Date: Tue, 8 Mar 2005 09:41:53 +0100 From: Anne Reboul Subject: Doctor Dolittle's Delusion: Animals and the Uniqueness of Human Language
AUTHOR: Anderson, Stephen R. TITLE: Doctor Dolittle's Delusion SUBTITLE: Animals and the Uniqueness of Human Language PUBLISHER: Yale University Press YEAR: 2004
Anne Reboul, Research Team "Linguistics, Pragmatics and Cognition", Institute for Cognitive Sciences, CNRS, France
The present book presents both a very accessible, readable, extremely well- informed and wide ranging (from bees to humans) introduction to animal communication as well as a very scholarly discussion of what makes human language different from any of the other existing communication systems in the animal world. At the same time, it includes a careful and honest discussion of what could be seen as hybrids between animal systems of communication and the human language, i.e., the experiments aiming at developing something like human language in nonhuman animals, from Alex the parrot to Koko the gorilla, passing through the chimpanzees Washoe and her conspecifics and Kanzi, the bonobo who still represents the most successful of such experiments in apes.
This is a book well-worth reading: for the uninformed reader, it's a wealth of information on animal communication relying on the latest literature and covering all aspects of communication, from acoustics to contents of signals, and including a very clear discussion of human language (from the vantage point of generative grammar). For the informed reader, the book is still well-worth reading for the discussions, which are both well-argued and very fair. There is no hasty dismissal of claims relative to the diverse abilities of such and such a species but a careful exposition of the similarities and dissimilarities between animal and human communication. As a sample example, Anderson notes that all the apes who were taught ASL did not develop the features of ASL that make it a language in the human sense, i.e., morphological and syntactic features: he points out however that they were taught whatever ASL they learnt by people for whom ASL was a more or less badly learnt foreign language rather than their mother tongue and who did not themselves exemplify those crucial morphosyntactic features. This does not mean that the apes could learn ASL in the strict sense (as spoken by deaf people), but it does show, as Anderson clearly says, that the design of the experiments was certainly not optimal if the goal was, as advertised, to show that apes could master human language.
I'll begin by a short synopsis before returning to what I consider as the most important points of the book.
The book is organized in eleven chapters preceded by a preface. The first chapter, "Animals, Language, and Linguistics" , introduces linguistics as the scientific field that can say what is or not a language, reviews popular claims about animal communication, either natural or artificially induced (talking apes) and outlines the contents of the rest of the book.
The second chapter, "Language and communication", contrasts language and communication as well as communication as usually understood, on the basis of linguistic communication (clearly intentional), and what takes place in animal communication (where intentionality is at best debatable). It then turns to the characteristics of language as enumerated by Hockett (1960), noting that though some are common between language and animal communication, others are not.
Chapter 3, "On studying cognition", opens with a discussion of the distinction between apparent (termite mounds) and real complexity of structure (language) and of the dangers of over- as well underestimating cognitive abilities, both human and nonhuman. It then turns to communication as a double way into animal mind: through the study of natural animal communication (what is communicated among conspecifics and non-conspecifics); through direct communication with animals who have been taught a means of communicating with humans. Additionally, studying the difference between animal and human communication also provides valuable insights about human cognition and its (linguistic) specificities. Anderson adds a few words of caution relative to overestimation of animal capacities, advocating Morgan's law (never attribute to a higher faculty what can be explained through the exercise of a lower faculty), before turning to natural language, its modular nature as evidenced through double dissociation and its complexity.
Chapter 4, "The dance "Language" of honeybees", tells the story of the well-known honeybee dance, from its first description by Karl von Frisch to the most recent experiments. After a careful description, Anderson points out that though communication is undeniably taking place, the honeybee dance is clearly not intentionally communicative, is entirely innate and, though unbounded in the sense that it can transmit an infinite number of messages, quite restricted as to what is being communicated, i.e., the location of food. Additionally, though the dance is sequential, it is not compositional in the way human languages are. Finally, the dance is iconic, rather than arbitrary. Anderson then turns to the question of whether the dance involves displacement. Though it might seem so, he concludes that "the parameter of the dance that indicates distance to the food does not represent a genuinely external property, but reflects the bee's subjective experience in flight" (88).
Chapter 5 is devoted to "Sound in frog and man". Male frogs produce mating calls, which allows females to choose their partners on the basis of the depth of their voices, indicative of their size, based on the simple principle that "Bigger is better". Anderson then turns to a detailed and very clear exposition of the acoustics of frog calls, then turning to the properties of the vocal tract in humans, as well as to their auditory system, noting that speech is perceived in a different way from other auditory stimuli. He introduces the motor theory of speech perception: "we are competent hearers (...) because we are also talkers" (115), before turning to phonology, concluding the chapter with an analogy between songbirds and humans. "The capacity to acquire the phonology of a language is just as deeply rooted in our own biology" (127) as the capacity of songbirds to acquire their songs is rooted in their biology.
Rather naturally, the next chapter turns to "Birds and babies learning to speak". Some species of songbirds learn their songs in a manner highly reminiscent of the way children acquire language, i.e., "song develops in a way that requires interaction with early experience (rather than being entirely innate)" (129). In contrast to bird calls, songs are an expression of territoriality, being both warnings to other males and attractive to females. Song learning and production is quite different from call learning and production. Calls are always innate and use different brain structures. Song by contrast is not entirely innate, but its learning, though it involves experience, is strongly biologically constrained. Some lateralization is implied in song and "the motor control areas involved in song production are also involved in song perception" (143). Song learning exhibits a critical or sensitive period, just as language acquisition. Finally, songs may be said to have phonology, but not syntax. Anderson then describes the chronology of language acquisition, from cooing to syntax and concludes that "nestlings and babies both grow up in a specific way, determined in essence by the fact that they are birds and humans, respectively" (165).
Chapter 7, "What primates have to say for themselves", turns to natural communication in primates. Anderson begins with the remark that though monkeys use mainly vocalizations and relatively little gestures, exactly the reverse is true of apes. Calls in monkeys and prosimians (e.g., lemurs) are not restricted to alarm calls, though the alarm calls of some species such as vervets have been extensively studied. The three distinct calls of vervets seem to be referential in as much as they don't express degrees of danger or of alarm, but refer to different types of predators. What is more, vervets also use the alarm calls of non-conspecifics (i.e., the superb starling) and show themselves sensitive to the semantic differences between their and other species' calls. Other monkeys, such as Campbell's monkeys seem to evidence a combination of two elements, the first apparently modifying the alarm call to mitigate it. Anderson then turns to a comparison between call production and interpretation in human and nonhuman primates, insisting on the specificity of the upper vocal tract in humans. Regarding perception, nonhuman primates show evidence of lateralization for the perception of conspecific vocalizations. Regarding learning, the production of monkey vocalization seem innate: the calls are produced from birth on, though some fine tuning takes place as to when exactly the different calls must be produced. Finally, alarm call production is sensitive to audience, which is an additional reason not to consider it as an expression of an internal state. Turning to apes, chimpanzees do not seem to have alarm calls differentiated in terms of predators, though this can be explained through the differences in selective pressure on chimpanzees and monkeys. Pant-hoots identify individuals though they can hardly be regarded as referential. Chimpanzees also communicate through gestures, which "seemingly are flexible in formation and use" (195). Thus "that nonhuman primates can be trained to use manual gestures in a meaningful way is to some extent a natural extension of normal behavior" (196).
Chapter 8 returns to language, being devoted to "Syntax". As Anderson notes, "we find the same basic structural properties in every human language" (199). Those properties are those that make languages discrete combinatorial systems. These systematic properties can be evidenced by pro- forms substitution, displacement and agreement relations, the discussion of which leads Anderson to the autonomy of syntax.
Chapter 9 argues that "Language is not just speech". It is devoted to a discussion of ASL, showing that ASL, though using gestures rather than phonemes, is no less a language than English, French or any sound using language. Though signs may be originally based on iconicity, they very quickly reduce iconicity, thus exemplifying the arbitrariness typical of human languages. Again, though ASL does not use sounds, one can find in it the equivalents of phonemes, morphemes, tense and aspect markers, as well as the same syntactic properties evidenced by spoken languages.
Chapter 10, "Language instruction in the laboratory", leaves the field of natural communication to describe the experiments which from the late 60s have attempted to teach animals, mostly apes, human languages, mainly ASL. Given that signed languages are languages, if an ape could be shown to have mastered it (including its fundamental linguistic properties), it would show that "human language is within the cognitive capacities of an animal" (266). Though most such experiments used ASL, some used keyboards or tokens. As Anderson notes, the star of the show is undoubtedly the bonobo Kanzi though it is still doubtful whether he "has truly acquired the structural core of a human language" (268). Anderson reviews the results of these experiments, beginning with those based on ASL. He notes that these experiments were seriously flawed given that ASL was not a first language for any of the experimenters, which means that they learnt it as a foreign language as adults and had no true mastery of its fundamental linguistic properties. Thus, they could not in any way transmit those properties to the apes who were never exposed to linguistically correct utterances. Another doubtful area concerns the type of combinations that the animals produced: none of them seem to approach anything like syntax, neither do these utterances manifest displacement, other than in relation to food. The most interesting case, however, is Kanzi, though his abilities should be considered in interpretation rather than in production (Kanzi communicates with keyed lexigrams). The comparison made between the understanding by Kanzi and a 18 months old child of English is indeed impressive in terms of cognitive abilities, showing that Kanzi seems to operate on a fairly sophisticated word-chain model, though his performance was poor for sentences involving grammatical words (e.g., prepositions and conjunctions). However, it does not show that Kanzi has mastered anything like the syntax of natural language. Regarding mastery of the lexicon, though non-iconicity is not in doubt, displacement and non-instrumentality may not be present. Anderson then turns to the remarkable Alex, an African grey parrot to whom Irene Pepperberg has taught a communicative system based on natural language (English). Again his receptive abilities are more developed than his productive abilities, though he, like Kanzi, seems to operate on a word chain system. His words are non-iconic, though displacement and non- instrumentality seem doubtful. The most remarkable thing about Alex's communicative abilities is that they have been used to show that he has built higher-level categories such as shape, color and number. This was the goal of Pepperberg, who wasn't trying to show that Alex could learn English, but rather to make "interspecies communication" possible.
In the final and 11th chapter, "Language, biology and evolution", Anderson insists on an argument proposed by Chomsky: "animals could not have the cognitive capacity to learn a language (...) in that they never display this capacity in nature" (307). Given that language is evolutionarily advantageous, the idea that animals are capable of language but do not use language is nonsensical. Thus, "language as we know it is a uniquely human capacity, determined by our biological nature" (307). Anderson then turns to the fashionable but disputed problem of the origins of language, noting that there are two answers, depending on whether one says that "the language ability arose during the course of evolution [as opposed to] through evolutionary mechanisms" (308). Chomsky is a representative of the first position, while Pinker, Jackendoff and others are representatives of the second, emphasizing communication as a selective pressure. Anderson then turns to speech, noting that though the descent of the larynx has been exaggerated since, as shown by Fitch, quite a few species descend their larynx when they vocalize, still the permanently low position of the human larynx may be the result of the evolution of language rather than the reverse. Anderson then rehearses the scenario advocated by Bickerton and adopted with modifications by Jackendoff, in terms of a protolanguage from which universal language eventually developed. Finally, Anderson argues that what makes language a unique and powerful tool is not only the possibility for symbolic reference, but syntax and since syntax is not evidenced in any other species, "it appears that the syntactic principles of Universal Grammar are a part of specifically human biology" (324).
The main interest on the book is not, I think, that Anderson's views on animal communication are original (indeed, his aim is a synthesis) or that his ideas about language are original (they can be found, mutatis mutandis, in quite a lot of linguistic books). Rather the uniqueness of Anderson's attempt resides in the fact that he clearly succeeds in illuminating language through his study of animal communication and animal communication through his study of language. If anything, his book endeavors and succeeds to show that each communicative system in the animal (including humans) world is unique and uniquely adapted to the needs of the species that has it. This is certainly also true of human language, though whether the needs it satisfies are communicative is debatable. It might be that language indeed has a more cognitive than communicative function, but this is another question.
Thus, reading Anderson's book is a highly satisfying experience. Each chapter (and some subsections) opens with a quotation taken from the Doctor Dolittle stories and the book as a whole is very readable. For a user-friendly introduction to animal communication and the uniqueness of the human language, one can hardly do better than read Anderson.
Hockett, Charles F. (1960) Logical considerations in the study of animal communication. In: Lanyon, W. E. & Tavolga, W. N., Animal sounds and animal communication, Washington D.C., American Institute of Biological Sciences, pp. 392-430.
ABOUT THE REVIEWER:
ABOUT THE REVIEWER
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 philosophical subjects.