CHAPTER 21

LANGUAGE IN THE LIGHT OF EVOLUTION*

ERIC H. LENNEBERG

LIMITATIONS ON INFERENCES FROM ANIMAL COMPARISON

Most vertebrate species emit some kind of acoustic signal, and the sensory receptors of each species are sensitive to the broadcasts of their own kind. The ubiquity of this phenomenon suggests that some biological functions are served by it. However, we do not know whether these functions are the same for all species—in fact, there is good evidence against this. Acoustic broadcasting may serve to warn territorial intruders, to call and attract sexual mates, to threaten adversaries, to lure bait, to call the young, to transfer information; it may function to strengthen social cohesion in large groups, or to prevent the breaking up of single couples only; it may have the effect of arousing or of lulling; it may be directed at members of other species, at members of the same species, at only certain individuals, or only to the self, as in echolocation.

Animal communication does not merely fascinate us as a zoological phenomenon; it also lures us to believe that appropriate comparative studies will reveal to us the origin of human communication. The rationale here is approximately this: Since Darwin has shown that man is not the product of special creation but has descended from more primitive animal forms, neither his structure nor his behavior are special creations either. His forms of communication must have descended from primitive animal forms of communication, and a study of the latter is likely to disclose that there is indeed a straight line of evolution of this feature. This type of reasoning I shall call the continuity theory of language development. First, let me examine this critically. Then I will propose a discontinuity theory and show that it is not only biologically acceptable but, in fact, is more in line with present theories in developmental biology than the former theory (Roe and Simpson, 1958; Simpson, 1949; Haldane, 1949; Rensch, 1954).

Straight-Line Evolution of Language with Only Quantitative Changes

This type of theory rests on the belief that there is no essential difference between man’s language and the communication of lower forms. Man’s noises just sound different, and his repertoire of messages is merely much larger than that of animals, presumably because of a quantitative increase in nonspecific intelligence. Theorists of this persuasion might picture the development of communication systems in the animal world as a straight road toward language with various animal communication systems as early way stations. Human language is thought to be much more advanced, perhaps by virtue of some kind of “proliferation of elements” (more memory units, or more classification devices, or more computing elements).

It can be only this kind of implicit belief that encourages investigators to count the number of “words” in the language of gibbons, or to look for “phonemes” in the vocalizations of monkeys or songs of birds, or to collect the “morphemes” in the communication systems of bees and ants. The technical definition of these concepts is such that they cannot be applied to anything but natural, human languages. Not all students of animal communication are engaged in such endeavors; but often enough the underlying faith still appears to be the same, since much time and effort is spent to teach parrots, dolphins, or chimpanzee infants to speak English.

At the root of the notion that human language is merely quantitatively different from animal “language” is the idea that all animals have something that might be called “nonspecific intelligence,” but that man has much more of this endowment and that this intellectual potential happens to be useful in the elaboration of a universal biological need for communication. Animals are thought to be unable to learn to understand English because of an insufficiency of this intellectual capacity. There are grave difficulties with this reasoning.

Intelligence or intellectual capacity is difficult to define in the context of general zoology. Insofar as intelligence is a measurable property within our own species (and there are those who have their doubts about this), we have seen that it correlates poorly with language capacity. Within certain IQ ranges there is virtually no correlation whatever; and in the extreme low range, where there is an apparent correlation, it is rare to find individuals who have not even the capacity to understand simple spoken language. Most idiots and even imbeciles may be given verbal commands, and many also acquire, spontaneously, the use of some words or even simple phrases. When the concept of intelligence must be applied to disparate species, the problem of scaling and measurement is enhanced greatly. Clearly, intelligence is not a physical property that can be measured objectively. It is always tied to specific tasks and to the frame of reference of a given species. When we test different species by requiring animals to solve a certain problem, the similarity in task is seen by us, the human experimenter, but different species are likely to “interpret” an apparently similar task in their own, species-specific mode.

Comparing the intelligence of different species is comparable to making relative measurements in different universes and comparing the results in absolute terms. When we say that a cat is more intelligent than a mouse and a dog more intelligent than a cat, we do not mean that the one can catch the other by superior cunning but that one solves human tasks with greater ability than the other. The mode of “interpreting” a problem situation becomes more and more similar to that of humans as the experimental animal is phylogenetically closer to man. But from this we cannot infer that language acquisition is just another problem-solving experiment and that phylogenetic proximity to man increases the capacity for language.

In man, the ability to acquire language appears to be relatively independent of his own ability to “solve problems,” i.e., of his type of “nonspecific intelligence.” Why, therefore, should we expect that an animal’s ability to solve human problems is relevant to his ability to acquire verbal behavior? In most animals the “cognitive strategy” for solving a given problem is quite different from that used by man (Uexküll, 1921).

It may seem as if the cross-species comparisons of cognitive function and behavior by Harlow (1949 and 1958), Schrier et al. (1965), David D. Smith (1965), and by Rensch (1959, 1964) and their students were contradictions to my assertions. Actually, they are not contrary evidence but evidence that is not relevant to language acquisition. Let us picture the various skills that are relevant to communication as overlapping maps. There are some common skills as well as specializations. Let us assume that man’s language is closely tied to his cognitive structure. One might diagram the cognitive structures of other animals also as overlapping but not coterminous maps so that each has its own peculiar deviation. Suppose that language is in that part of man’s cognitive realm which most diverges from the “common region.” We see that it may take more than overlap to be capable of learning to speak.

Straight-Line Evolution of Complexity by Stepwise Accretion

Proponents of theories of this type admit of qualitative differences between human and animal communication, but they also believe that the extant communication behavior of animals has a discernible and continuous history. Language is seen as a complex of more or less independent features each with its own history. In the course of evolution more and more features developed and were added to the structure of communication behavior, but, because of the various fates of individual species and phylogenetic offshoots, there are a number of “missing links” or empty cells, so to speak.

Thus, one finds zoologists who are concerned with what they consider to be the biological prerequisites for speech and language and who search for each of these prerequisites independently throughout the animal kingdom. For instance, O. Koehler (1951, 1952, 1954a, b) believes that there are at least nineteen biological prerequisites for language. Because of felicitous circumstances all of these nineteen prerequisites are present in man. Except for one or two, the prerequisites are common zoological characteristics which man has preserved owing to his animal nature. No animal has at its disposal even all of those prerequisities which are not specifically human. A given species may have just a few of them—not enough to learn to understand or to speak—whereas a few species have so many of the prerequisites that they are either able to reach the lowest stage of human language-1earning (parrots) or engage in behavior that is an excellent parallel of human language (such as v. Frisch’s honey bees). Koehler proposes that one of the first prerequisites of language is the existence of concepts (unbenanntes Denken), i.e., unnamed thoughts. In a great number of experiments he has shown that many birds and mammals are capable of “counting” at least to three and many up to seven or even eight but none beyond. This suggests to Koehler that a number concept is present and that this concept is practically universal among higher animals. The language of the bees and the research on bird navigation indicate that spatial concepts are also widespread. Thus all or most animals have unbenanntes Denken according to Koehler. Man has a peculiar skill in attaching symbols or names to these concepts which Koehler considers to be the essence of language. But he feels that even in this skill man is not totally alone. Parrots can also name concepts, that is, are supposedly able to learn the meaning of a few words; and rudiments of the same skill are also seen by Koehler in aspects of the bees’ communication system.

Speech-motor skills are innate in man, but biologically they are no innovation, because some animals can learn to say things. Also, the ontogenetic development of vocalizations in man has parallels in birds; just as birds go through characteristic song stages after hatching, human infants go through characteristic stages of vocalization. The onset of words Koehler explains by an essential law of effect. The infant notices the results or effects of his crying and babbling and thus begins to make use of these vocalizations in order to bring about certain consequences. The early history of vocalizations and the beginning of language, he thinks, are identical with developments in certain birds and are thereby evidence for the biological nature of these phenomena. Man and his language differ from animals and their communication (1) by degree of certain universal skills, particularly a nonspecific learning ability and (2) by accretion of new skills such as man’s ability to combine and permute the named concepts, that is, words. So much about Koehler’s views.

I share some of Koehler’s beliefs, but not all of them. That is beside the point. What is important here is that his views necessarily imply that language is not a unique and integral behavioral development but a conglomeration of skills and abilities each of which has its own, independent phylogenetic history. Except for one aspect of language, verbal behavior is a continuation and amplification of ubiquitous zoological properties. There is a suggestion here of continuity with only a few recent innovations that lifted an earlier type of communication into the realm of human language. I reject this type of continuity theory on several counts.

1. The prerequisite skills for language can in only a few cases be shown to have a fully documented phylogenetic history that reaches up to Homo sapiens. Actually, this is the exception. Continuity theories are bolstered by citing examples from all over the animal kingdom in complete disregard for phylogenetic proximity to man. One parallel comes from birds; the next from insects; another from fish; still another from aquatic mammals. Frequently only one species within a given genus or family even possesses the trait, indicating clearly that we are dealing with species specificities, probably all of comparatively recent date. The reason the examples are so disparate is that parallels are rare. This suggests accidental convergence (if, indeed, it is even that) rather than milestones within one continuous phylogeny.

2. Far from being proof that there is anything resembling a continuous history, the examples customarily cited might also serve as evidence of discontinuities of skills and behavior patterns because of the sporadic occurrence on the branches of the phylogenetic tree.

3. Language is not a loose association of relatively independent abilities. There is no evidence that language comes about by a gradual accretion of skills. If it did, we should be able to see all but a few of such skills in our closest relatives; further relatives should show fewer, and so on down the line of evolution. Nothing like this seems to be the case.

4. What is thought to be the beginning of language in parrots, monkeys, or dolphins, is empirically totally different from the beginnings of language in the human infant.* At the most primitive stages of language acquisition, man does not imitate sounds, words, or sentences, but generates novel sound sequences that are recognized as speech and language because the rules of generation bear certain formal similarities to those of the standard language. A healthy child does not ordinarily parrot (or at least no more often than at special occasions). The outstanding characteristics of language are the all-pervasive principles of productivity. These principles are totally lacking from the examples of animal communication.

Another line of thought, different from O. Koehler’s in approach but similar in theoretical structure, was contributed by Hockett (1960) and applied to animal communication by Altmann (in press) and other zoologists (see also Marler, 1961). Hockett also begins with an analysis of language in terms of what I shall call for the moment essential attributes, and he then examines a great variety of animal communication systems with a view to discovering how many and which of the essential attributes of human language are discernible in the communication of other species. In contrast to Koehler, his attributes are almost entirely of a logical nature (i.e., not physiological or psychological). He calls them design features, a terminology that expresses well the intent of the investigation: it is a study of the efficiency and effect of the communication system, the result and outcome, so to speak, of behavior rather than the mechanism of the behavior itself. I believe this approach is an innovation in biological investigations, and it is apt to focus attention on many interesting aspects of communication, including the underscoring of parallel but different developments and phenotypic convergences by very different means. For instance, some of the design features that characterize language (Hockett distinguishes thirteen) are also characteristic of the so-called language of the honeybee (broadcast, rapid fading, total feedback, and perhaps specialization and discreteness), but the physical means used for the incorporation of these design features into “bee language” are quite different from those of human language.

This is an important point. A study of design features may give us insight into some of the biases that enter into the process of natural selection, into the biological usefulness of certain features of animal communication, but it is not relevant to the reconstruction of phylogenetic history. For the latter we are interested only in the relation of types of anatomical structure (including molecular structure) and physiological function (including motor coordination and sensory acuity), but we disregard the usefulness or efficiency of these features to the contemporary form. Thus, whatever similarities exist on the surface between dolphins and fish, shrews and rodents, bats and birds, pandas and bears must be ignored in our attempts to reconstruct the respective phylogenies; and, in fact, the more abstract and pragmatic our criteria for comparison are, the less relevant will they be to a reconstruction of phylogenetic history. For instance, among the most abstract pragmatic criteria is successful adaptation to the environment; this may be accomplished by an apparent infinity of means. If we could rank-order adaptation in terms of success, it might tell us something about life in general, but it would tell us little about phyletic descent.

The converse of this argument is also true. Suppose we are interested in locomotion. While it may be quite revealing to study such logical design features as (1) range of speeds, (2) the radius of an individual’s movements, and (3) endurance, it will be immediately obvious that commonality in any of these criteria does not define phyletic relatedness. Thus we must guard against the application of Hockett’s design features—interesting as they are for certain purposes —to any argument concerning the evolution of language.

Justification for a Discontinuity Theory of Language Evolution

Search for True Antecedents. A discontinuity theory is not the same as a special-creation theory. No biological phenomenon is without antecedents. The question is, how obvious are the antecedents of the human propensity for language? It is my opinion that they are not in the least obvious. The fact that most vertebrates vocalize is not very informative. There is every indication that different species have adapted this common feature to very different functions, that they have carved out specializations from a common potential. No living animal represents a direct primitive ancestor of our own kind, and, therefore, there is no reason to believe that any one of their traits is a primitive form of any one of our traits.

The noise-making aspect of language is only one feature, and, at least today, an incidental feature, of our form of communication (the deaf have language without noise-receiving or making). Clearly there are other processes involved in language, and the history of these is no less important than that of vocalization. It is, for instance, entirely possible that certain specific principles of categorization and recombination which we encounter again and again in the perception of speech as well as in its production, in phonology, in syntax, and in semantics, are modifications of physiological principles evident in motor coordination. The ability to name may be related to perceptual and modified neurophysiological processes. Certain innate neurophysiological rhythmic activities might have been adapted to subserve speech in a highly specialized way. These remarks are speculative and merely serve to point out that the range of possible antecedents is vast and that in addition to the abstract and logical aspects of language (most often discussed now in the literature) there are also physiological prerequisites for speech and language.

Phylogenetic Change. If our time perspective is deep enough, all species are in a continuous state of change with respect to structure and function. But there is great variation in rates of change. Some species have not changed their structure appreciably for many millions of years; others have undergone relatively rapid change preceded or followed by periods of relative constancy. Also, individual aspects of species may have different rates of evolution. For instance, skeletal structure may very well be less susceptible to modifying influences than behavioral adaptations. But change, that is, evolution, is a continuous, ever-present phenomenon in all respects of life.

Evolutionary change is attributed to two general types of principles. First is the relative instability of genetic, replicative processes within cells; this provides the raw materials for potential change. Second, selective biases operate to allow some variations to remain while others are eliminated. The variations due to imperfections in the replication process need not have any direction, though ideal and perfect randomness is also very unlikely because the molecular structure of the genie material probably always favors some variations over others (cf. Waddington’s (1957) principle of “canalization”). However, it is nowadays assumed by most students of the genetic basis of evolution that the variations due to imperfection of replication would be rich enough to have a self-canceling effect, counteracting one another in such a way that the end result of evolution would be erosion of all species specificities and general leveling of all characteristics (death of species, in other words) were it not for the biases of natural selection, which tend to preserve some variations but not others.

The source material for reconstruction of the history of changes is primarily fossilization; most aspects including “soft” anatomy, physiological functions, and behavioral traits do not fossilize, and so we have a very imperfect record of evolutionary changes of species in their actual integrated, whole form and function.

The continuity of evolutionary change is not identical with the phenomenon of creation of new species or branching. The latter may be viewed as a possible but not necessary by-product of the former. If the replicative process is sufficiently unstable and selection pressures are sufficiently high and specific, the species may undergo rapid alterations, but this does not result in branching out into a radiation of new species unless certain conditions affecting population mechanics are present. The latter are variables which are quite independent of variables controlling mutations or the nature of selection pressures (for details see Mayr, 1963). Branching is merely a phase in the total history of changes and probably in many instances a relatively short-lived one.

If evolutionary change proceeded at a constant rate, then the gaps between extant species would be directly proportional to the age of the branching that separated any two species. Since branching occurs at irregular times, we would expect a great variation in the size of gaps between any two species and the existence of discontinuities should not surprise us in the least. In fact, the rate of evolution varies quite considerably, and this enhances the spread in size of discontinuities even further. The phylogenies of taxonomie groups may be reconstructed by ordering the size of the discontinuities. This principle, together with the fossil history, yields a hypothetical history of man from which one may expect considerable gaps between many aspects of Homo and the other Hominoidea. According to Dobzhansky (1962) and most primatologists, the present properties of our species are the result of so-called phyletic evolution, that is, changes without branching. There was enough time in the course of mans speciation for such changes to occur.

Although it is quite conceivable that behavioral propensities yield more readily to selection pressures and are, perhaps, also more easily affected by genetically conditioned variations, thus changing at a more rapid rate than skeletal structure, for instance, we must still remember that all evolutionary changes affect animals as a whole. Entire patterns of life, so to speak, are altered, but at each time slice there is, necessarily, full integration and mutually adaptive interaction of all of the animal’s features; it is the condition for viability and successful continuation of the species. This consideration has an important consequence for reasonable expectations of the phylogenetic history of one specific trait, such as human language. Individual traits of an extant species can never have a continuous history because they do not evolve independently from the rest of the animal. Thus we see that there is every reason to believe that animal communication is a discontinuous affair and that logical commonalities among communication systems are not necessarily indicators of a common biological origin.

Sharing of Traits. These assertions are not contradicted by the wealth of evidence that some sort of symbolic behavior can be demonstrated in a wide variety of animals, that the communication of affect is very common, or that territoriality, protection of the young, or maternal behavior is frequently accompanied by vocalizations. It is a truism to point out that animals share certain traits; it follows directly from the tree-1ike relationship between species. Notice, however, that the phylogenetic relationship between species cannot, in most cases, be represented by a single, unique tree diagram that accounts for absolutely all of the commonalities and all of the specific differences. A tree that characterizes the relationship of skeletal structures of certain species fairly well may differ in some (usually small) respects from a tree that characterizes the relationship between given protein structures (Goodman, 1963). This is due to a number of circumstances; for example, certain aspects of life do not allow as many (or any) variations as others; or there may be only one or very few possible biochemical solutions to a given problem posed by the environment so that a similar condition comes about more than once throughout the animal kingdom; or certain features are lost or added by individual species.

PROBLEMS CONCERNING THE RECONSTRUCTION OF THE HISTORY OF LANGUAGE

There were days when learned treatises on the origin of language were written based on nothing more than imagination. The absence of ascertainable facts rendered these essays disreputable early during the rise of empirical sciences. For some time the topic became taboo in respectable scientific circles. But recently it seems to have acquired new probity by adumbrating the speculations with empirical data. Let us test the soundness of the various types of arguments by examining the corroborative evidence in terms of relevance to the problem of the phylogenetic history of language.

Arguments Based on the History of the Brain and Skull

In contrast to the prior part of this chapter, in which attempts to derive the biological origin of language from a comparison of animal communication were discussed, we shall now deal with efforts to reconstruct language history through a reconstruction of brain history.

Since the brain does not fossilize, its history is based on secondary indications, either by a comparison of the brains of present-day animals or by a study of the bony enclosure of the brains of extinct forms. We must examine these two sources individually.

Comparison with Brains of Contemporary Animals. From what we have said above it is obvious that we need not compare the brains of any other species but those of primates, because the hope of encountering behavioral mechanisms that are directly related to human language fades with phylogenetic distance from man. There is fair agreement on the relationship between species within the order, so that we may confine ourselves to the family most closely related to us, namely the great apes (Pongidae). Unfortunately there are still vast gaps in our knowledge of the comparative neuroanatomy of these forms. The literature on cerebral cortex (von Bonin and Bailey, 1961), basal ganglia (Feremutsch, 1961), thalamus (Feremutsch, 1963), and autonomic nervous system (Wrete, 1962) has recently been reviewed, and the general phylogenetic trends are discussed there (D. Starck, 1965; see also Connolly, 1950).

However, the history of the human brain is far from clear. Many aspects, and perhaps the most important ones for an understanding of language, remain completely unexplored from a comparative point of view; an example is the connectivity between cells and cell aggregates. Furthermore, not all details of the brains examined fall into a linear order of evolution, because we are not dealing with primitive and primordial but with contemporary and adapted brains. Some of the peculiarities of the human brain are predictable on the basis of allometry and may therefore be attributed simply to general growth factors instead of to behavioral specializations. The greatest problem in this connection arises from the uncertainty about neurological correlates of language. It is not at all clear that the capacity for language depends on any grossly observable structural peculiarity of the brain —not even the central region of the left cerebral cortex; because if this part of the brain is removed surgically early enough in life, language may develop without impairment through specialization of other areas. Thus it is perhaps only molecular structure which affects function, especially relative speeds of conduction in various parts of fiber systems, that is relevant to brain history of language capacity. At the present time nothing is known about this.

There are some very general, methodological questions regarding the usefulness of comparative neuroanatomy for an elucidation of the phylogenetic emergence of behavior. In a sense species make different “use” of particular brain structures in the elaboration of their speciesspecific behavior. The anatomy of the visual system of mammals differs quantitatively rather than qualitatively. However, the destruction of large parts of the area striata appears to have different consequences for chimpanzee than for cats, and the peculiarities of pattern perception and the recognition of similarities characteristic for a given species can never be explained on the grounds of neuroanatomy. This type of phenomenon makes it difficult to say that “language could only have come about after a certain type of fiber-connection had developed or a given cortical area had expanded.” Language is the end product of many interacting processes depending on a variety of cerebral mechanisms. But the now-existing associations of language with central-nervous-system peculiarities do not mean an evolutionarily inevitable and necessary relationship. Similar behavior might have come about in different combinations of ways. Earlier forms of communication might have implicated other brain characteristics. It is an ad hominem argument to say that language is “due” to a given brain development, just as it would be inaccurate to say that man is a poor swimmer because of his lack of fins, scales, and fish brains. This is not the reason (many animals live in water without these attributes). There is no other reason but that the phylogenetic history of man did not adapt him for aquatic life. We may say that today the capacity for language is dependent upon a human brain, but we cannot write a causal history of this relationship (see also Overhage, 1959).

In short, the evolutionary history of mans cerebral capacity for language cannot be easily elucidated by examining the brains of other living species. Probably man was separated from other primates long before his brain began to evolve in the direction of the language prerequisites.

Relevance of the History of the Skull. Could the paleontological history of the human skull give us clues about the emergence of language? Critical evaluation of the evidence speaks against this possibility. Instead of brains from fossil men, we have only an array of skull fragments. From these fragments endocasts are made, that is, plaster casts of the concavities. Unfortunately, the patterns of cortical sulci are not well delineated on the calvarium. The main landmarks inside the extant bones are the meningeal vessels, which, in modern man, do not bear a constant relationship to fissures and sulci of the subpial cortical surface. Thus, endocasts give no language-relevant information. They tell us something about the brain’s approximate size and shape, but nothing about cortical fields, subcortical connections, or other internal structure.

If a group of experts were handed fragments of the cranium (or the entire exenterated skull, for that matter) of modern man, say, an individual who died three years before under unknown circumstances, and were asked to say whether or not that person had acquired language, they could only give an answer in terms of statistical probabilities, namely, the incidence of persons who do and who do not have language in the presence of normal bone formation. Their judgment would be based entirely on their knowledge of conditions prevailing today among the population at large. The bones themselves could give them no clue as to the language capacities of the deceased. He might have been a normocephalic retardate or a great orator, or, if the skull is subnormally small, it might have belonged to an African pygmy or a birdheaded dwarf—in either case in possession of language—or it might have been a microcephalic individual with no more than a few words at his command. If one cannot make unfailing deductions from recent bones, one can hardly presume that inferences are possible about fossil men who lived under almost totally unknown circumstances.

The extraordinary size of modern man’s brain and the cause of its relatively rapid phylogenetic development have captured the imagination of virtually every student of human descent. This evolutionary event has been generally linked to our capacity for language. Although I tend to be skeptical about this relationship, based on the evidence mentioned elsewhere, I admit that there can be no proof for the historical independence of brain size and language.

The oldest fossil which some authorities are willing to regard as the first human ancestor is Australopithecus, who lived one to two million years ago (there is considerable uncertainty about the age). His brain was about as large as that of a modern gorilla, but he was slighter in build than the modern ape, so that he might have had a relatively large brain. If we are right in classifying him into a family distinct from the Pongidae, we must also accord him some peculiar brain functions. If the modern apes do not talk, this is no evidence against the possibility that Australopithecus had some potential for a primitive form of speech-1ike communication. A study of his skull cannot decide the question. Later, more distinctly hominid fossils, such as Java man and Peking man, had successively larger cranial capacities. Toward the end of the third glacial, fairly suddenly, a new race appears with brains as large as ours: Neanderthal man. It appears to be the consensus today that this form was a racial specialization that became extinct without affecting the ancestral line from which we have come. Our direct forbears, the Cro-Magnon race, emerged about fifty thousand years ago. Their brains had the same size as Neanderthal’s; in fact, on the average, the cranial capacity was slightly larger than that of modern man. Cro-Magnon’s skull had one characteristic in which it departed from all other types of human fossils: its shape. The skull was shorter but higher and there was a forward shift of its center of gravity; it was balanced differently on the spinal column.

Haldane (1949) and Mayr (1963) have pointed out that the rise of modern man, particularly the increase in cranial capacity, has occurred within a spectacularly short period. While evolutionary changes are often measured in terms of millions of years and a time span of a hundred million years for some given change is quite common, the most significant hominid feature, the increase of the brain, took place within a mere few hundred thousand years. One may wish to speculate on the target of the selection pressures at work here. What was so advantageous about a large brain and skull?

It is tempting to relate the size of the brain to man’s two most outstanding characteristics: his capacity for language and his general cognitive capacities. Intuitively this relation may be reasonable. But it is important to remember that it rests on no more than just that: intuition. There is no way of demonstrating that cognitive or language capacities either required or resulted from a rapid increase in the number of brain cells.

One common line of argument in favor of relating brain size to intelligence and language is based on observations on feeble-minded individuals. Here it is not uncommon that abnormally small brains are correlated with a lowering of intellect, and language-1earning ability may be affected also. However, feeble-minded patients are not replicas of primitive human races; they are not a viable subspecies. Their constitution and growth patterns are deeply abnormal; their brain functions have not developed properly; and little is known about the quantitative aspects of their brain-cell populations. These deviants can tell us nothing about evolutionary history of the brain. Furthermore, there are abnormally small human brains (namely, of nanocephalic dwarfs) that are capable of learning language, and the surgical removal of up to one-third of the cerebral mass early in childhood does not restrict the capacity for language acquisition. Dart (1956) cites further anthropological data which show that individuals with much smaller brains than those normally seen in Europe, America, and Asia can learn language, and Schultz (1962) describes the skull of a gorilla that had a 752 cm³ capacity, i.e., within the range in which several microcephalic humans have an appreciable degree of linguistic facility. There is, then, little that compels us to think of language or communication in general as the prime selection target responsible for the present size of our brain (for a well-argued opposite point of view see Hockett and Ascher, 1964). To postulate a given brain size as the Rubicon for the capacity to speak (Keith, 1948) does not appear to be justified.

There is, however, another line of argument that induces many scholars to suspect a close relationship between brain size and intelligence. It is based on purely logical considerations; in fact the reasoning underlying it is by analogy. The capacities of an electronic computer or desk calculator are directly related to the number of its constituent elements. This engenders the belief that an increase in the number of units in the brain has a similar consequence. However, evidence for this is surprisingly poor. Perhaps Lashley’s early observations might be cited in support of this contention, namely that the quantity of cerebro-cortical destruction is inversely related to the complexity of task and pattern perception an animal is capable of. But it is not clear how these studies relate to human intelligence. For instance, Teuber (1959) and Ghent, Mishkin, and Teuber (1962) have compared intelligence test scores of war veterans at the time of their recruitment with scores on the same test several years later and after extensive brain injuries; there was no significant difference between before and after cortical destructions. Nor did short-term memory change as a consequence of frontal-1obe tissue destruction.

If the increase in neuronal elements did bring about an increase in capacity (and this remains a reasonable assumption), we are still incapable of defining the notion capacity. There is no clear indication whether it is related to storage, to simultaneous processing, to internal efficiency of processing, to more advantageous utilization of input, or to speeding up of processing time, etc. We do not know how any of these purely theoretical aspects of capacity are related to particular quantitative dimensions such as brain weight, cell counts, and neurodensity. It is not good enough to say, “there is nothing else that a large brain could be good for than to bring about greater intelligence and language!” This is merely a reflection of our ignorance of the causes of particular evolutionary changes.

Since this point is of no small consequence for our image of man, his capacities, and his place in the primate order, one last consideration on the nature of cognition may be in place. Suppose one explores cognitive capacities of animals and man by systematic measurements of a great number of aspects of psychological activities. Let it include various types of memory, pattern recognition, associative capacities, generalization, and propensity for inference. Each type of measurement constitutes a dimension with which we can construct a multidimensional, mathematical space; let us call it the generalized cognition space. The total of the capacities that characterize a given species, that is, its species-specific cognition, now becomes a locus in the cognition space. Cognitive evolution could be expressed, in such a space, as vectors; the locus of an earlier form is thus connected with that of a later form. The directionality and length of the vector represent the peculiar changes that took place in the course of evolution to bring about the species-specific cognition of a given animal. In terms of this conceptualization, could we expect all evolutionary cognitive changes to be vectors that have the same direction? Such a supposition seems absurd. On the other hand, the enlargement of the brain is a widespread and recurring phenomenon. It appears to be contrary to our empirical findings that an increase in size changes cognitive capacity in a specific direction. Thus, man’s peculiar type of intelligence is not the “logical or necessary” outcome of the enormous growth of his brain, and his capacities today could not have been predicted simply from a knowledge of the evolutionary trend of the change of brain volume.

In short, we do not know why the brain increased so rapidly in size. Since man is distinct from other Hominidae in many ways, we cannot reconstruct which feature added most to selection pressures and which came about through pleiotropic effects (Caspari, 1958). While it is entirely possible that the emergence of language and intelligence are historically related to the increase in size of the brain, the case is certainly not yet irrefutably proved, and the various arguments adduced for one or the other position are too weak to allow us to date the onset of language from fossil remains.

Arguments Based on Other Skeletal Features

Most relevant here are the shape of the jaws and oral cavity, the suspension of the tongue, and the shape and mechanisms of pharynx and larynx. Unfortunately, with the exception of the first two items, these structures are not preserved in the fossils, and the reconstructions are so speculative that they need not be considered here. The mandible, with its absence or presence of a chin and the shape of the denture surely has an influence on the acoustic production of sounds. But all one may deduce from this evidence is that the vocalizations of fossil men did not bear any close acoustic resemblance to the speech sounds of any modern tongue. But we may not dismiss the possibility that the early vocalizations might already have had ethological or biological characteristics that foreshadowed modern languages in some way. None of the language aspects discussed can be reconstructed from the paleontological findings.

Racial Diversification and the Emergence of Language

All races appear to have the same biological potential for the development of culture and the acquisition of language. Thus we must assume that the evolutionary events favoring culture and language go back to the common ancestor of all modern races. This would mean that the age of language is no less than say 30,000 to 50,000 years. Credence is lent to this hypothesis not only on the grounds of racial evidence; the cultures associated with the fossils of this period give evidence of the development of a symbolic medium other than language: graphic representation. The cave drawings of that time are extremely skillful and, what is more important, they are highly stylized and, in a sense, abstract. Thus it is likely that the cognitive processes of Cro-Magnon had a number of characteristics in common with modern man.

The possibility that language is of much older age is not precluded. One authority (Coon, 1962) has advanced the hypothesis that the races have individual ancestries that go back as far as Australopithecus. Mayr (1962) has pointed out that this thesis is far from substantiated but also not entirely impossible. If this were so, language or its prerequisites could have been present as long as half a million years ago. Another theoretical possibility is that the biological matrix for language is of great age but that earliest fossil man did not yet “utilize” it fully. This brings to mind Waddingtons (1957) epigenetic landscape representing the notion of canalization. The evolutionary trend might have entered a particular groove which canalized the subsequent developments and thus made language the necessary outcome, owing to a peculiar evolutionary antecedent. We see that consideration of modern races sets a time at which we might reasonably assume language to have been in existence. But it does not enable us to carry the dating of its emergence any further.

Cultural Status as Evidence for Language

Do the cultural remains of prehistoric man furnish clues for the dating of the development of language. If one could be certain that language is the necessary concomitant of either tool-making, or social organization, or cultural complexity, one could make fairly precise statements about the time of the birth of language. Unfortunately, such certainty does not exist for the first two phenomena, and even the third gives but vague indications.

The use of objects as instrumentalities for behavior is not reserved to man among the primate order (Miyadi, 1964; Goodall, 1963; Birch, 1945; W. Köhler, 1927). Apparently a primitive capacity for the use of tools is common to several primate species and is therefore not necessarily tied to the human form of communication. However, even the earliest forms of Homo must have made a very different and much more extensive use of tools than any subhuman primate today. Miller (1964) has proposed that the use of tools and the use of language demand very similar, biologically given capacities. I consider this to be a fruitful way of looking at language but, at the same time, it must be stressed that it does not compel us to assume simultaneous emergence of the two skills. One may have been present before the other, and there is no way to decide which might have had the lead. Nor does the nature of the tools or the state of primitivity allow us to postulate concomitant levels of primitivity for a form of communication. (Miller suggested no such correlation.)

Degree of social organization must be related to efficiency of intraspecies communication. This is almost a truism. But the communication may, and does, take on an infinite variety of forms. Our speculations about the beginnings of language are seriously handicapped by two unknown factors. First, we can make only the vaguest of assumptions about the social structure and organization of prehistoric men; second, their forms of communication might have been highly developed but very different in nature and principle from our present form. Therefore, Dart’s (1959) postulate that evidence of hunting and fishing technology makes the possession of language a reasonable assumption cannot be accepted uncritically. Perhaps a certain activity calls for good communication, but whether this was a direct, primitive antecedent of what we now call language is uncertain.

The most difficult evaluation of evidence for the existence of language is cultural complexity. At one time our phylogenetic ancestors must have had a truly primitive form of culture. On the other hand, the neolithic cultures of fifty thousand years ago may not have been any less complex than the most primitive living cultures today, say in central Brazil or New Guinea. When did “complexity of culture” arise? Can we be certain that the prehistoric cultures were as primitive as their physical remains today would indicate? The older the culture, the more tenuous must be our guesses. Even the notion of “complexity” is itself a source of inaccuracy. One cannot measure degrees of complexity of culture. There is a further difficulty in using culture as an indication for language.

Today we may study cultures that are essentially neolithic in their state of development, as well as cultures advanced enough to split the atom and explore interplanetary space. Surprisingly, the natural languages spoken throughout this range of cultures appear to be based on similar principles. It is an empirical fact that today neither the tools commonly occurring in a given culture nor the social structure associated with that culture can give us clues about the complexity of structure of the language now spoken by the individuals of that culture. Natural languages cannot be ordered in terms of complexity. A complex task should be more difficult to learn than a simple one and therefore take more time and effort. But all natural languages are learned with the same ease by children of a certain age, which seems to confirm the “equal complexity” hypothesis.

It is reasonable to assume (though not absolutely necessary) that Cro-Magnon man, whose material culture might not have been too far removed from the most primitive present-day cultures, and who had all the physical characteristics of modern man, was in possession of language as we know it today. There is nothing that requires us to think of his language as substantially more primitive than ours, or to postulate any “uk-uk-theory.”* It is likely that Cro-Magnon and Neanderthal men were speaking creatures. We have no means of deciding whether earlier races had a form of communication that was in any way similar to that of Cro-Magnon. Attempts at dating the origin of modern types of language development seem unwarranted.

SUMMARY

The biological history of language is “covert”; its evolution is hidden in the series of transformations, structural and functional, that took place in the course of the formation of modern man. It is tied to the history of physiological adaptations, of cognitive specializations, of sensory specifications.

Our present capacity for language may well go back to species-specific alterations in genetic material (intracellular changes) which, however, affected certain rates and directions of growth during ontogeny, producing a peculiar ontogenetic phase of an optimal confluence of various abilities; the capacity for language might thus be a secondary consequence of a new synchronization of maturation rates of diverse structures and their functions, an assumption which must not be confused with the postulation of “genes for language.”

The biological history of language cannot be revealed through a random comparison with animal communication; this is particularly so if the basis of comparison is pragmatic or “logical” and without regard to the animals’ phylogenetic relation to man. Comparison of language with animal communication beyond the order of primates is dangerous because of the phenomenon of convergence.

Reconstruction of the origin of language is impossible except for some very simple determinations. This is due to the following limitations: (1) the size and shape of the brain furnish no secure clue about the capacity for language; (2) given morphological peculiarities of the central nervous system do not bear a fixed relationship to behavior; the same cerebral feature may subserve somewhat different aspects of behavior in different species, and vice versa; the relation of behavior to certain aspects of the brain may have undergone several changes during the course of evolution of modern man; (3) even if we had direct knowledge of social structure or cultural complexity of the societies of various fossil men, we could not draw conclusions about language as we know it today. Different types of communication might have prevailed at those times.

The identical capacity for language among all races suggests that this phenomenon must have existed before racial diversification. There is nothing unbiological about recognizing language as unique behavior in the animal kingdom; such uniqueness is to be expected from the very processes underlying speciation.

* For a more complete treatment of the ideas presented here see my Biological Foundations of Language, New York, Wiley, 1967; the present article is excerpted (with permission of the publisher) from Chapter VI.

* It is true that human ontogeny need not be a recapitulation of the evolutionary events that led up to the formation of language capacity. On the other hand, the first stages of language acquisition in the child are the only types of language that we may confidently label as primitive beginnings. We have no other empirical data from which we could infer language primitivity in a phylogenetic sense.

* By “uk-uk-theory” I mean all of those accounts in which the beginning of language is characterized as the discovery that the original animal evocations (such as a supposed “uk-uk”) could be used for transmitting information to other individuals.

REFERENCES

Altmann, S. A., (in press). The structure of primate social communication.

Birch, H. G., 1945. The relation of previous experience to insightful problem-solving. J. Comp. Psychol., 38:367-383.

Bok, S. T., 1959. Histonomy of the Cerebral Cortex. Elsevier, Amsterdam.

Bonin, G. v., 1950. Essay on the Cerebral Cortex. Springfield, Illinois, Thomas.

Bonin, G. v., 1963. The Evolution of the Human Brain. University of Chicago Press, Chicago.

Bonin, G. v., and P. Bailey, 1961. Pattern of the cerebral isocortex. In: H. Hofer, A. H. Schultz, and D. Stark, (eds.) Primatologia: Handbook of primatology, Vol. II, part 2, fasc. 10. Basel, Karger.

Caspari, E., 1958. Genetic basis of behavior. In: A. Roe and G. G. Simpson (eds.), Behavior and Evolution, New Haven, Yale Univ. Press.

Connolly, C. J., 1950. External Morphology of the Primate Brain. Springfield, Illinois, Thomas.

Coon, C., 1962. The Origin of Races, New York, Knopf.

Darlington, C. D., 1947. The genetic component of language. Heredity, 1:269-286.

Dart, R. A., 1956. The relationships of brain size and brain pattern to human status. S. Afr. J. Med. Sci., 21:23-45.

Dart, R. A., 1959. On the evolution of language and articulate speech. Homo, 10:154-165.

Dobzhansky, T., 1962. Mankind Evolving. New Haven, Yale University Press.

Feremutsch, K., 1961. Basalganglien. In: H. Hofer, A. H. Schultz, and D. Starck (eds.), Primatologia: Handbook of primatology, Vol. II, part 2, fasc. 8. Basel, Karger.

Feremutsch, K., 1963. Thalamus. In: H. Hofer, A. H. Schultz, and D. Starck (eds), Primatologia: Handbook of primatology, Vol. II, part 2, fasc. 6. Basel, Karger.

Garstang, W., 1922. The theory of recapitulation: a critical restatement of the biogenetic law. J. Linn. Soc. London, 35:81-101.

Ghent, L., M. Mishkin, and H.-L. Teuber, 1962. Short-term memory after frontal-1obe injury in man. J. comp. Physiol. Psychol., 55:705-709.

Goldschmidt, R., 1938. Physiological Genetics. New York, McGraw-Hill.

Goldschmidt, R., 1952. Evolution as viewed by one geneticist, Amer. Scientist, 40:84-135.

Goodall, J., 1963. My life among wild chimpanzees. Natl Geogr., 124(2) :278- 308.

Goodman, M., 1963. Man’s place in the phylogeny of the primates as reflected in serum proteins. In: S. L. Washburn (ed), Classification and Human Evolution. Aldine Publ., Chicago.

Haldane, J. B. S., 1949. Suggestions as to quantitative measurement of rates of evolution. Evolution, 3:51-56.

Harlow, H. F., 1949. The formation of learning sets. Psychol Rev., 56:51-65.

Harlow, H. F., 1958. The evolution of learning. In: A. Roe and G. G. Simpson, Behavior and Evolution. New Haven, Yale University Press.

Hockett, C. F., 1960. Logical considerations in the study of animal communication. In: W. E. Lanyon and W. N. Tavolga (eds.), Animal Communication. Washington, D.C., Am. Inst, of Biol. Sci., pp. 392-430.

Hockett, C. F., 1960. The origin of speech. Scientific American, Sept., 203:89-96.

Hockett, C. F., and R. Ascher, 1964. The human revolution. Current Anthrop., 5:135-168.

Keith, A., 1948. A New Theory of Human Evolution. London, Watts.

Koehler, O., 1951. Der Vogelgesang als Vorstufe von Musik und Sprache, J. f. Ornithol, 93:1-20.

Koehler, O., 1952. Vom unbenannten Denken, Verh. Dtsch. Zool. Gesell., in Freiburg, 202-211.

Koehler, O., 1954a. Vom Erbgut der Sprache, Homo, 5:97-104.

Koehler, O., 1954b. Vorbedingungen und Vorstufen unserer Sprache bei Tieren, Verhandlungen Dtsch. Zool. Gesell, in Tübingen, 327-341.

Köhler, W., 1927. The Mentality of Apes. London.

Marler, P., 1961. The logical analysis of animal communication. J. Theoret. Biol, 1:295-317.

Mayr, E., 1958. Behavior and Systematics, in A. Roe, and G. G. Simpson, Behavior and Evolution, Yale Univ. Press., New Haven.

Mayr, E., 1962. Origin of the human races (book review of Coon’s Origin of Races), Science, 138: (19 Oct.) 420-422.

Mayr, E., 1963. Animal Species and Evolution. Belknap Press of Harvard University, Cambridge, Mass.

Miyadi, D., 1964. Social life of Japanese monkeys. Science, 143:783-786.

Overhage, P., 1959. Das Problem der Anthropogenese, in A. Haas, Das stammesgeschichtliche Werden der Organismen und des Menschen, Vol. 1, Herder, Basel.

Rensch, B., 1954. Neuere Probleme der Abstammungslehre, F. Enke, Stuttgart.

Rensch, B., 1959. Trends towards progress of brains and sense organs, Cold. Spr. Harb. Symp. quant. Biol., 24:291-303.

Rensch, B., 1964. Memory and concepts of higher animals, Proc. Zool. Soc. Calcutta, 17:207-221.

Roe, A., and G. S. Simpson, 1958. Behavior and Evolution, New Haven, Yale Univ. Press.

Schrier, A. M., H. F. Harlow, and F. Stollnitz, (eds.), 1965. Behavior of Nonhuman Primates. New York, Academic Press.

Schultz, A. H., 1956. Postembryonic age changes, in H. Hofer, A. H. Schultz, and D. Starck (eds.), Primatologia: Handbook of Primatology, Vol. I, Basel, Karger.

Schultz, A. H., 1962. Die Schädelkapazität männlicher Gorillas und ihr Höchstwert., Anthrop. Anz., 25:197-203.

Simpson, G. G., 1949. The Meaning of Evolution, New Haven, Yale University Press.

Smith, D. D., 1965. Mammalian Learning and Behavior, Philadelphia, Saunders.

Starck, D., 1965. Die Neencephalisation. In: G. Heberer (ed.), Menschliche Abstammungslehre. Stuttgart, Fischer.

Teuber, H.-L., 1959. Some alterations in behavior after cerebral lesions in man. Evolution of Nervous Control, AAAS, Washington, D.C.

Uexküll, J. v., 1921. Umwelt und Innenwelt der Tiere, Berlin.

Waddington, C. H., 1957. The Strategy of Genes: a discussion of some aspects of theoretical biology, London, George Allen & Unwin.

Woodger, J. H., 1945. On biological transformations, in W. E. Le Gros Clark, P. B. Medawar (eds.), Essays on Growth and Form presented to D’Arcy W. Thompson, Oxford, Clarendon.

Wrete, M., 1962. Autonomes Nervensystem, in H. Hofer, A. H. Schultz, and D. Starck (eds.), Primatologia: Handbook of primatology, Vol. II, part 2, fasc. 11, Basel, Karger.