Animal Communication and Human Language
Stephen R. Anderson
[Prepared for Cambridge
Encyclopedia of the Linguistic Sciences]
An understanding of the communicative capacities of other animals is important on its face both for an appreciation of the place of human language in a broader context, and also as a prerequisite to discussion of the evolution of language. On closer examination, however, the differences between human language and the systems of other animals appear so profound as to make both projects more problematic than they appear at first.
In the 1950s and 1960s, ethologists like Konrad Lorenz and Niko Tinbergen revolutionized behavioral biologists’ views of the cognitive capacities of animals, but consideration of animal communication focused on the properties of quite simple systems. A prime example of “communication” in the texts of the time was the stickleback. A crucial component of the mating behavior of this common fish is the pronounced red coloration of the male’s under-belly when he is in mating condition, which furnishes a signal to the female that she should follow him to his pre-constructed nest, where her eggs will be fertilized. On this model, communication was viewed as behavioral or other signals emitted by one organism, from which another organism (typically, though not always, a conspecific) derives some information. The biological analysis of communication thus came to be the study of the ways in which such simple signals arise in the behavioral repertoire of animals and come to play the roles they do for others who perceive them. Those discussions make little if any contact with the analysis of human language.
In the intervening half century, we have come to know vastly more about the nature and architecture of the human language faculty, and to have good reason to think that much of it is grounded in human biology. One might expect, therefore, to find these concerns reflected in the behavioral biology literature. A comprehensive modern textbook on animal communication within this field (e.g. Bradbury & Vehrenkamp 1998) reveals a great deal about the communicative behavior of many species and its origins, but within essentially the same picture of what constitutes “communication” in (non-human) animals, confined to unitary signals holistically transmitted and interpreted. Little if any of what we have come to know about human linguistic communication finds a place here.
Biologists have not in general paid much attention to the specifics of linguistic research (though their attention has been caught by the notion that human language is importantly based in human biology), and are often not as sophisticated as one might wish about the complexity of natural language. But the consequences of this may not be as serious as linguists are inclined to think: in fact, the communicative behavior of non-humans in general is essentially encompassed within the simple signal-passing model. The complexities of structure displayed by human language are apparently quite unique to our species, and may not be directly relevant to the analysis of animal communication elsewhere.
Communication in the sense of emission and reception of informative signals is found in animals as simple as bacteria (“quorum sensing”). Most familiar, perhaps, are visual displays of various sorts indicating aggression, submission, invitations to mate, etc. In some instances these may involve quite complex sequences of gestures, reciprocal interactions, and the like, as in the case of the nesting and mating behavior of many birds. In others, a simple facial expression, posture, or manner of walking may provide the signal from which others can derive information about the animal’s intentions and attitudes.
These differences of internal structure are of course crucial for the correct expression and interpretation of a particular signal, but they play little or no role in determining its meaning. That is, the individual components of the signal do not in themselves correspond to parts of its meaning, in the sense that varying one sub-part results in a corresponding variation in what is signaled. Animal signals, however complex in form (and however elaborate the message conveyed), are unitary wholes. An entire courtship dance, perhaps extending over several minutes or even longer, conveys the sense “I am interested in mating with you, providing a nesting place and care for our offspring.” No part of the dance corresponds exactly to the “providing care” part of the message; the message cannot be minimally altered to convey “I am interested in mating, but not in providing care for our offspring,” “I was interested in mating (but am no longer)...” etc. Variations in intensity of expression can convey (continuous) variations in the intensity of the message (e.g., urgency of aggressive intent), but that is essentially the only way messages can be modulated.
The most widely discussed apparent exception to this generalization is the “dance language” of some species of honeybees. The bees’ dance conveys information about (a) the direction, (b) the distance, and (c) the quality of a food source (or potential hive site), all on quasi-continuous scales and each in terms of a distinct dimension of the dance. Although the content of the message here can be decomposed, and each part associated with a distinct component of the form of the signal, there is no element of free combination. Every dance necessarily conveys exactly these three things, and it is only the relative value on each dimension that is variable. As such, the degree of freedom available to construct new messages is not interestingly different from that involved in conveying different degrees of fear or aggression by varying degrees of piloerection.
Visual displays do not at all exhaust the modalities in which animal communication takes place, of course. Auditory signals are important to many species, including such classics of the animal communication literature as frog croaks and the calls and songs of birds. In some species, portions of the auditory spectrum that are inaccessible to humans are involved, as in the ultrasound communication of bats, some rodents, and dolphins, and the infrasound signals of elephants. Chemical or olfactory communication is central to the lives of many animals, including moths, mice and lemurs as well our pet cats and dogs. More exotic possibilities include the modulation of electric fields generated (and perceived) by certain species of fish.
In some of these systems the internal structure of the signal may be quite complex, as in the songs of many oscine songbirds, but the general point made above still holds: however elaborate its internal form, the signal has a unitary and holistic relation to the message it conveys. In no case is it possible to construct novel messages freely by substitutions or other ways of varying aspects of the signal’s form.
In most animals, the relation of communicative behavior to the basic biology of the species is very direct. Perceptual systems are often quite precisely attuned to signals produced by conspecifics. Thus, the frog’s auditory system involves two separate structures (the amphibian papilla and the basilar papilla) that are sensitive to acoustic signals, typically at distinct frequencies. The frequencies to which they are most sensitive vary across species, but are generally closely related to two regions of prominence in the acoustic structure of that species’ calls. Mice (and many other mammals) have two distinct olfactory organs, projecting to quite distinct parts of the mouse brain. The olfactory epithelium is responsive to a wide array of smells, but the vomeronasal organ is sensitive exclusively to the pheremones that play a major role in communication and social organization. In this case, as in many, many others, the perceptual system is matched to production in ways that optimize the organism’s sensitivity to signals that play a crucial ecological role in the life of the animal.
The essential connection between a species’ system of communication and its biology is also manifested in the fact that nearly all such systems are innately specified. That is, the ability to produce and interpret relevant signals emerges in the individual without any necessary role of experience. Animal communication is not learned (or taught), but rather develops (in the absence of specific pathology, such as deafness) as part of the normal course of maturation. Animals raised under conditions in which they are deprived of exposure to normal conspecific behavior will nonetheless communicate in the fashion normal to their species when given a chance.
Exceptions to this generalization are extremely rare, apart from human language. Vocal learning, in particular, has been demonstrated only to a limited extent in cetaceans and some bats, and more extensively in three of the twenty seven orders of birds. The study of birds, especially oscine songbirds, is particularly instructive in this regard. In general, their song is learned on the basis of early exposure to appropriate models, from which they in turn compose their own songs. There is much variation across species, but a clear generalization emerges: for each species, there is a specific range of song structures that individuals can learn. Experience plays a role in providing the models on which adult song is based, but (with the exception of a few very general mimics, such as the lyrebird) this role is quite narrowly constrained by the song-learning system of the individual species.
Like the systems of communication of other animals, human language is deeply embedded in human biology. Unlike others, however, it provides an unbounded range of distinct, discrete messages. Human language is acquired at a specific point in development from within a limited range of possibilities, similar to the acquisition of song in birds. Unlike the communicative signals of other species, human language is under voluntary control, with its underlying neurobiology concentrated in cortical structures, as opposed to the sub-cortical control characteristic of those other species that have been studied in this regard.
Human language is structurally a discrete combinatorial system, in which elements from a limited set combine in a recursive, hierarchical fashion to make an unlimited number of potentially novel messages. The combinatorial structure of language is governed by two quite independent systems: a small inventory of individually meaningless sounds combine to make meaningful words, on the one hand (PHONOLOGY), while these words are combined by a quite different system to make phrases, clauses, and sentences (SYNTAX). These properties (discrete combination, recursive hierarchical organization, and “duality of patterning”) are not simply idiosyncratic ornaments that could in principle be omitted without affecting the overall communicative capacity of the system. Rather, they are what makes large vocabularies practical and unbounded free expression possible. Contrast the unlimited range of potentially novel utterances which any (normal) speaker of a language can produce, and another speaker of the same language comprehend, with the strictly limited range of meaningful signals available to other organisms. No other form of communication found in nature has these properties. Human language, and especially its syntactic organization is quite unique in the animal world.
Furthermore, efforts to teach systems with these essential properties to other animals have not succeeded. Despite widespread claims to the contrary in the popular literature, there is no evidence that any non-human animal is capable of acquiring and using such a system. This should not be seen as particularly surprising: if language is indeed embedded in human biology, there is no reason to expect it to be accessible to organisms with a different biological endowment, any more than humans are capable of acquiring, say, the echolocation capacities of bats, a system which is equally grounded in the specific biology of those animals.
Human language is often considered as simply one more instantiation of the general class of animal communication systems. Indeed, like others it appears to be highly species-specific: although relevant experience is required to develop the system of any particular language, the overall class of languages accessible to the human learner is apparently highly constrained, and the process of language learning more like genetically governed maturation than like learning in general. The structural characteristics of human language are, however, quite different from those of other communication systems, and it is the freedom of expression subserved by those distinctive properties that gives language the role it has in human life.
— Stephen R. Anderson
Works Cited and Suggestions for Further Reading
Anderson, Stephen R. 2004. Doctor Dolittle’s Delusion: Animals and the Uniqueness of Human Language. New Haven: Yale University Press.
Bradbury, J. W. & Sandra Vehrenkamp. 1998. Principles of Animal Communication. Sunderland, MA: Sinauer.
Hauser, Marc D. 1996. The Evolution of Communication. Cambridge, MA: MIT Press.