinct' already begs many questions. Chomsky's concept of Universal Grammar is well known: the brain must contain a recipe or program that can build an unlimited number of sentences from a finite list of words; the program may be called a mental grammar; children - 'grammatical geniuses' - must innately be equipped with a plan common to the grammars of all languages that tells them how to extract the syntactic patterns from the speech of their parents. However Pinker did not share Chomsky's scepticism about whether Darwinian natural selection can explain the origins of the language organ. This paper seeks to identify where, at a number of important points, Pinker's account seems unsatisfactory, for example: the idea that language could have developed, like the eye, by minute steps, under the pressure of natural selection, the idea that eventually neuroscientists will be able to locate a 'language organ', or behavioural geneticists discover a grammar gene, the postulation of a uniform distinct language of thought, mentalese, to be translated into any particular spoken language, his discussion of the arbitrariness of the sign, his account of the acquisition of language by children.


Steven Pinker's The Language Instinct was published in 1994 and was received as one of the most complete and carefully argued accounts of the evolution of language. In speaking about language as an 'instinct' he recognised that the term is no longer thought appropriate in modern biology but said that he was following in the footsteps of Charles Darwin who had described language as half-art and half-instinct; Darwin's account of the gradual evolution of instincts generally by natural selection could be applied also to the human acquisition of the capacity for language. Darwin was writing at a time when the modern science of language did not exist; he claimed no particular expertise in the discussion of language generally or of particular languages. By contrast Pinker accepts Chomsky's current theoretical account of language and particularly Chomsky's concept of Universal Grammar. The essential feature of this in its present form is termed by Chomsky the Principles and Parameters approach, that is, the underlying structures of language, the grammar, are innate and the same for all humans; different languages are the result of ascribing binary values to a small set of parameters. The simplest illustration of a parameter is the choice of Head first or Head last; depending on which choice is made, a language is either SOV or SVO with many associated orderings in other aspects of syntax. For Pinker, following Chomsky, syntax is the key productive aspect of language and lexicon is subordinate. If, as Pinker intends, one seeks to present a persuasive account of the evolution of language, it is of the first importance to settle how best language should be characterized.

In a critique of The Language Instinct (1994), there are several complexly interlocking issues. The first is whether the account of language given currently by Chomsky, and accepted by Pinker, is adequate or plausible. The second is whether a gradualistic account of the evolution of the Chomskyan language system is conceivable. The third is Pinker's treatment of related questions such as the acquisition of grammar by children, the acquisition of lexicon. The fourth is the plausibility in terms of brain evolution, brain structure and function of Pinker's approach. The fifth is, as Pinker puts it, if Chomsky rejects the idea of the evolution of language by natural selection, what alternative is there?

On the first issue, the adequacy and plausibility of Chomsky's account, books have been written and controversy rages. So a contemporary linguist, Givon (1984), speaks about Chomsky's utter disregard for the nature and significance of cross-language typological variability which allowed him, on the basis only of English syntax, to make sweeping assumptions about typological-syntactic universals; Givon accordingly rejects all the tenets of the transformational-generative tradition. Here I will only attempt to note the main features of the Chomskyan theory, as far as possible using Pinker and Chomsky's own words so that one can get a clearer idea of what system it is that Pinker believes has evolved gradually by natural selection. A preliminary observation is that for Chomsky, and for Pinker, the issues of the biological basis of language and the acquisition of language by children are closely linked; Chomsky's ideas of Universal Grammar are very much framed to account for the rapid acquisition of language by children.

3. Pinker's Ideas

There are some problems in presenting his ideas concisely and clearly. Comments on different topics, instinct, syntax, lexicon, language acquisition are scattered across the chapters; the first rather pedestrian task is to bring the related ideas together. All page references unless otherwise indicated are to The Language Instinct (Pinker 1994); I include them where the wording is Pinker's own or a very close paraphrase of it.

3.1 Chomsky's Universal Grammar

Pinker's presentation is largely contained in Chapter 4 'How Language Works' and the following points are mainly taken from that. For Pinker, Chomsky's writings are classics. Chomsky's claim that, from a Martian's-eye-view, all humans speak a single language is based on the discovery that the same symbol- manipulating machinery, underlies the world's languages. Universal Grammar is like an archetypal body plan found across vast numbers of animals in a phylum (238-9), a common plan of syntactic, morphological, and phonological rules and principles, with a small set of varying parameters. Once set, a parameter can produce far-reaching changes in the superficial appearance of the language. One of the most intriguing discoveries is that there appears to be a common anatomy in all phrases in all the world's languages. Phrase structure is the kind of stuff language is made of; traces, cases, X-bars, and the other paraphernalia of syntax are colourless, odourless, and tasteless, but they, or something like them, must be a part of our unconscious mental life (124). The universal plan underlying languages, with auxiliaries and inversion rules, nouns and verbs, subjects and objects, phrases and clauses, cases and agreement, and so on, seem to suggest a commonality in the brains of speakers, because many other plans would have been just as useful (43).

3.2 Chomsky and the acquisition of language by children

Children must innately be equipped with a plan common to the grammars of all language, a Universal Grammar, that tells them how to distil the syntactic patterns out of the speech of their parents. The unordered super-rules (principles) are universal and innate; when children learn a particular language, they do not have to learn a long list of rules, because they are born knowing the super-rules (112). All they have to learn is whether their particular language has the parameter head-first, as in English, or head-last, as in Japanese. If the verb comes before the object, the child concludes that the language is head-first as if the child were merely flipping a switch to one of two possible positions. The way language works is that each person's brain contains a lexicon of words and the concepts they stand for (a mental dictionary) and a set of rules that combine the words to convey relationships among concepts (a mental grammar) (85).

3.3 Chomsky as the starting point
3.3.1 A general comment

There are risks in taking Chomsky's current theories as the basis for an attempt to present a plausible account of the evolution of language. The most striking aspect of the history of Chomsky's linguistic theories is how rapidly and frequently they have changed over the years since his first work Syntactic Structures appeared in 1957 and the transformational-generative approach was born. Most of the key features of that approach have now been abandoned; deep structure from being the foundation of theory has shrunk and virtually disappeared, the idea of transformation has been abandoned; whilst language is still regarded, in a broad sense, as a generative process (new sentences created from a limited set of words and syntactic processes), the technicalities of generation have also disappeared; Chomsky has moved from a system which placed exclusive emphasis on syntax to one which begins to recognize the importance also of lexicon, moving from the transformational- generative approach to government and binding to principles and parameters. Specifically, Pinker explains that Chomsky wants to eliminate the idea that there is a special phrase structure underlying a sentence called d-structure, a single framework for the entire sentence into which the verbs are then plugged. The suggested replacement is to have each verb come with a chunk of phrase structure preinstalled; the sentence is assembled by snapping together the various chunks. Pinker comments that 'deep structure' is a prosaic technical gadget in grammatical theory, not what is universal across all human languages; many linguists - including, in his most recent writings, Chomsky himself - think one can do without deep structure per se. In a recent interview (Grewendorf 1993), Chomsky was asked whether generative grammar had gone astray at some point; he admitted that in retrospect there had been some wrong turnings and that a really significant change took place about 1980; this, unlike earlier work in generative grammar, constituted a major break and dispensed entirely with both rules and constructions which he described as 'taxonomic artifacts' of early generative grammar; there have been a lot of changes in the theory since 1980.

3.3.2 Pinker's own problem with Chomsky

Whilst accepting Chomsky's current principles and parameters approach Pinker makes some effort to distance himself from Chomsky, no doubt partly because he does not wish to be committed to deriving Chomsky's concepts in detail from evolutionary natural selection (he makes no attempt to do this) but also because there is the major difficulty that Chomsky himself has consistently rejected the idea that language could have evolved by natural selection. Pinker says that Chomsky's arguments about the nature of the language faculty are based on technical analyses of word and sentence structure, often couched in abstruse formulations; his discussions of flesh-and-blood speakers are perfunctory and highly idealized; "Chomsky's theory need not be treated ... as a set of cabalistic incantations that only the initiated can mutter" (104). Pinker admits to being deeply influenced by Chomsky. "But it is not his story exactly... Chomsky has puzzled many readers with his skepticism about whether Darwinian natural selection ... can explain the origins of the language organ he argues for" (24). Chomsky and some of his fiercest opponents agree on one thing, that a uniquely human language instinct seems to be incompatible with the modern Darwinian theory of evolution, in which complex biological systems arise by the gradual accumulation over generations of random genetic mutations that enhance reproductive success." (333) Chomsky thinks that "to attribute this development to 'natural selection' .. amounts to nothing more than a belief that there is some naturalistic explanation for these phenomena... it is not easy even to imagine a course of selection that might have given rise to them" (354).

3.4 Language organ and/or language instinct?
3.4.1 The analogy with the evolution of the eye

Rejecting Chomsky's scepticism, Pinker suggests that language should be considered as an evolutionary adaptation like the eye, its major parts designed to carry out important functions. Chomsky speaks about the 'language organ' and language is assumed to be a distinct brain module. The evolution of the eye was a central debating point between proponents and opponents of Darwinian natural selection; Darwin himself offered a gradualistic account which has been taken up and refined many times, most notably recently by Richard Dawkins (1986). A plausible account has been given of how even a rudimentary eye could be selected and increase the fitness of the individual in whom the advance took place; each small improvement in the functioning of the eye would promote survival of the individual and the individual's offspring carrying the gene for the improved eye. Maynard Smith and Szathmary in their recent The Major Transitions in Evolution (1995) also pick up the analogy between development of the eye and the development of language. This is perhaps not surprising since admittedly they relied heavily on two sources, Bickerton's (1990) book and Pinker and Bloom's 1990 article which foreshadowed The Language Instinct.

3.4.2 The analogy with the evolution of instincts

Language is a distinct piece of the biological makeup of the brain (18). The universality of complex language is the first reason to suspect that language is the product of a special human instinct. If language is an instinct, it should have an identifiable seat in the brain, and perhaps even a special set of genes that help wire it into place (45-46). If language is like other instincts, presumably it evolved by natural selection, the only successful scientific explanation of complex biological traits (354), the only alternative (Pinker's emphasis) that can explain the evolution of a complex organ like the eye or like language. Each step in the evolution of a language instinct, up to and including the most recent ones, must enhance fitness (366- 367), the gradual accumulation over generations of random genetic mutations that increase reproductive success (333). The language instinct is composed of many parts: syntax, with its discrete combinatorial system building phrase structures; morphology, a second combinatorial system building words, a capacious lexicon; a revamped vocal tract; phonological rules and structures; speech perception; parsing algorithms, learning algorithms (362).

3.4.2.1 Pinker on the evolution of language by natural selection

Chapter 11 is entitled The Big Bang. Pinker admits that there are genuine problems in reconstructing how the language faculty might have evolved by natural selection (365), problems particularly with any gradualistic account. The problems relate both to finding a plausible account for the transitional stages between the first human articulations and the complexities of language as it exists, in all its varieties, today and also to accounting, in biological and genetic terms, for the acquisition of language by children.

To attribute the basic design of the language instinct to natural selection is not to indulge in just-so storytelling (364). Possibly there was the first grammar mutant, that is the first individual undergoing a genetic change which produced some capacity, however limited, for syntax; the neighbours could have partly understood what the mutant was saying just using overall intelligence. If a grammar mutant is making important distinctions that can be decoded by others only with uncertainty and great mental effort, it could set up a pressure for them to evolve the matching system that allows those distinctions to be recovered reliably by an automatic, unconscious, parsing process (365). Selection could have ratcheted up language abilities by favouring the speakers in each generation that the hearers could best decode, and the hearers who could best decode the speakers. Intermediate grammars are easy to imagine (366). Pinker suggests, following Bickerton, that the languages of children, pidgin speakers, immigrants, tourists, aphasics, telegrams, and headlines show that there is a vast continuum of viable language systems varying in efficiency and expressive power, exactly what the theory of natural selection requires (366). However " Bickerton makes the jaw-dropping additional suggestion that a single mutation in a single woman, African Eve, simultaneously wired-in syntax, resized and reshaped the skull, and reworked the vocal tract" (366). Syntax is a Darwinian 'organ of extreme perfection and complication' (124).

For the origin of language, in all its complexity, Bickerton's suggestion is as improbable as the idea (advanced by Hoyle as a criticism of evolutionary theory and discussed by Richard Dawkins) that hurricanes might by chance assemble a jetliner from a scrapyard containing the aircraft parts. Stone Age people have been found with high-tech grammars (409). If the first trace of a protolanguage ability appeared in the ancestor at the split between chimps and human branches there could have been on the order of 350,000 generations between then and now for the ability to have been elaborated and fine-tuned to the Universal Grammar we see today. Language could have had a gradual fade- in. There were plenty of organisms with intermediate language abilities but they are all dead (345-346). The utility of language development is obvious; people everywhere depend on cooperative efforts for survival, forming alliances by exchanging information and commitments; this puts complex grammar to good use (368). "But could these exchanges really have produced the rococo complexity of human grammar?" (368) A cognitive arms race could easily propel a linguistic one. In all cultures, social interactions are mediated by persuasion and argument (368). Anthropologists have noted that tribal chiefs are often both gifted orators and highly polygynous; this is how linguistic skills could make a Darwinian difference in a world in which language in relationships played a key roles in individual reproductive success (369).

3.4.2.2 Children's acquisition of language

Grammar: Pinker comments on "the mystery of how children's grammar explodes into adultlike complexity in so short a time" (112). "Do grammar genes really exist or is the whole idea just loopy?" (322). Children's rapid acquisition of syntax is possible because they are born with the super-rules hard-wired into their brains; all the child has to do is to attach the right values to the parameters which determine what the structure of the local language is by listening to the speech of their parents (22).

Lexicon: The other startling aspect of children's acquisition of language is the acquisition of the words of the local language. One extraordinary feature of the lexicon is the sheer capacity for memorization that goes into building it (typically more than 60,000 words) (149). Preliterate children must be lexical vacuum cleaners, inhaling a new word every two hours, day in, day out (151-152). A name is rapidly acquired because of the harmony between the mind of the child, the mind of the adult, and the texture of reality (157). Somehow a baby must intuit the correct meaning of a word; humans are innately constrained to make only certain kinds of guesses about how the world and its occupants work. The word-learning baby has a brain that carves the world up into discrete, bounded, cohesive objects and into the actions that they undergo; the baby forms mental categories that lump together objects that are of the same kind; babies are designed to expect a language to contain words for kinds of objects and kinds of actions - nouns and verbs (153). There really are things and kinds of things and actions out there in the world, and our mind is designed to find them and label them with words (153). Since word boundaries do not physically exist, it is remarkable that children are so good at finding them (267).

Each person's brain contains a lexicon of words and the concepts they stand for (a mental dictionary); the mental dictionary seems like nothing more than a humdrum list of words, each transcribed into the head by a dull-witted rote memorization (126). A word is a pure symbol; the relation between its sound and meaning is utterly arbitrary (151-2) , a wholly conventional pairing of a sound with a meaning. 'Dog' means dog only because every English speaker has undergone an identical act of rote learning in childhood that links the sound to the meaning (83-84).

4. Critical Issues

The above conflation of extracts from The Language Instinct show some of the difficulties that any theory of the gradual evolution of language has to face. I would pick out as the key issues, both for the evolution of language and for the directly related question of the biological basis for children's acquisition of language:

1. The genetic basis of the language capacity. He posits language evolution by minimal steps with, in some sense, the existence of grammar genes and grammar gene mutations.

2. The relation of inclusive fitness and language evolution. How if the total system of language, and of languages, evolved by minimal changes, minimal additions to the system, could these minimal changes have increased the fitness (reproductive success) of the individuals who first manifested them?

3. Language as a property of the social framework, of the individual only as a member of the group. A minimal change in language by an individual is of no value unless it is a change which is shared in comprehension and production by other members of the group.

4. 'Gradual' evolution of grammar (syntax). Pinker and Chomsky concentrate on syntax in the narrow sense of phrase structure. They underrate the complexities of grammar in a more traditional sense, the complicated tense, declension and classification systems of many languages. In these languages refinements of use and meaning are achieved through lexical variation, complex modifications of root-words or through individual function words.

5. The source of the lexicon. Pinker and Chomsky treat the remarkable evolution of the lexicons of many different languages as a relatively trivial matter but it is not.

6. Acceptance of the lexicon within the group. Evolution of the lexicon means both the addition of new words to refer to new aspects of the natural or social environment and the modification of words to express different relationships between one word and another. If, as Pinker (following Saussure) suggests, all this lexical evolution is arbitrary, with no relation between the sound-structure of a word and its meaning, then by what process can the lexical evolution have taken place since it must depend on the adoption of the arbitrary word or word-form by members of the speech-group?

7. Children's rapid acquisition of the lexicon. Acquisition of lexicon by children is described as a simple matter of rote-learning, or alternatively as a pre-ordained matching of labels to pre-existing neurally-based concepts in the infant.

8. The relation between language evolution and brain evolution, whether the neural basis of the language capacity is a single module (language as a brain organ).

4.1 Language as instinct or organ?

Instinct: Pinker comments that to use the term 'instinct' in relation to language is 'quaint'(18) but it is worse than quaint. It immediately establishes a misleading picture of the nature of the language capacity. No doubt Darwin spoke about language as being half-art and half-instinct, that is, language was not an instinct like those observed in bees, rabbits or birds. To call any aspect of human behaviour instinct can mean no more than that it has a biological basis, a genetic basis. The confusion is made worse when language is also treated as an organ like the eye. An organ is not an instinct but a structure. Both 'organ' and 'instinct' are misleading when applied to language. Language is both physiologically and neurologically based; it depends on the articulatory structures for producing speech-sound, on the brain for the muscular control of those structures and for the relation between the speech-sound and the percept or action to which the speech-sound relates. The idea that language is an instinct and thus must have a specific seat in the brain suggests that the language capacity must be localised in some area of the brain; evidence from recent research using PET and MRI brain- imaging (see, for example, the PET images in chapter 5 'Interpreting Words' in Images of the Mind Posner and Raichle 1994) that many parts of the brain are involved in the production of a spoken utterance; different areas of the brain are activated for different aspects of speech. Whether or not 'instincts' in general can be traced to a special set of genes, and this appears questionable even if one accepts the loose use of the term instinct, the idea that the whole of what is required for language could ever be traced to a limited number of genes seems totally implausible.

Organ: Pinker treats language as a complex organ like the eye. Language is not an organ and is not like the eye. The eye is a diversified physical structure which is of no use without the neural connections within the brain which interpret the patterns of light falling on the retina. If Pinker had said that language is like the visual system as a whole, or like the perceptual capacity as a whole, including the brain connections which integrate these systems, this would have been more plausible. What is more like the eye is the whole articulatory system with the auditory structures but these are structures which evolved to serve functions completely distinct from language; the new aspect of language is the use the brain makes of these pre-existing structures.

4.2 Language complexifying by natural selection?

When Pinker says flatly that natural selection is the only successful explanation and that natural selection is gradualistic, this is too sweeping. In the Darwinian system, natural selection is seen as the operative force which has quite recently been given a more precise application by the introduction of the concept of inclusive fitness, the genetic interpretation of Darwin's original concept. However, in the case of humans there can also be cultural selection, behavioural selection at the group level, where the patterns of behaviour adopted are not tied to individual genetic differences. Even in terms of genetic evolution, natural selection does not simply mean that every system, every aspect of behaviour and use of any structures, must be the result of gradual change, genetic change, directed solely to that use or behaviour. Darwin himself recognised that complex structures which evolved by gradualistic natural selection, could in different environments be put to new uses quite different from those which gradual natural selection had first produced. The example Darwin gave was the transformation of the swim-bladder into the lung; whether the lung or the swim-bladder came first and was in some fish converted into the other, the point remains the same: that a complex structure developed to serve one function was transferred to serve a quite different function. Other examples can be found in the development of limbs for locomotion into wings for flying or fins for swimming, of muscles into electric organs in some fish, of gills into the structures of the ear, of the tracheae into wings in insects. The essential point is that complexity developed for one function could come to serve as complexity for a quite different function. More significantly, in thinking about language and indeed other functions, the neural organization for supporting one function, e.g. respiration, swallowing, mastication, locomotion, became adapted to serving the new function, e.g. speech-sound production, flying, swimming.

4.3 Language complexity driving inclusive fitness?

Pinker says that each step in the evolutionary development of language must have enhanced fitness; he has to say this to attribute the evolution of language to gradualistic natural selection of language as a distinct organ, instinct or function. That the advance of language from the most primitive articulation to the perfection of the systems of fully developed languages was due to its contribution to fitness seems unlikely. Fitness as a concept applies to the individual and not to the group or society; fitness depends on genetic change in the individual. Pinker makes no more than a rhetorical attempt to justify the idea that even the limited aspects of language on which he and Chomsky concentrate could have brought added fitness to any individual. If one confines attention to the evolution of grammar alone, it is hard to believe that the complexities of the Greek, Russian or German grammatical systems, the development of the subjunctive, the middle and passive voices, perfective and imperfective aspects of the verb, case systems, could be the product of minute changes resulting from genetic mutation, could be linked to genetic change in an individual which increased that individual's reproductive success, that individual's inclusive fitness.

Pinker recognises that his approach may be criticised as Just-So story telling and denies that this is so. However at many points this is exactly what it is. He antedates the first rudiments of language as far back as the split between the human and chimpanzee lines, making available he says genetic change over some 350,000 generations for the refinement of the language capacity. He speaks about the first 'grammar mutant'. What conceivably would the first grammar mutant be, what genetic mutation on Pinker's scheme would account for this? This seems a phrase for which Pinker provides no content. Whatever 'grammar mutant' may mean, a mutation in an isolated individual could have no more effect on the social development of language than mutation producing any other abnormality. He suggests that neighbours of grammar mutants would come to understand through using their overall intelligence, but they could not decode the behaviour of the grammar mutant unless they already had brains adapted to appreciate the refinement the grammar mutant was producing. This is a quite superficial attempt to tackle what over the centuries had been seen as the great problem about any suggestion that language, or any aspects of language, could have been invented by an individual; language and languages are social constructs, not the special capacity of any individual. Neighbours (without language capacity) could no more by general intelligence decode the grammar mutant than we could by general intelligence decode the meaning of bird-song. Pinker goes further and makes the extraordinary suggestion that the grammar mutation in the individual would create pressure on the neighbours to evolve a matching system, a parsing system which would enable them to comprehend, and use, the genetic language change in the first individual. This offers the bizarre picture of the individual in whom the capacity for the subjunctive developed, using this capacity, transmitting it over generations to his offspring, whilst others, at the incredibly slow pace of genetic change over generations, evolved to the point where they in their turn were able to use the subjunctive; it is equally implausible to think that a similar process could apply to other grammatical aspects, the refinements of the case system, the development of modal forms and so on. Neighbours could not by general intelligence bring about genetic change in themselves as a basis for an improved language capacity. Pinker suggests that speakers that others could best decode would be favoured but, ex hypothesi, others would not have undergone the genetic changes required to make use of the advances in language made by the best speakers. He makes the point that language would be socially useful, social interactions are mediated by persuasion and argument, more specifically that complex grammar would be put to good use in exchanging information and commitments. Obviously language is useful for humans in social interaction but whether the complex grammatical structures are required or give any particular added evolutionary benefit to the individual or the group seems quite uncertain. Language, he says, could have advanced as a result of a cognitive arms race propelling a linguistic arms race; a cognitive arms race presumably means that more intelligent, more perceptive, more creative individuals survived and achieved greater inclusive fitness, that is, their more intelligent, perceptive and creative children also survived and they competed among themselves in generating beneficial changes in language. What exactly is meant by 'a linguistic arms race'? An Oxford or Cambridge debating society? A parliament? An election campaign? Applied to communities of hunter- gatherers, or warring tribes (depending on which view one takes of the early social states of human beings) the idea seems improbable. Pinker then produces his crowning suggestion, that tribal chiefs are judged by anthropologists to be both polygynous and gifted orators and accordingly one might suppose they achieved the reproductive success through their superior linguistic skills. The instruments of tribal chiefs in achieving reproductive success, or success in other ways, are not their language skills but others such as size, physical strength, rapidity and ruthlessness of action, kinship.

4.4 Intermediate grammars?

In saying that grammars of intermediate complexity are easy to imagine, Pinker adopts Bickerton's suggestion (also taken up by Maynard Smith and Szathmary) that grades of protolanguage might resemble pidgin, the speech of tourists, immigrants, aphasics, wolf-children. Pinker makes no attempt to give any specific illustration of this. The deviant forms result from the degeneration of already existing fully developed languages; to suppose that in the total absence of structured language these forms could come into existence is highly unlikely. The fact that, as Pinker says, Stone Age people may have 'high-tech grammars' (presumably meaning elaborately-structured forms) speaks against rather than for the idea that language may have evolved as a distinct function (instinct, organ) by gradualistic natural selection. How could these 'primitive' peoples by genetic change evolve their high-tech languages, when in many cases they did not succeed in evolving a number system going beyond One, Two, Three? Were their chiefs busy developing case and tense systems, phrase structures, word-order, nouns, verbs, prepositions, conjunctions? Pinker suggests that the gradual evolution of language may have taken place over an extremely long period allowing for 350,000 generations of random genetic change producing minor changes in language. To produce complex and refined languages over this period by a succession of random genetic changes seems quite as implausible as the suggestion by Bickerton that the language capacity might have been the result of one massive super-mutation in a single individual (African Eve). Because no plausible account can be given of the step-by-step growth of language by random genetic change, Pinker proposes that there must have been thousands of organisms, and presumably thousands of speech-communities, with intermediate capacities between animal grunts and the Greek, Chinese, or German languages, but sadly all these organisms and communities disappeared without trace. There is no evidence for this and it remains the sheerest speculation.

4.5 Evolution of language acquisition by children?

Pinker is more specific about the capacities the gradual evolution of language by natural selection should have provided to enable children to acquire the local language as rapidly as they do. He suggests that children's brains must be structured from birth in particular ways, first by having grammatical super-rules hard-wired, with provision for a few parameters to be given specific values derived from the child's exposure to the local language, and secondly by having the ability to attach the locally-correct labels to the objects and actions of the local environment. Pinker does not discuss exactly what super-rules are hard-wired and how this hard-wiring might be produced in neural organisation. As regards parameter setting, the suggestion by Pinker (and Chomsky) is that children decide whether the local language is SOV or SVO by observing whether the verb comes first or last in the sentence - but how do children know what word is a verb and what word is not? The treatment of the acquisition of lexicon by the child is no more satisfactory than the discussion of the growth of lexicon as part of the evolutionary process. Children's acquisition of lexicon, of the names for things and actions, can, he says, only be the result of rote-learning of the arbitrary relation between a unitary pattern of speech-sounds (the word) and a discriminable object or action. But children do not learn new words by rote memory; rote memory would mean that their mothers tell them ten times that 'that creature is a dog'; learning words by children is not like learning the multiplication table, learning telephone numbers, learning the Kings and Queens of England or learning the catechism (for none of which children show any special ability). Children learn new words incredibly rapidly in a way which adults cannot match. Some other explanation than rote learning is required to account for the acquisition of lexicon by children (and there are similar problems about the acquisition of an ever-growing lexicon by primitive adults in the gradualistic evolutionary scenario). Pinker notes that word boundaries do not physically exist (in the instrumental record of utterances) but are found by children; he rightly regards this as remarkable but attempts no explanation how it is possible. There is parallel problem of the absence of sharp concept boundaries to which words are to be related, the gavagai problem discussed by Quine (Quine 1960) which Pinker attempts to deal with. Pinker sees no real alternative to explain the rapid acquisition of words by children to the idea the human mind is designed to find objects and label them; there is a pre- established harmony between the mind of the child and the texture of reality (157); the child somehow has the concepts available before experience with language and is basically learning labels for them. What exactly does this mean? How can the pre-existing harmony exist between concepts which vary between physical and cultural environments and words which vary between the many different languages? Some explanation is needed but pre-existing harmony, pre-existing conceptual structure, is not a clear or persuasive or testable suggestion.

5. Summing up

Pinker's The Language Instinct does not offer a satisfactory account of language evolution. Piattelli-Palmarini (1994: 339) says that Pinker's account (developed with Paul Bloom) is the best, yet still unconvincing, adaptationist reconstruction. It aspires to give an account of the evolution of language but is marred by its concentration on the Chomskyan account of the nature of language, an account constructed largely on the basis of the English language which ignores or downplays the lexical and syntactic complexities of other languages. There is much little-examined speculation, about the time when the first rudiments of language might have emerged, about the manner in which children can acquire lexicon, about the anthropological basis for improved language capacity, about the role of genetic mutation in bringing about changes in language structures, about the possibility of survival benefit for the individual flowing from mutations affecting language competence and performance. The principal error is the failure to treat adequately the social character of language, as a possession of the group, of the speech community, and not simply of the individual. Language development and change as increasing the inclusive fitness, that is serving the long-term reproductive success of the individual, is simply not a plausible proposition, prehistorically or historically. The other major error is concentration on the evolution of syntax, grammar (in a narrow Chomskyan sense or in a broader more traditional sense) and treating superficially the vital role of the development of lexicon, both as a representation of the perceived world and as an instrument for syntactic manipulation of utterances through function words, inflections etc. The final and perhaps most important error is a mistaken view of natural selection as limited to gradualistic change in a complex structure serving a specified function; natural selection also operates through serendipitous transfer of complexity developed for one function to a new function, typically the move from swim-bladder to lung, from webbed foot to wing, from gill to structures of the ear and so on. The root problem with Pinker's presentation is the imprecise, metaphorical or rhetorical use of terms: organ, instinct, natural selection.

6. Another direction?

In this paper I have presented a summary account of Chomsky's approach to language which is the foundation for Pinker's treatment of language evolution in The Language Instinct . Next I have brought together, in his own words, the main points from Pinker's exposition and proposed criticisms both of the general approach and of a number of specific points. I have mentioned only very briefly the identification of language as the fifth major transition in evolution in the recent book of Maynard Smith and Szathmary. I have also referred to Givon's dismissal of the entire Chomskyan position and to the characterisation by Piattelli-Palmarini (a convinced Chomskyan) of Pinker and Bloom's adaptationist account as the best so far but still unacceptable. If both the Chomskyan approach to language and the gradualistic account of language evolution are rejected, what alternatives are there? Any theoretical approach to language has to go wider than phrase structure and cope with the elaborated systems of grammar and lexicon found in many world languages.

Is there then nothing of value to be extracted from The Language Instinct or from the largely derivative account of language evolution given by Maynard Smith and Szathmary? There may be something of value when they consider how language might be represented in the brain. Here some of Pinker's incidental remarks, and suggestions by Chomsky and his other followers, may offer clues to a more plausible approach to the evolution of language:

Chomsky: It has also been suggested that the properties of language derive in some fundamental way from properties of the visual system (Grewendorf 1994: 391) These skills may well have arisen as a concomitant of structural properties of the brain developed for other reasons (quoted by Pinker 1994: 362). Organs develop to serve one purpose and, when they have reached a certain form in the evolutionary process, became available for different purposes, at which point the processes of natural selection may refine them further for those purposes. (Chomsky 1988: 167)

Pinker: Language could have arisen, and probably did arise, from a revamping of primate brain circuits that originally had no role in vocal communication); it is the precise wiring of the brain's microcircuitry that makes language happen; brains can be rewired only if the genes that control their wiring have changed. The ancestral brain could have been rewired only if the new circuits had some effect on perception and behavior (350, 364).

Maynard Smith and Szathmary : there is not only a formal similarity between the construction of sentences and the performance of manual tasks, but there may be a common physiological basis for the two abilities; one could suppose that language is a spandrel, that is, an unselected by-product of design for some other purpose; there is a formal similarity between 'action grammar' and protolanguage. (Maynard Smith and Szathmary 1995: Chapter 17)

Bickerton: True language had to wait on a change in neural organisation that caused us to slot meaningful symbols into formal structures and to do so quite automatically; the capacity to construct sentences could in principle have derived from some previously established function, unlikely, however, unless there already existed some structure and/or function preadapted for syntax, so that syntax simply utilised existing neural structures. (Bickerton 1990: 130-131)

The idea that language may have been modelled on or directly derived from pre-existing brain systems has been explored by a number of writers. The possibilities include modelling on tool use (Greenfield and others), modelling on the visual system (Givon), modelling on throwing action (Calvin), modelling on motor control (Studdert-Kennedy, Lieberman, Allott). The earliest suggestion on these lines was by Karl Lashley (1951) who discussed the generality of the problem of syntax and drew attention to the parallels between the syntax of language and the syntax of action; there has since been considerable discussion of the grammar of action and of the grammar of vision Richard Gregory (1976). It is not possible in this paper to present these alternatives at any length but the following paragraphs briefly describe them.

Greenfield's 1991 paper "Language, tools and brain: The ontogeny and phylogeny of hierarchically organized sequential behavior" postulated an evolutionary homologue of the neural substrate for language production and manual action which provided a foundation for the evolution of language before the divergence of the hominids and the great apes. The role of toolmaking as a precursor for or as coevolving with language has been extensively discussed. Perhaps it should be treated as the first approach to investigating the relation between language and the cerebral motor control system.

Studdert-Kennedy suggested that linguistic structure may emerge from, and may even be viewed as, a special case of motoric structure, the structure of action. For language, the goal is to derive its properties from other, presumably prior, properties of the human organism and its natural environment; we should try to specify the perceptual and motor capacities out of which language has evolved; evidence from brain stimulation (notably the work of Kimura, Ojemann and Mateer) almost forced the hypothesis that the primary specialisation of the left hemisphere is motoric rather than perceptual; language would be drawn to the left hemisphere because the left hemisphere already possessed the neural circuitry for control of the fingers, wrists and arms, precisely the type of circuitry needed for control of the larynx, tongue, velum, lips and of the bilaterally innervated vocal apparatus (1983: 5, 329).

Ojemann and Mateer (1979, 1991) identified common cortical sites for sequencing motor activity and speech; language arises at least in part in brain areas that originally had a predominantly motor function; the development of language seems to have incorporated brain mechanisms originally developed for motor learning.

Givon in his 1994 paper for the Berkeley meeting of the Language Origins Society took the system of visual perception as the basis on which language emerged in a process of coevolution; in this the evolution of language was linked directly to the development of the visual system. He discussed the correspondences between visual and linguistic information and suggested that language processing piggybacked on visual processing; in evolution there had been an early co-existence of auditory-vocal and visual-gestural codes; the rise of visual-gestural coding provided a neuro-cognitive preadaptation for a shift to audio-oral coding because of the adaptive advantages it offered, freeing the hand and body for other activities, transcending the immediate visual field. He developed these ideas in the light of recent evidence from PET scans and otherwise of brain localisation of particular aspects of language processing in relation to visual and auditory brain organization.

Lieberman (1984, 1991) has presented a motor theory of the origin of syntax. According to this, the evolution of speech and language follows from Darwinian processes; organs that were originally designed to facilitate breathing air and swallowing food and water were adapted to produce human speech. The development of language was an instance of the mechanisms of preadaptation which besides examples such as swim-bladders and lungs, produced the sometimes surprising preadaptive bases of various specialized organs, for example, milk glands from sweat glands, the bones of the mammalian middle ear from the joint of the lower jaw. The initial stage in the evolution of the neural bases of human language appears to have involved lateralized mechanisms for manual motor control, facilitating precise one-handed manual tasks. Brain mechanisms that allow the production of the extremely precise complex muscular manoeuvres of speech, the most difficult motor control task that humans perform, may have provided the preadaptive basis for rule-governed syntax which may reflect a generalisation of the automatic schema first evolved in animals for motor control in tasks like respiration and walking. A change in brain organization that allowed voluntary control of vocalization is the minimum condition for vocal communication.

Calvin (1989) has argued the case for an even more specific preadaptation for the neural machinery underlying language in the neural circuitry required for planning sequential hand-movements such as hammering and throwing. Since hand-arm sequencing circuitry in the brain has a strong spatial overlap with where language circuitry is located in the left brain, perhaps the same massively-serial architecture can do double-duty for language and planning ahead. The well-formed sentence and the reliable plan of action have some strong analogies to more familiar darwinian successes, a matter of what Charles Darwin called 'conversion of one function to another' or metamorphosis of function. To describe the original function from which the conversion of function was made, the better word is exaptation because of the 'preconceived' connotations of preadaptation. A given piece of anatomy can have more than one function. The conversion of function, Calvin argues, is an excellent candidate for how beyond-the-apes language abilities originated. Hominid-to-human language is a 'free' secondary use of neural sequencing machinery that was primarily shaped by the food-acquisition uses of ballistic movement skills.

The motor theory of language evolution and function proposes as a universal principle that the structures of language (phonological, lexical and syntactic)were derived from and modelled on the pre-existing complex neural systems which had evolved for motor control, the control of bodily activity. Motor control at the neural level requires pre-set elementary units of action which can be integrated into more extended patterns of action - neural motor programs. These in turn have to be linked to and integrated with one another by 'syntactic' neural processes and structures. On this theory, given that speech is also essentially a motor activity, language made use of the elementary pre-set units of motor action to produce the equivalent phonological units (phonemic categories); the neural programs for individual words were constructed from the elementary units in the same way as motor programs for bodily action are formed from them (in both cases a neural program is formed in direct relation to the perceived structure of the external world); the syntactic processes and structures of language proper were modelled on the 'syntactic' rules of motor control.

Chomsky, Pinker and Bloom, Piattelli-Palmarini argue against preadaptation on the basis of the visual or motor systems on grounds which are directly related to their perhaps idiosyncratic formal analysis of language, with its emphasis on syntax. So Piattelli-Palmarini (in his 1994 paper which rejected Piaget's view that language was derived from or related to motor schemata) said that the form of linguistic principles is very specific e.g. c-command, X-bar, PRO, projection of a lexical head, trace of a noun-phrase etc. and went on to say that there is no hope, not even the dimmest one, of translating these entities, these principles, and these constructs into generic notions that apply to language as a 'particular case'; nothing in motor control even remotely resembles these kinds of notions; concrete linguistic examples (drawn from Chomskyan theory) make it vastly implausible that syntactic rules could be accounted for in terms of sensorimotor schemata (Piattelli-Palmarini 1994: 324). Chomsky in Language and Problems of Knowledge said the visual system is unlike the language faculty in many crucial ways; though there are some similarities in the way that the problems can be addressed, in relation to vision and language, the visual faculty does not include the principles of binding theory, case theory, structure dependence, and so on. The two systems operate in quite different ways (Chomsky 1988: 159, 161).

7. Conclusion

Pinker, Chomsky and Piattelli-Palmarini, in rejecting a preadaptive or exaptational basis for the evolution of language in the visual or motor systems of the brain because it is impossible to see how such as a basis could accommodate the formalisms of transformational-generative grammar, government and binding, or principles and parameters, ignore the unwelcome possibility that there is something fundamentally wrong with the linguistic theories, not with the Darwinian process by which there can be conversion of function from an already existing complex neural system for perception or action to serve as the basis for speech and language function. Chomsky is left in the awkward position of being unable to conceive of a Darwinian origin for language even though he asserts that it must have a biological basis; this leads Pinker to propose a gradualistic account of language evolution as the product of a series of minimal genetic and language changes, which is implausible in accounting for the step-by-step accretion of the elements required for Chomskyan phrase-structure theory, and even less plausible to account for the development of other complex grammatical and lexical features of world languages. The way out of the impasse is to see the evolution of language as a system founded on, reflecting and expressing the pre-existing complexities of the perceptual and motor systems of the brain.

REFERENCES


Allott, R. 1989. The Motor Theory of Language Origin. Lewes: Book Guild.
Allott, R. 1992. The Motor Theory of Language: Origin and Function. In Language Origin: A Multidisciplinary Approach ed. by Jan Wind et al. NATO ASI. Dordrecht: Kluwer Academic Publishers.
Allott, R. 1994. Gestural equivalence (equivalents) of language. Paper for Language Origins Society meeting, Berkeley.
Allott, R. 1995a. Motor theory of language origin: The diversity of languages. In Studies in language origins Vol. 3 ed. by Jan Wind et al., pp. 125-160. Amsterdam: John Benjamins.
Allott, R. 1995b. Syntax and the Motor Theory of Language. In Syntactic Iconicity and Linguistic Freezes ed. by M.E. Landsberg. Berlin: Mouton de Gruyter.
Bickerton, Derek. 1990. Language and Species. Chicago: University of Chicago Press.
Calvin, William. 1992. Evolving mixed-media messages and grammatical language: Secondary uses of the neural sequencing machinery needed for ballistic movements. In Language Origin: A Multidisciplinary Approach ed. by Jan Wind et al. NATO ASI. Dordrecht: Kluwer Academic Publishers.
Chomsky, N. 1957. Syntactic Structures. The Hague: Mouton.
Chomsky, N. 1988. Language and Problems of Knowledge. Cambridge, Mass.: MIT.
Darwin, Charles. 1859. The Origin of Species. Intro by L. Harrison Matthews. London: Dent Reprint 1971 6th edition 1882
Darwin, Charles. 1871. The Descent of Man. Reprint with Introduction by J.T. Bonner and R.M. May. Princeton, N.J.: Princeton University Press.1981
Dawkins, Richard. 1986. The Blind Watchmaker. Harlow, Essex : Longman.
Givon, T. 1994. On the Co-evolution of Language, Cognition and Neurology. Paper for 10th meeting of Language Origins Society, Berkeley.
Givon, T. 1984. Syntax Vol. I. Amsterdam: John Benjamin.
Greenfield, P.M. 1991. Language, tools and brain: The ontogeny and phylogeny of hierarchically organized sequential behavior. Behavioral and Brain Sciences 14: 531-595.
Gregory, R.L. 1976. Concepts and mechanisms of perception. London: Duckworth.
Grewendorf, Günther. 1994. Interview with Noam Chomsky. Linguistische Berichte 153: 386-395.
Kimura, D. 1976. The neural basis of language qua gesture. In Studies in Neurolinguistics. ed. by H. Whitaker, pp. 145-156. New York: Academic Press.
Lashley, K.S. 1951. The problem of serial order in behavior. In Cerebral Mechanisms in Behavior. ed. by L.A. Jeffress, pp. 112-135. New York: Hafner.
Lieberman, P. 1984. The biology and evolution of language. Cambridge, Mass.: Harvard University Press.
Lieberman, P. 1991. Uniquely Human: The Evolution of Speech, Thought, and Selfless Behavior. Cambridge, MA: Harvard University Press.
Maynard Smith, J. and Eörs Szathmary. 1995. The Major Transitions in Evolution. Oxford: W.H. Freeman.
Ojemann, G. and C. Mateer. 1979. Human language cortex: Identification of common sites for sequencing motor activity and speech. In Hearing Mechanisms and Speech ed. by U. Creutzfeldt, H.Schreich and C. Schreiner, pp. 192-220. Berlin: Springer.
Ojemann, G.A. 1991. Cortical Organisation of Language and Verbal Memory Based on Intraoperative Investigation.In Progress in Sensory Physiology 12 ed. D. Ottoson, pp. 193-230. Berlin: Springer-Verlag.
Piattelli-Palmarini, Massimo. 1994 Ever since language and learning: Afterthoughts on the Piaget-Chomsky debate. Cognition 50: 315-346
Pinker, S. 1994. The Language Instinct. London: Penguin.
Pinker, S and P. Bloom. 1990. Natural language and natural selection. Behavioral and Brain Sciences 13: 707-84.
Studdert-Kennedy, M. ed. 1983. Psychobiology of Language. Cambridge, Mass.: MIT.

Studdert-Kennedy, M. 1986. Development of the speech perceptuo-motor system. In Precursors of early speech. ed. by B. Lindblom and R. Zetterstrom, pp. 206-217. New York: Stockton Press.