The evolution of language is a hot topic. New books on the subject garner attention in the popular media, and their authors are sought for appearance on televised documentaries. Currently, the dominant view is that language evolved wholly within the hominid (human ancestral) lineage, whether beginning early in that lineage at millions of years ago (Pinker 1994) or only more recently with our own species (Bickerton 1995, Noble and Davidson 1996). Human language is thus sharply different from all types of animal communication. The contrasting, minority position (Savage-Rumbaugh et al. 1993, King 1994, Armstrong et al 1995) allows deeper roots for language and precursors to features of human language in animal communication systems.
Variations on these two views have been repeated for centuries. Theorists in the first group seize any new scrap of information about the unique properties of human language to bolster their discontinuity view, whereas theorists in the second group search for data from the animal world to bolster their continuity view. The whole enterprise thus begins to resemble an endless ping-pong match with back-and-forth debate but little productive dialogue.
A book of a different nature, having no steadfast commitment to spinning a scenario about human language is, in this present climate, both rare and much-needed. In The Evolution of Communication (TEOC), Marc Hauser is interested in understanding how human language arose, but for him that's only one of many fascinating questions about communication. In the best tradition of Niko Tinbergen, the Nobel Prize-winning ethologist who worked in mid-century, Hauser adopts four perspectives on the evolution of communication: the mechanistic, the ontogenetic, the functional, and the phylogenetic. Taken together, these perspectives "represent a coherent theoretical and methodological framework for both studying and explaining communication" (p. 2). As Hauser puts it, "My approach in this book, therefore, is to provide a framework for studying communication from a broad comparative perspective, among both closely and distantly related species to inspect how adaptive solutions to specific problems have evolved" (p. 5). He puts it even more succinctly a few pages later: "What is sorely needed is for communication to be treated on a broader phylogenetic scale" (p. 10).
Wisely, however, given the enormity of his topic, Hauser does focus his treatment. He writes only about natural communication systems that use auditory, visual, or audiovisual (and not chemical, tactile, and electrical) signals. He concentrates his analysis heavily on birds, anurans, and primates, using case studies of these creatures so that his readers "can learn to appreciate 'the best of from a given research area and hopefully understand how a problem is solved in the best of all worlds'" (p. 14). Using these guidelines, he organizes TEOC into 8 chapters: the evolution of communication: historical overview; conceptual issues in the study of communication; neurobiological design and communication; ontogenetic design and communication; adaptive design and communication; psychological design and communication; comparative communication: future directions.
Each chapter is divided into many titled subsections.
Hauser tends to write in outline style, sometimes sacrificing elegance for clarity: "At present, there seems to be evidence for all three perspectives, and I see at least two reasons for such evidence to exist. First. . ." After pages and pages of this, the reader may feel she is reading over old college lecture notes. But it's easy to forgive Hauser given the massive amount of material he synthesizes in this volume; the bibliography alone, at 73 pages of small-print entries (too many to count!), is a useful research tool.
What kind of order does Hauser impose on all this material? At the outset he notes his "bias" for discussing concepts like information and signal in terms of their "functional design features. Thus information is a feature of an interaction (i.e., not an abstraction...) between sender and perceiver. Signals carry certain kinds of informational content, which can be manipulated by the sender and differentially acted upon by the perceiver. Signals have been designed to serve particular functions, and the functions they serve must be evaluated in light of both production and perception constraints" (p. 6). Theevolution of signals is thus a subject at the heart of the book, but Hauser also takes care to differentiate among signals, cues, and signs.
Cues too have been designed to be informative, but unlike signals, they are permanently "on" (in the way an animal's coloration is often invariantly "on") and they carry no extra cost to the organism. Signs differ from both signals or cues in conveying information but not being designed for that purpose. A potential prey animal might, for instance, learn to "read" signs of danger in the environment, such as a predator's footprints, but those footprints were not made in order to convey information. By distinguishing among these three types of communication, Hauser lays the groundwork for a framework that analyzes not only the production and perceptual constraints of signallers but also the role of intention, deception, and the like in the evolution of communication.
Rather than attempting a straightforward summing up of results, I will focus in the remainder of this essay on four ways in which, for me, TEOC succeeds in illuminating the study of the evolution of communication. First, the compendium of facts is simply fun to read, especially for those who might tend to assume that the communication systems of primates are the best-designed or most efficient or most complex. At this relatively basic level of "natural history appreciation," I was interested to learn, for example, that in echolocation, the echoes returning to the bat supply it with information as detailed as the species of insect detected (p. 162). Nor had I known that in some mixed-species flocks of birds, some species specialize in alarm-calling and others in food-getting, but when competition for insect prey gets tight, the alarm-callers utter alarms (in the apparent absence of any need to do so), causing the food-getters to look up. The alarm callers then take this opportunity to snatch food (see p. 422 for Hauser's description of these data, taken from research by Munn).
Hauser describes intriguing examples such as these for a purpose, of course. TEOC's second strength is the way in which it highlights, in frequent summary sections, the most significant patterns in the evolution of communication. One example is the short section called "synthesis of studies on food-associated signals" (pp. 447-450). In it Hauser presents a chart of data about food-associated signals in various species, and from these data concludes that despite interspecific variation, some similarities emerge with regard to the function of such signals. Food-associated calls are involved in recruiting group members to the food source; a relationship exists between call rate and some measure of food quality, calling behavior is guided by social factors such as the benefits of forming large groups (p. 450). By providing such summaries, Hauser makes it easy for his readers to formulate ideas or questions of their own about the evolution of signalling, for instance about possible selection pressures.
Writing about calls given in a variety of contexts, Hauser asks whether they have informational content and tackles the difficult question of how to distinguish among semantic, referential, and representational signals (p. 507-8). Signals are semantic if they provide listeners with information about objects and events in the environment; signals are representational if listeners create some kind of mental picture of the object or event eliciting the signal; signals are referential if they are reliably associated with objects and events in the world. They three types are thus (to my understanding) not mutually exclusive, in that, for example, representational and referential signals would also seem to be semantic.
As Hauser notes, evidence for mental imagery in nonhuman animals is hard to come by, so representational signals are difficult to measure. As a result, the term referential is now favored by ethologists in discussing the information content of signals. They modify the term slightly, however, to talk about functionally referential signals, "to make clear that non-human animal calls are not exactly like human words, but rather appear to function in the same way... Although none of the functionally referential signals identified thus far have the expressive power of human words or, for that matter, the honeybee's dance language, such signals are more than mere expressions of emotional state" (p. 509). Hauser then reviews evidence for referential signals in animals as diverse as domestic chickens and vervet monkeys (he did field research in Kenya on the latter species). He emphasizes that the precise meaning of referential signals remains unclear, but what is also clear from this section of the book is that ethologists now know enough to deny the oft-repeated claim that animals signal only about their emotions.
Hauser goes much further with the referential material, linking it to questions of cognition. He asks, for example, how animals categorize the world, and discusses experiments designed to find out. I will return to these below, but will now consider the third strength of TEOC. As hinted above, it presents a refreshing approach to the evolution of language, both in considering what other theorists have written and in presenting Hauser's own thoughts on the topic. Hauser notes that we can learn the most about the "supremely difficult problem" (p. 32) of language evolution by focusing on the various theorists' strengths rather than on their differences: "I consider this to be a more productive approach, for it highlights an often-neglected feature of scientific progress: the conceptual development of thought based on a continuously developing set of empirical building blocks" (p. 33). Hauser then discusses, in turn, the views of linguists Chomsky, Bickerton, Lieberman, Hockett, and Pinker, and biologists Marler and Smith. Admirable as this stance is, Hauser can go too far when seeking strengths. He describes Pinker's perspective as "powerful and lucid, and (it) fits beautifully with the conceptual goals of this book, focusing on design features and Darwinian processes of evolution" (p. 49). Yet what happens, in PinkerŐs work, to the broad phylogenetic approach that Hauser touts? Pinker (1994) concludes that chimpanzees don't have language, but never seriously analyzes their systems of communication. He reduces their natural communication to hoots and shrieks (p.342), and dismisses the accomplishments of the enculturated apes used in research programs designed to assess their linguistic skills. He writes, for instance, that Kanzi - the enculturated bonobo who produces meaningful utterances using lexigram symbols on a computer, and who understands spoken English (Savage-Rumbaugh et al. 1993) - "learned to bang on visual symbols on a portable tablet" (1994:341). Pinker fashions himself a gradualist, championing natural selection as the agency of language evolution, but nonetheless insists on a sharp demarcation between the behaviors of apes and hominids.
Awe of Pinker aside, Hauser's balanced assessment of the evolution of language in this section and throughout TEOC leads to some important insights. No other subject so firmly divides the language theorists as syntax, which is seen as by most discontinuity theorists as a biologically based instinct, and by most continuity theories as another feature of language that could have evolved gradually. Hauser steps back from the fray, reminding us that the calls and songs assumed by researchers to be the fundamental units in an animal's repertoire may or may not actually be that. These assumptions are only rarely "substantiated by the relevant perceptual tests, designed to assess what individuals perceive as meaningful units. Until such experiments have been conducted, however, we will remain in the dark on a number of interesting problems, especially the problem of syntactically governed organization of elements within the repertoire" (p. 68). Hauser offers a few of his own ideas about evolution of language. Hominids were, early on, capable of rudimentary referential signalling. How this ability evolved into full human language, with its huge lexicon and "recursive system that could combine entries in the lexicon into an infinite variety of meaningful utterances" (p. 649), is the interesting question. The driving mechanism, Hauser posits, is the ability to imitate: "Imitation allows individuals to quickly pick up the meaning of a new signal and also provides a foundation for generating variation" (p. 650). Imitation can introduce new ways of doing things because individuals vary in their ability to precisely reproduce the behavior or skill in question. Hauser sees the ability to imitate, at least vocally, as a true divider between nonhuman primates and hominids.
TEOC's fourth successful feature is its call for a new research program concerning the evolution of communication. Hauser sprinkles ideas for hypotheses to test throughout the book, but in the final chapter, which is subtitled "future directions," he explicitly lays out a thoughtful agenda for research to come.
In general, Hauser wants "to develop tools that allow for direct interspecific comparisons, while remaining sensitive to the ecological and social factors that both were and currently are responsible for the design features of each species' communicative repertoire" (p. 609). Particularly pressing problems include ones I have mentioned above - determining call meaning and the functionally meaningful units of communication for animals without language. Hauser sees a marriage of ethology and neuroscience as the best way to proceed, aiming for "a more socio-ecologically rooted research agenda, driven by questions of design and the ability toprocess biologically meaningful stimuli" (p. 610).
One promising tool for the future is what's called the preferential looking procedure, characterized by Hauser as providing "an elegant entry into the mind of the nonlinguistic organism" (p. 634; he discusses the procedure at length in chapters 7 and 8). Developed for testing of prelinguistic human infants, this procedure relies on the fact that humans form expectations about the way things happen in our world. When our expectations are violated, we seek explanations - and we look longer at the event in question that causes the violation. Human infants under a year, for example, look longer when a solid object, a ball, is made to look as if it passes through a table, an impossible motion, than when the ball moves in a possible, normal way.
Variants of the preferential looking test have already been carried out on baboons and macaques (the latter done by Hauser and colleagues). The work with baboons (Cheney et al. cited in Hauser p. 557), for instance, asks whether impossible vocal behavior elicits a different response than possible vocal behavior. In playback experiments, the impossible vocal behavior was represented by a low-ranking female's grunt followed by a higher ranking female's fear bark. (Since this sequence has never been heard by researchers, it is considered impossible - females of high rank just donŐt give fear calls to females of lower rank.) The possible form added a third call to the end of this sequence: the grunt of an even higher-ranking female. As Hauser explains, "The inclusion of this third call made the sequence possible, since fear barks are responded to by grunts. Results show that individuals looked significantly longer following presentation of impossible vocal sequences than possible ones" (p. 558). Baboons apparently have, then, some notion of cause-effect, some understanding that cannot be accounted for simply by associational rules (for a full explanation of the logic of this conclusion, see p. 559). Hauser suggests specific ways in which the preferential looking test can be applied to learn more about primate cognition in a variety of species and behavioral contexts.
By making these types of recommendations, Hauser acknowledges the tight link that emerges in some signalling systems between communication and cognition. He acknowledges too that many discontinuity theorists argue that language and thought are co-extensive, meaning that before language arose, nothing akin to what we call "thinking" could have existed (e.g., Bickerton 1995, Noble and Davidson 1996). Yet Hauser concludes (p. 561) on the basis of preferential looking tests that nonhumans have "complex thoughts" without language, and on the basis of another series of experiments, that there is "evidence of abstract relational concepts in the absence of language, at least human language" (p. 561).
This latter experiment (Hauser et al in prep, cited p. 563) was carried out with cotton top tamarins. In part of this series, monkeys were trained on a task involving two pieces of cloth, physically separated, with food on one. The monkeys had to pull the correct cloth to get the food, with only one pull allowed per trial. In the novel trials after training, the monkeys reached almost 100% accuracy in pulling the cloth to get the food "even when such parameters as food distance, size, position, shape, and color were manipulated, in addition to alterations in the properties of the cloth..." The conclusion that this experiment allows Hauser to reach - that language and thought are not co-extensive - is consistent with those of other behavioral primatologists who review field data on primate cognition (e.g., Byrne 1995), and are highly significant for understanding the evolution of behavior in general.
As I have tried to demonstrate, TEOC is crammed full of information - including whole topics I've not even mentioned here, e.g., human signed languages, facial expressions across taxa, and evidence of functional vs intentional deception in animals. Hauser's organization of the material reviewed is masterful; his original ideas are not only solid, they are a must for anyone engaged in comparative work in communication. Further, in TEOC Hauser proves that writing about the evolution of language can be based in a broad ethological framework with insightful results. Read all 653 pages of text; you'll be well rewarded.
I thank Sherman Wilcox of the University of New Mexico for comments that improved this review.
Armstrong, David E., Stokoe, William C. and Wilcox, Sherman E. Gesture and the Nature of Language. Cambridge: Cambridge University Press, 1995.
Bickerton, Derek. Language and Human Behavior. Seattle: University of Washington Press, 1995.
Byrne, Richard W. The Thinking Ape: Evolutionary Origins of Intelligence. Oxford: Oxford University Press, 1994.
King, Barbara J. The Information Continuum: Social Information Transfer in Monkeys, Apes, and Hominids Santa Fe: School of American Research Press, 1994. Noble, William and Davidson, Iain. Human Evolution, Language and Mind. Cambridge: Cambridge University Press, 1996.
Pinker, Steven. The Language Instinct: How the Mind Creates Language. New York: William Morrow and Company, 1994.
Savage-Rumbaugh, E. Sue, Murphy, Jeannine, Sevcik, Rose A., Brakke, Karen E., Williams, Shelly L. and Rumbaugh, D. M. "Language comprehension in ape and child." Monographs of the Society for Research in Child Development 58, 1993 (3-4).
Barbara J. King is an Associate Professor of Anthropology at the College of William and Mary, Williamsburg, Virginia, USA. She has studied baboons in Amboseli National Park, Kenya, and apes in various captive settings. In addition to the book The Information Continuum, she has published two articles on the evolution of language in the journal Language & Communication. She is currently working on a book that analyzes how language theorists interpret data from primatology.