J. Linguistics 39(2003).
DOI: 10.1017/S0022226703252294
© 2003 Cambridge University Press
Chris Knight, Michael Studdert-Kennedy & James R. Hurford (eds.), The evolutionary emergence of language: social functions and the origins of linguistic form. Cambridge: Cambridge University Press, 2000. Pp. xi + 426.
Reviewed by PHILIP LIEBERMAN, Brown University
Over the course of decades spanning a century, linguists avoided addressing the question of how human linguistic ability might have evolved. The apparent problem was the tendency towards unsupported speculation in the heady period in which evolution first became part of the conceptual framework of human thought. This remained the case until the last decade, even through the Chomskian ‘revolution’, although the centerpiece of the theories promulgated by Noam Chomsky is a hypothetical innate ‘Universal Grammar’ that determines the syntax of all human languages. And syntax was taken to be the defining quality of human linguistic ability. An innate organ of the mind governing syntax must be instantiated in the brain through the expression of genes. Although this amounts to a strong biological claim, the evolution of the putative Universal Grammar was, like most aspects of biology, ignored by virtually all linguists and consigned to fringe groups outside the mainstream of linguistic inquiry.
In recent years that situation has changed and a series of conferences on the evolution of language has been organized by linguists. The volume in question is a collection of some of the contributions to the Second International Conference on the Evolution of Language, which took place in 1998. Not surprisingly, the focus of the conference was on syntax, which since Chomsky’s early publications has generally been taken by linguists to be the factor that differentiates human language from the communications of other species. However, though the conference papers were delivered by many intelligent scholars, for the most part the studies reported in this volume exemplify the hermetic nature of linguistic research. The findings and procedures that have exemplified evolutionary biology since the time of Charles Darwin, and which one might expect to be applied to the study of the evolution of human language, are conspicuous by their absence. Unfortunately, this absence results in claims that have little or no grounding in data, often contravened by facts that are biological certainties.
Virtually all aspects of the theory that Darwin proposed in 1859 are consistent with current data. One of the principles introduced by Darwin amounts to the claim that, unless there is compelling evidence to the contrary, one should not assume that events in the past followed different principles. Historical evidence shows that languages slowly metamorphose into different forms over time, that the connections can be traced, and that changes do not derive from genetic distinctions. It is, for example, clear that the indigenous inhabitants of China are no more genetically predisposed to learn a Chinese language than English, Latin, Walpiri or Sanskrit. And within any given group of human beings who live in a particular region and speak a particular language a range of genetic diversity exists. As Darwin (1859) noted, natural selection acts upon the variation that is always present in the state of nature. We know that natural selection will act to enhance fitness in human populations. Tibetans, for example, generally are adapted to extract more oxygen from air at high altitudes than other human populations. Natural selection over the course of 40,000 years has resulted in the survival in Tibet of those individuals who had the genetic disposition for more efficient oxygen transfer. If the hypothetical genetically determined Universal Grammar actually existed, we should have found that natural selection for the particular parameters which yield Tibetan linguistic competence should also have occurred within the same time span, given the indisputable contributions of language to biological fitness. In light of the substantial differences between Tibetan and English we would expect some delays or deficiencies in the acquisition of English by children of Tibetan ancestry who are born and raised in the United States, if their ‘language’ genes had been optimized for Tibetan, but that is not the case.
It is possible that the common elements that one can see in seemingly unrelated languages may derive from cultural transmission and from the indisputable fact that all human beings have a common ancestor. Except for a few isolated scholars, evolutionary biologists generally accept the theory that modern human beings evolved in the last 200,000 to 150,000 years in Africa. The first human group must have had a particular language. Over the course of time the group or groups of people speaking the ancestral language dispersed and different languages gradually formed. The human groups who left Africa and populated most of the world took this original language or group of related languages with them. A concerted attempt to reconstruct the vocabularies of languages in prehistory has been underway for decades. Unfortunately the findings of this endeavor (e.g. Ruhlen 1996), namely that words rather than syntax appears to be conserved, are absent in this volume except for peripheral citations in two papers whose focus is syntax, namely those by Mark Pagel (‘The history, rate and pattern of world linguistic evolution’) and Frederick J. Newmeyer (‘On the reconstruction of “Proto-World” word order’).
Moreover, the claim that human syntactic ability is innate needs to take into account the fact that the genetic distinctions between humans and chimpanzees are slight. The common ancestor of humans and living apes lived no more than seven million years ago; the genetic distinction between human beings and chimpanzees is somewhere between 5 and 2 per cent, depending on how differences are tabulated. And advances in molecular biology confirm the continuity of evolution — much of the genetic endowment of humans is present in fruit flies. The question then arises, how could anyone account for the evolution of a detailed genetic program for complex syntax, the ‘Universal Grammar’, in this short span of time? The improbability of a genetic code that specifies the syntax of all human languages is reinforced by evidence from neurophysiological and behavioral studies, which show that the details of most motor acts, including walking, are not genetically specified — they are learned (see Lieberman 2000).
Moreover, syntax is not the unique ‘derived’ feature of human linguistic ability. Evolutionary biology differentiates ‘primitive’ features that are present in a species and species ancestral to it and related species. The possession of five digits, for example, is a primitive feature present in frogs and humans. Horses are differentiated by not having five toes. Comparative studies of living species, one of the most powerful tools of evolutionary biology, show that speech production is arguably the most derived feature of human language. The ape language studies of the Gardners and Savage-Rumbaugh & Rumbaugh show that apes can attain a vocabulary of about 150 words and can produce and comprehend simple syntax — but they cannot talk. In this sense, human lexical and syntactic ability are ‘primitive’ qualities, most likely present in the species ancestral to living apes and humans, arguing against the presence of a ‘protolanguage’ that lacked syntax.
In contrast, speech production is a ‘derived’ feature that characterizes human beings. The particular properties of speech permit communication at rates that exceed the fusion frequency of the auditory system, allowing long complex sentences to be comprehended within the attention span of verbal working memory. This would suggest that research on the evolution of the physiology and neural bases of human speech should concern linguists.
One of the papers in this volume, that of Andrew Carstairs-McCarthy (‘The distinction between sentences and noun phrases: an impediment to language evolution?’), touches on the proposal that the neural mechanisms involved in speech production may have been the starting point for syntax, a suggestion that has previously been noted and developed. In a meaningful sense, syntax codifies relationships that hold between various aspects of behavior, which in their totality achieve a particular purpose. Lashley (1951) realized that this is the case and suggested that the ‘roots’ of human syntactic ability most likely lie in the neural structures that regulate motor control. The relationship that may hold between the neural mechanisms regulating motor control and syntax is discussed in detail in Kimura (1979) and Lieberman (2000).
In the volume in question, three papers specifically address the evolution of human speech. First, Peter MacNeilage & Barbara Davis, in their paper ‘Evolution of speech: the relation between ontogeny and phylogeny’, propose that basic syllable structure derives from opening and closing one’s jaws. They further propose that the ontogenetic development of speech in infants may reflect the evolutionary pattern. Neither proposal is novel: for instance, Muller (1848) and Jakobson (1940) also pointed out the possible relationship between the sequence in which sound patterns are acquired and the frequency with which they occur in the various languages of the world. Nonetheless, the MacNeilage & Davis paper makes specific predictions that can be experimentally tested and it provides a refreshing starting point for further research.
Second, the paper contributed by Michael Studdert-Kennedy, ‘Evolutionary implications of the particulate principle: imitation and the dissociation of phonetic form from semantic function’, focuses on the ‘particulate’ nature of speech, citing phonetic transcriptions and alphabetic orthography, downplaying the fact that speech is encoded into syllable-sized units. Given Studdert-Kennedy’s long association with Haskins Laboratories, where many of the seminal studies on speech encoding originated, this omission is difficult to understand. Studdert-Kennedy cites alphabetic systems of orthography as evidence for the psychological reality of particulate phonemes, but he disregards the fact that other successful orthographic systems code entire syllables and words, as is the case for Chinese. And there is much evidence that syllables are primary units in both the production and perception of speech. For example, it is extremely difficult for monolingual speakers of English to produce words that begin with the final [ŋ] of the English word sing. Studdert-Kennedy also claims that meaning plays no role in the process by which children imitate sounds — a most unlikely conclusion. Even dogs pay attention to the meaning of a word: every dog owner knows that dogs quickly learn to respond to words such as ‘out’ or ‘biscuit’. In contrast, consider the third paper focusing on speech, by Marilyn Vihman & Rory A. Depaolis, ‘The role of mimesis in infant language development: evidence for phylogeny’. These authors disagree with Studdert-Kennedy and discuss the role of imitation in the acquisition of speech by children.
Chris Knight, in an interesting contribution on ‘The evolution of cooperative communication’, points to the role of play in the acquisition of language; however, he neglects the body of evidence documented by Greenfield (1991) which argues for a gradual transition from gestures to vocal signals in the course of human language development. Knight instead favors a model that leads to an abrupt saltation from ‘protolanguage’ to the sudden appearance of full human linguistic ability.
A number of papers propose various aspects of human behavior in which language may have enhanced Darwinian biological fitness, thereby resulting in natural selection for linguistic ability: cooperation (Knight); politics (Jean-Louis Dessalles, ‘Language and hominid politics’); ‘secret female languages’ (Camilla Power, ‘Secret language use at female initiation: bounding gossiping communities’); social transmission (James R. Hurford, ‘Social transmission favours linguistic generalisation’) and other social forces are all noted. All of these factors may have contributed to biological fitness — the survival of progeny and the ultimate evolution of human language through Darwinian natural selection. The synergy between linguistic ability and these activities may very well have enhanced fitness and led to natural selection for language. However, it is difficult to identify any aspect of human behavior (excepting suicide, protracted warfare, and tendencies in males towards misogyny) that would not profit from the exchange of information by means of language, thereby enhancing fitness.
In contrast, several papers take the position that natural selection played a minor part in the evolution of human linguistic ability. Although biologists — including Darwin (1859) acknowledge the role of chance events in shaping the course of evolution, natural selection acting on variation clearly is the key element in evolution, as Mayr (1982) notes. The argument that the evolution of human language is somehow different would have to be supported by compelling evidence. However, various papers in this volume present computational models rather than data to support their claim. For example, Simon Kirby, in ‘Syntax without natural selection: how compositionality emerges from vocabulary in a population of learners’, appears to claim that syntax develops because people acquire many words, but his model starts with individuals who are preloaded with context-free grammars at the start of the evolutionary process which leads to complex grammars. In short, Kirby starts with a built-in grammar to prove that grammar develops spontaneously. Bart de Boer, in ‘Emergence of sound systems through self organisation’, presents a different computational model, which claims that vowel distinctions ‘emerge’ without ‘evolution-based explanations for the universal tendencies of vowel systems’ (193). Studies that date back to Muller (1848), who noted that the vowels [i] [u] and [a] occur most frequently in human languages, refute this position. Paraphrasing George Orwell, some sounds are more equal than others. For example, Stevens’ seminal paper (1972) shows that these vowels have physiological and perceptual properties that are better adapted for vocal communication. Although de Boer references Stevens (1972), he ignores Stevens’ findings, as well as Lindblom’s and Labov’s studies of sound change. Olmsted’s (1971) study, showing that children first acquire the speech sounds which are most perceptually salient and which also occur most often in different languages, is neither referenced nor noted.
A tendency to rely on computer model studies rather than data documenting the utility of syntax characterizes other contributions. Jason Noble, in ‘Cooperation, competition and the evolution of prelinguistic communication’, makes use of game theory simulation to attempt to account for a transition from ‘prelinguistic’ to linguistic communication, disputing results from other simulations. Observations of the actual communicative abilities of apes and other animals which might provide insights into ‘prelinguistic’ communication are absent. David Lightfoot, in ‘The spandrels of the linguistic genotype’, takes as a given an innate Universal Grammar. Lightfoot claims that the putative ‘linguistic genotype’ is a ‘spandrel’ — derived fortuitously as the result of evolution directed towards a different end. Lightfoot’s linguistic analysis invokes traces — presently relegated to the dust heap in Minimalist linguistic theory. Derek Bickerton, in his contribution ‘How protolanguage became language’, discusses the transition from a ‘protolanguage’ that lacked syntax to ‘human’ language. However, as noted above, cross-fostered chimpanzees who have been raised with signed or manually expressed human language comprehend sentences with simple syntax and produce simple ‘utterances’ (Savage-Rumbaugh & Rumbaugh 1993, Gardner & Gardner 1994). It is most unlikely that any archaic hominid communicated with a ‘protolanguage’ lacking syntax. Bickerton also claims that the brains of other species are incapable of controlling actions that must be ‘maintained without external stimulation for long periods’ (272), thereby rendering them incapable of comprehending or producing sentences. A reading of Jane Goodall’s comprehensive account of chimpanzee behavior (1986) will inform the reader that this is not the case. Anticipatory tool preparation and chimpanzee warfare clearly involve long-term cognitive activity without external stimulation.
In short, the hermetic tradition of linguistic research pervades many of the papers in this volume. Computer simulations can be useful but they must take account of biological facts if they are to provide any insights on the course of evolution. And the broader framework of evolutionary biology must be considered. Studies that take account of the communicative and cognitive behavior of human beings and other species incorporating relevant anatomical, physiological, neurophysiological and genetic data as well as the archaeological record would better illuminate our understanding of the evolution of human language.
REFERENCES
Darwin, C. (1859). On the origin of species (Facsimile edn. 1964). Cambridge, MA: Harvard University Press.
Gardner, B. T. & Gardner, R. A. (1994). Development of phrases in the utterances of children and cross-fostered chimpanzees. In Gardner, R. A., Gardner, B. T., Chiarelli, B. & Plooj, R. (eds.), The ethological roots of culture. Dordrecht: Kiuwer. 223-255.
Goodall, J. (1986). The chimpanzees of Gombe. Cambridge, MA: Harvard University Press.
Greenfield, P. M. (1991). Language, tools and brain: the ontogeny and phylogeny of hierarchically organized sequential behavior. Behavioral and Brain Sciences 14. 531-577.
Jakobson, R. (1940). Kindersprache, Aphasie und allgemeine Lautgesetze. In Selected writings (vol. I). The Hague: Mouton. 328-401. [Translated by A. R. Keiler, 1968, as Child language, aphasia, and phonological universals. The Hague: Mouton.]
Kimura, D. (1979). Neuromotor mechanisms in the evolution of human communication. In Steklis, H. D. & Raleigh, M. J. (eds.), Neurobiology of social communication in primates. New York: Academic Press. 197-219.
Lashley, K. S. (1951). The problem of serial order in behavior. In Jefress, L. A. (ed), Cerebral mechanisms in behavior. New York: Wiley. 112-146.
Lieberman, P. (2000). Human language and our reptilian brain: the subcortical bases of speech, syntax and thought. Cambridge, MA & London: Harvard University Press.
Mayr, E. (1982). The growth of biological thought. Cambridge, MA: Harvard University Press.
Muller, J. (1848). The physiology of the senses, voice and muscular motion with the mental faculties (Translated by W. Baly). London: Walton & Maberly.
Olmsted, D. L. (1971). Out of the mouth of babes. The Hague: Mouton.
Ruhlen, M. (1996). On the origin of language: tracing the evolution of the mother tongue. New York: John Wiley & Sons.
Savage-Rumbaugh, E. S. & Rumbaugh, D. The emergence of language. In Gibson, K. R. & Ingold, T. (eds.), Tools, language and cognition in human evolution. Cambridge: Cambridge University Press. 86-108.
Stevens, K. N. (1972). The quantal nature of speech: evidence from articulatory-acoustic data. In David, E. E. & Denes, P. B. (eds.), Human communication: a unified view. New York: McGraw-Hill. 51-66.
Author’s address: Cognitive and Linguistic Sciences, Brown University, Providence, RI 02912-1978, USA.
E-mail: Philip Lieberman@brown.edu (Received 7 April 2003)