The first use of words by our early ancestors probably depended on four cognitive capacities: A rich conceptual understanding of the world around us; the ability to use and understand motivated signs, both icons and indices; the ability to imitate; the ability to infer the referential intentions of others. The latter three capacities are rare or absent in nonprimate mammals, but incipient in apes and well developed in modern humans. Before early humans could have begun to use words these capacities would have needed further development than is found in modern apes. It is not clear why selection favoured these skills more strongly in our ancestors than in the ancestors of apes.
AUTHOR'S RATIONALE FOR SOLICITING COMMENTARY: The search for the antecedents of language in primate communication may be misplaced; the launching of language may have had more to do with the unique cognitive abilities of our ancestors than with their earlier forms of communication. Based on insightful work of Paul Bloom and his colleagues on the cognitive capacities that underlie word learning, a plausible case can be made that the earliest use of word-like signals required selection that brought improvements in specific cognitive abilities: (1) a rich understanding of the external world; (2) the ability to infer the referential intentions of others; (3) the ability to imitate; (4) the ability to use motivated signs (i.e. icons and indices, whose form is not arbitrary); (5) the ability to use a sign in a wider variety of contexts than, for example, a vervet alarm call. These cognitive abilities are poorly developed in most mammals, incipient in apes, but well enough developed in humans to allow us to learn words.
Several questions arise from these considerations: (1) What encouraged selection for the cognitive capacities that are prerequisites to word learning even before words were used? (2) Why did these capacities develop more fully among our hominid ancestors than among the ancestors of the apes? (3) What were the mechanisms by which signs came to be used in such a wide range of contexts that it becomes reasonable to call them "words" and even "names"? The target article's goal is to elicit thoughtful responses to these three questions in the commentary.
1. The purpose of this contribution is to raise several questions about what may have been the earliest stage of language, the period when words, especially names for things, first began to be used. The bulk of the contribution explains why I believe the questions are interesting and significant. The importance I give to these questions may not be clear before the discussion, but it may still be useful to state them at the outset. My discussion, then, will lead to the following three questions: (1) What encouraged the evolution of the cognitive capacities that are prerequisite to word learning even before there were words to be learned? (2) Why did these capacities develop more fully among our hominid ancestors than among the ancestors of the apes? (3) By what mechanisms did signs come to be used in such a wide range of contexts that it becomes reasonable to call them "names"? I can speculate about possible answers to these three questions, but this paper is intended to raise the questions, rather than to answer them.
2. Linguists have often insisted that syntax is the most distinctive aspect of language, and this has encouraged the view that the appearance of syntax is the crucial step in the origin of language. Syntax would be useless without names for things, however, and names must have come first in evolution just as they come first for every child. It has always been tempting to compare names with the vocal signals of animals. Both are made with the voice and both convey meaning, but names differ from animal signals in several important ways: names are more dependent on learning and imitation than animal signals; they are conventional, which means that they can differ from one language and culture to another; they stand in mutual contrast, while many animal signals, notably those of the anthropoid apes, are graded (Marler 1976, Burling 1993); finally and most important, instead of being confined to a very specific situation, names can be used in an unlimited variety of circumstances. The vervet snake-alarm call can warn other animals that a snake has been spotted but it can be used for no other purpose. The English word "snake" can also act as a warning, but in addition, it can be used in such diverse sentences as "The snake has gone, Did you see the snake?, Where is the snake?, I don't like snakes", and so on without limit. Most often, no snake is even present when the word is used.
3. Captive chimpanzees have been observed to use a number of manual signals that share more features of words than do most other animal signals. The pokes by which infant chimpanzees indicate that they want to nurse are a good example (Tomasello, Gust, and Frost 1989). These pokes begin when the infant simply tries to push its mother's arm aside in order to get at the nipple. As mother and infant become accustomed to each other's behaviour, the push can be simplified and conventionalized until the infant finally needs only to poke the mother in a characteristic way, and she will adjust her position to allow her baby to nurse.
4. The most interesting aspect of these pokes is that different mother-infant pairs settle on different forms of pokes. This means that nursing pokes are conventionalized, while most primate signals are narrowly prescribed by inheritance. Nursing pokes also differ from most chimpanzee signals, and from our own giggles, laughs, and guffaws, in not grading into other signals (Marler 1976). Instead, they are sharply discrete, contrasting with other signals. Conventionalization and discreteness are not enough to turn nursing pokes into words, however, for they differ from words in two important ways. First, because they are not learned by imitation, they cannot be shared by an entire community or passed on to the next generation as part of a cultural tradition. Second, they cannot be used in the wide variety of contexts in which the words are used. Instead, they serve only as imperatives.
5. What would our early ancestors have needed in order to use signals that come closer to words than nursing pokes do? Hints may be found in the capacities needed by children in order to learn words. Bloom and Markson (1998) suggest that "children succeed at word learning because they possess a rich understanding of the external world, the ability to infer the referential intentions of others ('theory of mind') and, by the time the child is about two years of age, an appreciation of syntactic cues to word meaning." Of these three abilities, syntactic cues to word meaning could not have been used by the earliest word learners simply because there was no syntax quite yet. On the other hand, a rich understanding of the external world and the ability to infer the referential intentions of others must have been available from the start. To these two prerequisites to word learning, I would add two others. The earliest word using human beings must also have able to imitate and I believe they must have been able to use motivated signs (i.e. signs whose form is not arbitrary), including both icons and indices. It may be that imitation, motivated signs, and the ability to infer referential intentions can be subsumed under a broader concept of a theory of mind, but they seem sufficiently distinct to merit separate consideration. Each of these four cognitive prerequisites to word learning needs to be described briefly.
6. (a) A rich understanding of the external world. Before they start to learn words, children must be able to distinguish objects from one another and to group objects into classes. They must, for example, be able to recognize the similarity between Fido, Rover, and Rex which allows all three to be identified by the same word, "dog". At first, they may not get the classes exactly right. Many children start by using "dog" for any four legged animal, but this initial overgeneralization actually demonstrates a child's ability to classify even more clearly than would immediate conformity to the community's norm. Overgeneralization shows that children can invent their own classes, and only an understanding of the world and its objects would give children an idea of what sorts of things might be named. They will not be surprised to find words for "eye", "nose" and "mouth" but they will not expect distinctive words for "left eye" and "right eye". Animals must also deal consistently with various kinds of objects and actions in the world, of course, so the human ability to do so is, at least in some degree, a part of our mammalian heritage. The point is that even before children could learn names for things they would need to have some ideas about the things to which names can be given.
7. (b) Inference of referential intentions. Understanding that other people can attend to things, and that another person's attention can be directed differently than one's own, is a central component of any theory of mind. The full development of a theory of mind may require several years of childhood (Wellman 1993), but even very early word learners must recognize that two people can pay attention to different things. A child can follow the direction of a speaker's gaze, but will profit from doing so only if he understands that the speaker is able to attend to something different from what he has been attending to himself. Only by knowing what a speaker is attending to can a child correctly determine the referent of a previously unknown word (Bloom and Markson 1998).
8. (c) Icons and indices. The philosopher Charles Peirce famously classified signs into icons, indices, and symbols (Peirce 1940). Icons are signs that resemble the thing for which they stand. Pictures, diagrams, and onomatopoetic words are all icons. Indices are signs that point to their referent or that have some logical or physical association with the referent. Smoke is an index of a fire; my pointing hand can be an index of the thing to which it points. Both icons and indices can be said to be "motivated" rather than arbitrary because they have a logical association or physical similarity with the object for which they stand. Symbols, on the other hand, are signs that are arbitrary. Their forms are determined entirely by social convention. Modern spoken languages make only limited use of iconic or indexical signs, and most of our words are purely conventional symbols.
9. Elsewhere I have argued at length that icons and indexes very likely played a greater role in the earliest stages of language than they do in spoken language today (Burling, in press). All the early writing systems that we know about used pictographs, which are obvious icons (Boltz 1976, Kramer 1963), but with time the iconicity has been almost or completely lost. Present day sign languages of the deaf still use both indexical pointing gestures and hand shapes which iconically resemble the objects for which they stand, but the long term drift, even in sign language, has been away from iconicity and indexicality and toward increasing conventionality and arbitrariness (Klima and Bellugi 1979, Frishberg 1975). It seems likely that the earliest human language would also have made extensive use of both icons and indices. When first creating signs to be used in any conventional form of communication, iconicity and indexicality are the most obvious principles to exploit, and they were very likely more important when words were first used than they are in modern spoken language (See also, Armstrong, Stokoe, and Wilcox 1995).
10. (d) Imitation. The ability to imitate is essential to word learning today. It is implausible to suppose that anything we would want to call language could have been perpetuated in a community unless the individual speakers were able to imitate one another's signals. Imitation may not seem to be primarily a cognitive capacity, but imitation depends upon a recognition of a similarity between an observed act and one's own performance, and this recognition is certainly cognitive. The recognition of similarity in imitation recalls the recognition of iconic similarity, and perhaps imitation and iconicity rest upon the same or closely related cognitive capacities.
11. An understanding of the external world, the ability to infer another's referential intentions, the ability to use icons and indices, and imitation are all cognitive abilities. Early word learning would seem to have required all four, and this implies that they must all have developed even before word learning could begin. In order to behave appropriately, all mammals require a complex understanding of the external world. They must recognize objects and even classify the objects they encounter into familiar types if they are to act appropriately toward them. Perhaps we can attribute the rich human understanding of the external world to our mammalian heritage. The other three prerequisites to word learning, however, are rarely found among other mammals. Most mammals show no evidence of iconicity. Nor, with the remarkable exception of several species of cetaceans and seals (Tyack and Sayigh 1997), do most mammals show any ability for imitation, either vocal or gestural. It has proved very difficult to demonstrate a theory of mind even for monkeys (Cheney and Seyfarth 1990).
12. Apes, on the other hand, seem to show incipient ability in all three of these skills, although the exact extent to which these skills are developed among apes, and the status of a theory of mind more generally among apes, remain controversial. Some have been persuaded by the evidence of deception among chimpanzees (Byrne and Whitten 1988) that these animals must recognize that other animals can have different knowledge than they do themselves. Others, such as Heyes (1998) remain skeptical, but as the commentaries to Heyes's article make clear, her opinions hardly represent a consensus. There does seem to be general agreement that chimpanzees can follow one another's gaze, surely an important basis for recognizing what another animal is attending to. Apes have been observed to gesture iconically to one another (Tanner and Byrne 1996) and, as is well known, they are unusual in being able to recognize objects in pictures, a clear sign of iconic ability. The question of imitation by apes is complex and much debated (Heyes and Galef 1996, Heyes 1998), and some primatologists have been very sceptical of imitation, even among chimpanzees. This is partly because of a narrow definition of "imitation". Tomasello (1996), for example, wants to limit "imitation" to cases in which the imitator grasps the intentions of the observed animal. Mere copying of another's behaviour without using that behaviour to achieve the same goal is considered to be only "mimicry". By this definition, a parrot mimics human speech but does not imitate it. Similarly, a child who duplicates the gestures or speech sounds of another person, without recognizing their purpose or meaning, is only mimicking, not imitating. It does now seem clear that captive apes can learn to mimic gestures (Whiten and Custance 1996), even if imitation, in this special sense, has not been demonstrated. Vocal mimicry, to say nothing of vocal imitation, seems to play, at most, no more than a very minor role among any species of primate other than humans (Seyfarth and Cheney 1997).
13. Perhaps it is these capacities that have allowed some apes, notably a few bonobos, to learn words under human tutelage (Savage-Rumbaugh et al 1993; Savage-Rumbaugh and Lewin 1994), but their capacities do not seem to have reached the point where they can develop or sustain even a simple language in the wild. This would mean that the cognitive capacities of our ancestors would have had to move past the point that has been reached by modern apes before they could have sustained a tradition of words. My first question, therefore, is this: What evolutionary pressures encouraged selection for improved abilities in the four cognitive areas I have pointed to: an understanding of the world, the ability to infer another's intentions, imitation, and an ability to use icons and indices? These would have had to improve even before words began to be used.
14. It is not at all difficult to speculate about the reasons why these abilities would have been useful to our early ancestors, and why natural selection would have fostered their improvement, but I see no convincing reason why selection would have been more insistent among our own ancestors than among the ancestors of the apes. Inferring intentions, imitation, and the ability to use iconically or indexically motivated signs would appear to be as useful to tree-climbing apes as to the ground-dwelling and bipedal apes that became our ancestors, but only in our line did they improve sufficiently to allow names to be conventionalized and used. Since upright posture and the consequent freeing of the hands and arms for other uses than locomotion were the decisive changes that pushed the ancestors of apes and human beings to diverge, it is tempting to ask whether walking with erect posture, and the resulting opportunity for more manipulation and carrying, might have encouraged a greater reliance upon these cognitive skills. Could manipulation or throwing (Calvin 1983) have fostered cognitive skills that would have opened the way to word learning? Could even such a simple thing as a head balanced on the shoulders, where it no longer needs to be encumbered by enormous muscles, impose fewer limits on expansion of the brain? Would that permit a degree of cognitive development that remains closed to apes? I do not find such speculations satisfying, so my second question is this: Why did the cognitive capacities needed for language develop more fully among our own ancestors than among the ancestors of the great apes?
15. As I have already noted, words, in particular names for things, can be used in far more diverse ways than any known animal signals. A chimpanzee nursing poke, like a vervet snake alarm, is limited to a single situation. A human baby's "arms-up" gesture, in which the arms are held out from the body or stretched upward as a way of asking to be picked up, has the same kind of limitation. It is a conventional gesture with a clear and specific meaning, but it is no more a word than is a nursing poke. It is used in only one very specific circumstance and it could never be used in contexts that would be understood to mean "I don't want to be picked up" or "Did you pick up Billy?"
16. Children learn names for things before they use them in sentences, but even during the one word stage children are able to use their words to convey a variety of meanings. "Truck" might mean "that is a truck", "I want the truck", "Where is the truck?" and so forth. The intended meaning, of course, is made clear both by the situation in which the word is used and by the gestures and intonation that go along with it. Even a single word requires an infant to marshal a sort of proto-syntax--to combine the word with gestures and voice quality. If they were to be used more productively than allowed by the narrow limits of animals signs, even the first words in prehistory must have had some comparable support from the situation and from accompanying gestures or cries. This means that words can never have been used completely alone, and perhaps words could not have expanded to the full generalized usage of modern words until they could be joined together in pairs and groups.
17. Nevertheless, a case can be made that the capacity to use names in the unlimited variety of circumstances that we now use our words, was one of the most important steps, quite possibly the single most important step, in the evolution of the human ability to use language. Therefore, my third question is this: How did signs come to be used in such a wide range of circumstances that they deserve to be called names? How, in other words, did humans develop the ability to use names?
18. This target article began by indicating that it would raise questions rather than answering them. I hope that I have given sufficient background to suggest that my questions are interesting ones. I believe that reasonable answers to these questions would help us to understand how the earliest stage of language could have taken place.
ACKNOWLEDGEMENTS: The author would like to express his appreciation to the seven remarkably thoughtful and helpful reviewers whose opinions were solicited by PSYCOLOQUY: Ralph-Axel Mueller, Derek Bickerton, Paul Bloom, Iain Davidson, Barbara King, Chris Knight, and Michael Tomasello. They have pushed me to write a much better article than the one they first saw.
Armstrong, D. F., Stokoe, W. C., and Wilcox, S. E. 1995. Gesture and the Nature of Language. Cambridge: Cambridge University Press.
Bloom, P. & Markson, L. 1998. "Capacities underlying word learning." Trends in Cognitive Science, Vol. 2. No 2, 67-73.
Boltz, William G. 1986. "Early Chinese Writing." World Archaeology, 17, 420-436.
Burling, R. 1993. "Primate calls, human language, and nonverbal communication". Current Anthropology, 34, 25-53.
Burling, R. In press. "Motivation, conventionalization, and arbitrariness in the origin of language. In Barbara J. King (ed.). The Origins of Language: What Nonhuman Primates Can Tell Us. Santa Fe: School of American Research Press.
Byrne, R. W. & Whiten, A. (eds.). 1988. Machiavellian Intelligence. Oxford: Clarendon Press.
Calvin, W. H. 1983. "A stone's thrown and its launch window: timing precision and its implications for language and hominid brains." Journal of Theoretical Biology, 104, 121-135.
Cheney, D. L., and Seyfarth, R. M. 1990. How Monkeys See the World: Inside the Mind of Another Species. Chicago: University of Chicago Press.
Frishberg, N. 1975. "Arbitrariness and iconicity: historical change in American sign language." Language, 51, 696-719.
Hayes, K. J. & Nissen, C. H. 1971. "Higher mental functions of a home-raised chimpanzee." In A. M. Schrier & F. Stollnitz (eds.). Behaviour of Nonhuman Primates. New York: Academic Press.
Heyes, C. M. 1998. "Theory of mind in nonhuman primates." Behavioural and Brain Sciences, 21, 101-148.
Heyes, C. M. & Galef, Bennett G. Jr. 1996. Social Learning in Animals: The Roots of Culture. San Diego: Academic Press.
Klima, E. S., & Bellugi, U. 1979. The Signs of Language. Cambridge, Massachusetts: Harvard University Press.
Kramer, S. N. 1963. The Sumerians: Their History, Culture and Character. Chicago: University of Chicago Press.
Marler, P. 1976. "Social organization, communication, and graded signals: The chimpanzee and the gorilla". In P.P. Bateson and R. A. Hinde (eds.) Growing Points in Ethology. Cambridge: Cambridge University Press. 239-80.
Peirce, C. S. 1940. "Logic as semiotic: the theory of signs." In: J. Buchler (ed.). The Philosophical Writings of Peirce. Dover edition, 1955. New York: Dover. 98-119.
Seyfarth, R. M., and Cheney, D. L. 1997. "Some general features of vocal development in nonhuman primates." In Charles T. Snowdon & Martine Hausberger (eds.). Social Influences on Vocal Development. Cambridge: Cambridge University Press. 249-273.
Savage-Rumbaugh, E. S., Murphey, J., Sevcik, R. A. Brakke, K. E., Williams, S. L., & Rumbaugh, D. M. 1993. Language Comprehension in Ape and Child. Monographs of the Society of Research in Child Development. Serial No. 233. 58, 3-4.
Savage-Rumbaugh, E. S. & Lewin, R. 1994. Kanzi: The Ape at the Brink of the Human Mind. New York: John Wiley & Sons.
Tanner, J. E. & Byrne, R. W. 1996. "Representation of action through iconic gesture in a captive lowland gorilla". Current Anthropology, 37, 162-173.
Tomasello, M., Gust, D., & Frost, G. T. 1989. "A longitudinal investigation of gestural communication in young chimpanzees." Primates, 30, 35-50.
Tomasello, M. 1996. Do apes ape? In C. Heyes and B. Galef (eds.) Social Learning in Animals: The Roots of Culture. San Diego: Academic Press.
Tyack, P. L. & Laela S. S. "Vocal learning in cetaceans." In Charles T. Snowdon & Martine Hausberger (eds.). Social Influences on Vocal Development. Cambridge: Cambridge University Press. 208-233.
Wellman, H. M. 1993. "Early understanding of mind; the normal case." In Simon Baron-Cohen, Helen Tager-Flusberg, and Donald J. Cohen (eds.). Understanding Other Minds. Oxford and New York: Oxford University Press. 10-39.
Whiten, A. & Custance, D. 1996. "Studies of Imitation in Chimpanzees and Children". In C. Heyes and B. Galef (eds.) Social Learning in Animals: The Roots of Culture. San Diego: Academic Press.