Publishing Partner: Cambridge University Press CUP Extra Publisher Login
amazon logo
More Info

New from Oxford University Press!


Style, Mediation, and Change

Edited by Janus Mortensen, Nikolas Coupland, and Jacob Thogersen

Style, Mediation, and Change "Offers a coherent view of style as a unifying concept for the sociolinguistics of talking media."

New from Cambridge University Press!


Intonation and Prosodic Structure

By Caroline Féry

Intonation and Prosodic Structure "provides a state-of-the-art survey of intonation and prosodic structure."

Review of  The Evolution of Language out of Pre-language

Reviewer: Anne Reboul
Book Title: The Evolution of Language out of Pre-language
Book Author: T. Givón Bertram F. Malle
Publisher: John Benjamins
Linguistic Field(s): Psycholinguistics
Anthropological Linguistics
Language Acquisition
Issue Number: 14.1734

Discuss this Review
Help on Posting

Date: Tue, 17 Jun 2003 10:59:30 +0200
From: Anne Reboul
Subject: The Evolution of Language out of Pre-language.

Givon, Talmy and Bertram F. Malle, eds (2002) The Evolution of Language out of Pre-language. John Benjamins Publishing Company, paperback ISBN 1-58811-238-1, ix+392.

Reviewed byAnne Reboul, Institute for Cognitive Sciences, CNRS, France

This collection of papers is devoted to the subject of language evolution from a (mostly) functionalist point of view (language as a tool for communication rather than for thought). It includes papers by linguists, neurophysiologists, anthropological paleontologists and psychologists. It is on the whole easily accessible to graduate students, though it suffers from a lack of basic discussion: the functionalist perspective is adopted without so much as an argument.

The book is made of four parts: language and the brain; language and cognition; language and social cognition; language development.
Part one opens with a paper by one of the editors, T. Givon, on "The visual information-processing system as an evolutionary precursor of human language", and defends the corresponding (main) thesis that neural linguistic structures evolved from the visual information-processing system and the (secondary) thesis that the lexicon first emerged as a gestural code. Givon distinguishes two components of the human language, considered as a system for the representation and communication of experience: the cognitive representation systems ? conceptual lexicon, propositional information, multi-propositional discourse ? and the symbolic communicative codes ? peripheral sensory-motor codes and the grammatical code. The conceptual lexicon is a network of nodes and connections and corresponds to "semantic memory" while propositional information is encoded in clauses and corresponds to "episodic memory". The multi-propositional discourse can be seen as sequences of clause-chains (sentences combining several propositions). While peripheral sensory-motor codes map lexical-conceptual meaning onto phonological words, grammar is the latest evolutionary addition, as shown both by the order in first- and second-language acquisition and by the easy acquisition of symbolic systems by animals though grammar remains difficult if not altogether impossible for them. Pre-grammar is considered as proto-grammar and largely consists of what pragmaticians would call pragmatic rules (common sense rules more or less on gricean lines), which are to be found as much in non-linguistic as in linguistic communication. Givon then turns to neural substrates of language and recalls the distinction between the ventral stream for object recognition and the dorsal stream for space information processing, linking them directly, respectively, to the lexicon (semantic memory) and to propositions (episodic memory). Givon defends the thesis that language modules are distributed all over the brain. The main evolutionary hypothesis of the paper is double: the cognitive representation modules (semantic and episodic memories) arose pre-linguistically as visual information-processing modules and were the basis of a peripheral lexical code, which evolved from a visual-gestural beginning to a auditory-oral system. The grammar was a late addition to the human communicative system that may or may not have preceded the evolution from the gestural to the oral code. The paper ends with a defense of the hypotheses of a gestural code preceding the oral code and of a late evolving grammar.
The next paper by D. Tucker, proposes an evolutionary-developmental analysis of adaptive semantics. Its main idea is that "language evolved to transcend evolution" (51), allowing shortcuts in behavior changing through cultural transmission. Language is "embodied" in the left cerebral hemisphere. Evolution concentrated on "terminal additions", i.e. "recent forms could be added only to the terminal stage of embryogenesis" (55) which means that they concentrate on the telencephalic structures, the striatum and the limbic circuits. However, important preadaptations for human language go with social communication and are thus almost certainly precortical. The paper would have benefited from a more important number of illustrations and brain maps to make it more digestible.
The third paper is devoted to a description of the missing links in hominid fossils and is by C. Li. Li reviews recent discoveries in human paleonthology (e.g. Australopitecus ghari who used stone tools) and discusses their impact on the language evolution debate. He quite rightly emphasizes that the evidence is not enough to provide clues to the debate and advocates an integration of paleoanthropologic, neuroscientific, ethological, linguistic and evolutionary informations. According to him, rather than recursion, the criterion dividing pre-linguistic from linguistic communication is the dependence of the first of darwinian evolution and the dependence of the second on cultural transmission. Language evolution depends on the emergence of symbolic signals, referring in a context-independent way to concrete objects, though the cristallization of language can be complete only with the symbolization of abstract or absent entities. Once the lexicon reaches a critical mass, grammar will emerge in a few generations, as shown by creolization. However, "the cognitive prerequisite for the emergence of language is the ability to understand fully intentionality, causality and the mental states of conspecifics" (95), an ability shared by humans and chimpanzees. Li then evokes mirror neurons (which fire both in planning and in recognizing action) as a tool for learning (and for language acquisition) as well as the uniquely human migration of neuronal precursor cells from the telencephalon to the diencephalon in the foetus, explaining the wide differences between human and chimp brain despite the genetic similarities (99% of common DNA between the two species). Though Li's paper is enthusiastic and well-informed, it takes a lot for granted, notably the idea that language developped primarily if not exclusively as an instrument for communication. A sentence deserves quotation: "even from the perspective of intuition, it is difficult to imagine language as an instrument of thought before the dawn of civilization which is marked by such features as written language, urbanization, advanced technology, expanded population and complex social organization" (87). It might equally be said that it is difficult to imagine the development of written language, urbanization, advanced technology, expanded population and complex social organization if language is not or could not be used as an instrument for thought. This is very much an egg-or-hen argument and, as is well-known, such an argument does not prove anything.
Part 2 opens with a paper by J. Bybee on "sequentiality as the basis of constituent structure". Her hypothesis is "that sequentiality is basic to language and constituent structure emerges from sequentiality because elements that are frequently used together bind together into constituents" (109). Constituents are defined as sequences that can be used alone, can be replaced by pro-forms and can occur in various places in the sentence. Grammatical knowledge is procedural and thus unconscious. Bybee justifies her hypothesis from a study on the composition of English noun phrases, concentrating on eleven lexical nouns most frequently found in a corpus and examining the items immediately preceding and immediately following the nouns in question. She finds that the most frequent items before the nouns are determiners while those following the nouns are prepositions, verbs and relative clauses. "The primary counter-examples involve preceding prepositions, which seem restricted to certain high frequency phrases, and common verbs or auxiliaries following the noun" (120). This leads her to the claim that language learning consists in storing "repeated fragments of speech" (120), leading to the frequent production of prefabricated units in adulthood. Hierarchy is thus created out of automated sequences, a domain-general cognitive process. An additional argument is that chunking (e.g. auxiliary contraction) occurs in violation of constituent structure, but according to the predictions of the frequency hypothesis. "Thus constituency (?) is the convergence of two other factors and is itself not a basic structure. It is an emergent property of language" (130) and is basically epiphenomenal to the linguistic system. This leads to the view that the development of grammar is the result of domain-general abilities. Bybee's work is based on statistical analysis and the main problem, as often in the use of statistics, is the interpretation of the data. For instance, there is nothing surprising in the results of the noun phrases analyses given a generative view of language. The same thing goes for the argument that 55% of spoken and written texts are made of prefabricated units: there still remains 45% which are NOT made of prefabricated units, not such a feeble result after all! And finally, chunking in violation of constituent structure can also be explained in GG: the most likely hypothesis is that it is a phonological rather than syntactical phenomenon. Some support for that hypothesis can be found in the asymmetry between subject and object in liaison, in French for instance (impossible between a full NP subject and the verb or auxiliary, possible between the verb and the object).
The next paper deals with the internal structure of the syllabe from an ontogenetic perspective and is by B. Davis and P. MacNeilage. The idea is that speech could have evolved from non-speech vocal capacities, through coordination of mammalian laryngeal phonation and articulatorily alternations between open and closed mouth configurations. These can be found in infant babbling and thus this paper, as most in the volume, relies on the hypothesis that ontogeny recapitulates phylogeny. Rhythmic mandibular oscillations are the basis, simulating consonants and vowels. This is general through the acquisition of languages, however different they may be phonologically. Thus, "evolution may likely have preserved functional aspects of the internal structure of the jaw cycle (i.e. intrasyllabic regularities) while elaborating over time with additional adaptations seen in increase of independent articulator movements within syllables in the service of increase in message complexity" (149).
The next paper, by the same authors in reverse order, deals with the origins of intersyllabic complexity, that is, it justifies the content of the first paper by explaining why the relative phonological poverty of infant babbling leaves room for differentiated consonants and vowels in successive syllables. The order is usually labials-coronals because labials being more easy to produce are chosen to initiate the action, then are followed by the less easy coronals and finally with the still less easy dorsals. The trend for this preference is not found only in early words but is widely spread across languages, where the frequency order for initial position in words is labial>coronal>dorsal while it is reversed for final position. "Sound patterns of words are a result of two sometimes conflicting forces ? production constraints and the need for perceptual distinctiveness" (166). Thus, phonology evolved through "a self-organizing process of resolution of conflicts between production and perception constraints" (167). These two papers are well-argued and fairly convincing, but it is hard to see why their empiric claims should go against a generative account of language evolution or development. If anything, they would seem in agreement with some central tenets of optimality theory.
Chapter 7, by T. Givon and M. Barker, deals with the pre-linguistic origins of language processing rates. The paper is based on the stability of the temporal flow of word-processing (ca. 250 msecs) and clause-processing (ca. 1.0 sec.). The paper returns to the hypothesis of two neural systems for two types of memories for, respectively, the lexicon and clauses presented in Givon's opening paper to the present volume, introducing mirror neurons, to the debate. The difference in time between the processing of words and clauses led to the hypothesis that episodic memory of individual events and objects would decay rapidly if they were presented at rates inferior to the language processing of, respectively, clauses and words. This led to two experiments, which "fit, substantially if not absolutely, the predictions" (188). Thus, the two types of processing rates coincide. "Given the evolutionary precedence of visual information processing over language, an evolutionary interpretation of the amazing coincidence of the two types of rates is an eminently respectable theoretical gambit" (189).
G. Fenck-Oczlon and A. Fenck sign the next paper on the clausal structure of linguistic and pre-linguistic behavior. It begins with a plea for the co-evolution of language and the relevant cognitive capabilities, but concentrates on the cognitive pre-conditions for language development. A first requisite is mental propositions with a predicate-argument-structure, corresponding to "kernel-sentences". A trans-cultural phenomenon in human action is a segmentation in the range of a few seconds, and the hypothesis is that "intonation units [corresponding to propositions] in a language are a special case of action units" (221). Thus "the underlying span of (about two seconds and) about seven syllables has, according to our findings, the appropriate size for encoding one proposition" (226).
This leads us to part 3, which opens with a paper on the gradual emergence of language by B. Mac Whinney. Mac Whinney, though he acknowledges the species-specific character of language, insists that language did not evolve in and of itself but was the result of the six million years of evolution that produced modern humans. He distinguishes four major evolutionary periods, all of them accompanied by brain modifications: the acquisition of the bipedal gait, the development of group structure solidification through face to face interaction, the growth of social expression of language functioning ? prosody, chant, gesture, dance, etc. ?, and, finally the linking of language ? with the development of double articulation at phonological and syntactic levels ? and cultural production. "The four challenges [behind these evolutionary periods] are: bipedalism, social cohesion, mimetics, and systematization" (235).
Chapter 10, by the co-editor of the volume, B. Malle, is devoted to the co-evolution of language and theory of mind, both onto- and phylogenetically. Though language and theory of mind are not inseparable, they are strongly linked. Language developed for communication that itself is crucial to social coordination. Theory of mind is the ability to attribute mental states to others and to reason about them and develops in the child from the visual perception of agency and the development of the concept of intention. Though theory of mind in the full sense implies language, there are non-linguistic precursors to theory of mind (e.g. shared attention) as is shown by the fact that "language acquisition itself appears to rely on theory of mind skills" (269). This leads to the following picture: Theory of mind -> early language -> later stages of theory of mind. This suggests to Malle a coevolution hypothesis, which he discusses according to three models: coincidental coevolution, coevolution due to the emergence of a third factor, evolutionary escalation. He favors the third, though he recognizes that no clear evidence exists at present.
D. Baldwin signs the next paper on the rise of human intentional understanding, taken to be a system closely-allied to language. Language is riddled with intentionality, filling most descriptions of action with meaning and, as a tool for communication, is in contact with intentional understanding. And, as language, intentional understanding is a generative system. Acquisition of intentional understanding goes through structure detection and induction and the same is true for language acquisition.
Part 4 opens with a paper by M. Tomasello on the emergence of grammar in early child language, describing a model based on the following hypothesis: the acquisition of language proceeds through the functional analysis of utterances ? in terms of their communicative function ?, the creation of schemas ? both at utterance and at constituent level ?, the assembling of schemas into utterances. What this leads to is a bottom-up view of language acquisition in which the acquisition of syntax is similar to the acquisition of lexical morphology.
This is followed by a chapter by J. Morford, trying to answer the question of why exposure to language matters. Morford begins by distinguishing two categories of grammatical properties: "system internal" ones (within a single grammar, presumably involved in creolization), and "processing-dependent" ones, resulting from the automation of language processing. The second ones are crucially dependent on exposure. Morford discusses the cases of deaf children of normal parents and of the way their gestural production is organized, despite their lack of exposure to signed language. She reminds us of how difficult it is to learn a first language as a teenager through the by now classical example of Genie. Her conclusion is that "the same individual who innovates structure in a communication system will not have the opportunity to automate processing of that structure" (338), because automation needs exposure.
S. Goldin-Meadow, well-known for her work on deaf children, speaks of language creation, turning rather naturally to deaf children to highlight the linguistic creativity of language learning. She reports a study of severely deaf children raised by hearing parents with an oral method. She analyzes the gestures used by these children and notes that they show resilient linguistic properties, such as stable words paradigmatically and categorically organized and with grammatical functions, sentences marked by deletion, word order, inflection and recursion, and, finally, pragmatic properties of language use, such as indexicality, communication about absent or imaginary objects, narration, self-talk and metalinguistic uses. Such gesture systems develop over time, this being manifest in utterance growth, the apparition of morphology, grammatical categories. Goldin-Meadow compares this with language creation in adults (typically the hearing parents of deaf children, communicating through a sui generis gesture system), noting that though some grammatical structure (i.e. ergativity) emerges, morphology does not. She hypothesizes that two types of forces are at play: synchronic forces operating on communication and diachronic forces constraining the evolution of the system.
The final paper is by D. Slobin and probes the parallels between language evolution, acquisition and diachrony. In particular he examines four hypotheses that are well represented in the preceding papers: "1. The course of language development in the child has parallels with the development of language in our genus and species. 2. The sequence of development of linguistic forms in the child mirrors the diachronic development of those forms in the language that the child is acquiring. 3. Diachronic language change is, in part, the product of the learner. 4. The emergence of new languages ? creoles and Nicaraguan Sign Language ? is due to child learners" (376). He dismisses all of them with diverse arguments, e.g. the heterochrony of cognitive development between humans and chimps (making the comparison of toddlers with monkeys or apes irrelevant). His criticisms appear well founded and convincing, which throws doubts on much of the papers that appear in the rest of the book.


It is highly commendable in a book where most if not all of the chapters rely on the hypothesis that ontogeny recapitulates philogeny to include a chapter such as Slobin's which raises difficulties for such a view. However, it should be clear that, with the exception of deaf children raised by non-signing parents, there is absolutely no reason to suppose that acquiring an existing language through normal exposure at an individual level can be productively compared with evolving a language at a species level. Thus, the whole idea of the ontogeny of language acquisition recapitulating the phylogeny of language evolution seems, to say the least, debatable. At the most, it can give rise to what Slobin calls "speculative scenarios" which "we can never have sufficient evidence to scientifically evaluate" (389). Much of the book is (understandably) like that. The difficulty of studying language evolution, given that there are no fossil languages, leads rather naturally to the four hypotheses that Slobin rejects as the only possible approaches to the problem. Yet, if, as he convincingly claims, these hypotheses cannot be maintained, then maybe the very problem should be abandoned. In other words, though language evolution is a fascinating problem, it is far from clear that it can be scientifically studied.
Another problem stems from the very functionalist perspective of the book that pervades most if not all papers and leads to rather strong though not clearly justified assertions on the communicative function of language (see above, Li's quotation). A more balanced view, given that, as said above, there is no clear indication of the reason why language evolved in the first place (and of whether it evolved as the result of a specific mutation or just emerged as a byproduct of the evolution of other cognitive capacities) might be to suppose that language is both a communicative and a cognitive tool, facilitating not only communication but thought, an hypothesis which is supported by some studies in comparative psychology. For instance, Povinelli (2000) has argued that chimpanzees not only do not have the same folk physics and theory of mind as we do, but do not seem to be able to acquire the concepts underlying these cognitive abilities in human. He suggests that this is due to the lack of language in chimpanzees. And, as Malle recognizes, language, even though it may depend on theory of mind for its acquisition, seems to constrain later developments in theory of mind abilities. Thus, there may be some place for a dual view of the language function: both social ? for communication ? and cognitive ? for thought. Finally, though quite a few papers in the book claim that language developped because of the constraints placed upon communication by the growth in number of hominid groups, it should be said that no one really knows how big or small those groups were: for all we know, they might have the same size as contemporary groups of chimpanzees and chimpanzees seem to manage a fairly sophisticated social organization without language. Unless one wants to fall in the fallacy of considering present day so-called stone age people as similar to pre-sapiens hominids, there is no way other than sheer speculation and wishful thinking of describing the size or social organization of hominid groups when language appeared.
One line which, in the light of the above criticisms, might be pursued to study language evolution is simulation. Though simulation will not be able to determine which scenario actually occurred, it can at least help decide which scenarios are or are not possible. Thus simulation would at least be a test space for evaluation of concurrent scenarios.
Finally, there are some arguments which seem misguided. For instance, the first chapter (by Givon) relies fairly often on comparisons between language acquisition and learning to read which seems debatable as (first) language acquisition clearly is a natural phenomenon while learning to read, just as clearly, is not. What is more, it is not obvious that the structure of the adult brain as well as its activation as recorded in brain-scan studies can tell much about what would or would not be specific to language or about what is or is not innate in language acquisition.

D.J. Povinelli, "Folk physics for apes: The Chimpanzee's Theory of how the world works", Oxford, Oxford University Press, 2000.

BIOGRAPHICAL SKETCH Anne Reboul is a First Class Research Fellow at the French Center for Scientific Research (CNRS) in France. She has a Ph.D. in Linguistics (EHESS, Paris) and a Ph.D. in philosophy (University of Geneva, Switzerland). She has written some books, among which an Encyclopaedic Dictionary of Pragmatics and quite a few papers in French and English, on pragmatics and/or philosophic subjects. ACKNOWLEDGMENTS I have benefited from the help of the following colleagues in writing this review: Viviane Deprez, Jacques Moeschler and Tatjana Nazir. Needless to say, any errors are mine.