Susanne K. Langer "'The Origins of Speech and Its Communicative Function"

From The Quarterly Journal of Speech, Vol. XLVI, No. 2 (April 1960), 121-134.

“Susanne K. Langer is Professor of Philosophy, Connecticut College. She is author of a number of books, including Feeling and Form, Problems of Art, and Philosophy in a New Key. This paper was read at the University of Pittsburgh, under the auspices of the Department of Speech. It presents a detail of a larger study in the Philosophy of Mind, now in progress, supported by a research grant given to Connecticut College by the Edgar J. Kaufmann Charitable Trust [which study became her magnum opus, Mind: An Essay on Human Feeling.—A.F.] Professor Langer’s contribution is particularly timely in the centennial year of Charles Darwin’s Origin of Species, 1959-60.”

Posted June 20, 2008

The Origins of Speech and Its Communicative Function

Ever since the Darwinian theory of human evolution—tracing the descent of man from animal ancestors—has become generally accepted, the origin of speech has become more and more mystifying. Language is so much the mark of man that it was classically supposed to have been bestowed on him at his creation. But if he has not been created separately from the animals, but has arisen, as most of us now believe, just like them, from a more primitive animal ancestry, then surely at some time his own precursors did not speak. When, why and how did man begin to speak? What generations invented that great social instrument, language? What development of animal communica-tion has eventuated in human communication? What pre-Adamite thought of assigning a particular little squeak to a particular object as the name of that object, by which you could refer to it, demand it, make other people think of it? How did the other pre—Adamites all agree to assign the same squeaks to the same things? What has led to the concatenation of those primitive words in syntactically structured sentences of interrelated meanings? As far as anthropologists know, there is no human language that is not discursive—propositional—in form. Its propositions may be very different from ours, but their semantical structure is always equivalent to what we call a statement. Language always expresses relations among acts or things, or their aspects. It always makes reference to reality—that is, makes assertions or denials—either explicitly or implicitly. Some nouns imply relations, and where they do, verbs may not be needed. In classical Latin the verb is often understood through the inflections of nouns and adjectives. Verbs, in some languages, may imply their subject or object or even both, and make nouns all but unnecessary, as Whorf found in Hopi.¹

But no language consists of signs that only call attention to things without saying anything about them—that is, without asserting or denying something. All languages we know have a fairly stabile vocabulary, and a grammatical structure. No language is essentially exclamatory (like ah! and oh!), or emotional (like whining and yodeling), or even imperative.² The normal mode of communicative speech, in every human society, is the indicative; and there is no empirical evidence, such as a correlation of increasing discursiveness with increasing culture, to support the belief that it was ever otherwise.

Language may be used to announce one’s presence, to greet people, to warn, to threaten, to express pain or joy, or even for directing action.³ Whenever people speak of “animal language” they refer to such uses of observable signs among animals.⁴ Leaving aside, for the moment, the alleged “language” of social insects,”⁵ we may use the term vocal signs among animals.

Now, it is an obvious commonsense assumption that human language has grown from some such lower form of vocal communication. But common sense is a very tricky instrument; it is as deceptive as it is indispensable. Because we use it, and have to use it, all the time, we tend to trust it beyond its real credentials, and to feel disconcerted if its simple interpretations of experience fail. Yet commonsense conceptions of the nature and origin of human speech have always led into dilemmas, and will until the problem of its beginning and development has been generally given up.

Even methodology develops its commonsense principles. One of these is that, if you would find the important relationships between two phenomena, you should begin by checking what the phenomena have in common. So, in comparing the vocal communications of animals and men respectively, we find that all the things animals communicate by sound may also be communicated by human language; and it seems reasonable enough that those things which human language can do and animal vocalization cannot, have been added to the primitive animal language, to make the greatly elaborated system of verbal intercourse.⁶ But the finding of these common elements leads no further. Commonsense methodology, like the commonsense assumptions, produces nothing more than what we already knew—by common sense.

So it may be in order to question our obvious premises, and even depart from the method of seeking common factors in animal and human communication. Instead of noting points of similarity, let us consider the cardinal difference between human and animal language. That difference is in the uses to which utterances are put. All those functions that animal and human utterances share—calling, warning, threatening, expressing emotion—are essential uses of animal sounds, and incidental uses of human speech. The functions of animal vocalization are self-expression and sometimes, perhaps, indication of environmental conditions (like the bark of a dog who wants to be let in). The chief function of speech is denotation.

Animal language is not language at all; and, what is more important, it never leads to language.⁷ Dogs that live with men learn to understand many verbal signals, but only as signals, in relation to their own actions. Apes, that live in droves and seem to communicate fairly well, never converse.⁸ But a baby that has only half a dozen words begins to converse: “Daddy gone.” “Daddy come? Daddy come.” Question and answer, assertion and denial, denotation and description—these are the basic uses of language.

The line between animal and human estate is, I think, the language-line; and the gap it marks between those two kinds of life is almost as profound as the gap between plants and animals. This makes it plausible that we are not dealing with just a higher form of some general animal function, but with a new function developed in the hominid brain—a function of such complexity that probably not one, but many subhuman mental activities underlie it.

The complexity of living forms and functions is something that we are apt to underestimate in speculating on the origins of psychological phenomena. In textbook accounts the facts have to be generalized and simplified to make them comprehensible to beginners; but as soon as you tackle the monographic literature presenting actual cases of growth, maturation, and the conduct of life, and follow actual analyses of function and structure, especially in neurology, the complexity and variability of vital processes is brought home to you with great force. Consider only the chemical activities, that differ enough from anyone organism to another to produce the so-called “individuality factor.”⁹ Or think of the structural organization of the brain; in the small brain center known as the “lateral geniculate body” where the optic nerve ceases to be one bundle of fibres and fans out toward the cortex of the occipital lobe, anatomists have found scores of so-called “boutons,” points of reception or emission of electrical impulses, directly on nerve-cells, besides the synaptic connections of the branching axons and dendrites of those same cells.¹⁰ The potentialities of such a brain for different courses of activity run into billions and trillions, so that even if inhibiting mechanisms eliminate a hundred thousand connections at a time, the range of possible responses, especially in the crowded circuits of the forebrain, are as good as infinite.

It is very wholesome for a philosopher who tries to conceive of what we call “mind” to take a long look at neurological exhibits, because in psychological studies we usually see and consider only the integrated products—actions and intentions and thoughts—and with regard to speech, words and their uses. Words seem to be the elements of speech; they are the units that keep their essential identity in different relational patterns, and can be separately moved around. They keep their “roots” despite grammatical variations, despite prefixes and suffixes and other modifications. A word is the ultimate semantic element of speech. A large class of our words—most of the nouns, or names—denote objects, and objects are units that can enter into many different situations while keeping their identity, much as words can occur in different statements. This relation gives great support to the conception of words as the units of speech.

And so, I think, they are. But this does not mean that they are original elements of speech, primitive units that were progressively combined into propositions. Communication, among people who inherit language, begins with the word—the baby’s or foreigner’s unelaborated key word, that stands proxy for a true sentence; but that word has a phylogenetic history, the rise of language, in which probably neither it nor any archaic version of it was an element.

I think it likely that words have actually emerged through progressive simplification of a much more elaborate earlier kind of utterance, which stemmed, in its turn, from several quite diverse sources; and that none of its major sources were forms of animal communication, though some of them were communal.

These are odd-sounding propositions, and I am quite aware of their oddness, but perhaps they are not as fantastic as they sound. They merely depart rather abruptly from our usual background assumptions. For instance, the idea that a relatively simple part of a complex phenomenon might not be one of its primitive factors, but might be a product of progressive simplification, goes against our methodological canons; ever since Thomas Hobbes set up the so-called genetic method of understanding, we have believed that the simplest concepts into which we could break down our ideas of a complex phenomenon denoted the actual elements of that phenomenon, the factors out of which it was historically compounded. Locke’s construction of human experience from pure and simple sense data, Condillac’s fancied statue endowed with one form of perception after another, and in our own time Bertrand Russell’s “Logical atomism,” all rest on this belief.¹¹ But close empirical study of vital processes in nature does not bear it out. A great many advanced behavior patterns are elaborations of simpler responses, but some are simplifications of very complicated earlier forms of action. The same holds true of the structures that implement them. When the reflex arc was discovered, physiologists felt themselves in possession of a key to all animal response, for here was a simple unit that could be supposed to engender all higher forms by progressive elaboration. But Herrick and Coghill, through careful studies of salamanders in their larval stages,¹² found that the reflex arc is not a primitive structure ontogenetically at all, but is preceded by much more elaborate arrangements in the embryo, that undergo simplifications until a unified afferent-efferent circuit results. This finding was corroborated by Lorente de Nó.¹³

A principle that is operative in the development of an individual is at least possible in the larger development of a stock. There is nothing absurd about the hypothesis that the simple units in a very advanced function, such as human speech, may be simplifications within an earlier more intricate vocal pattern.

Most theories of the origin of language presuppose that man was already man, with social intentions, when he began to speak.¹⁴ But in fact, man must have been an animal—a high primate with a tendency to live in droves like most of the great apes—when he began to speak. And it must have been rather different from the ancient progenitors of our apes, which evidently lacked, or at least never possessed in combination, those traits that have eventuated in speech.

What were those traits? Speech is such a complex function that it has probably not arisen from any single source. Yet if it developed naturally in the hominid stock, every one of its constituents must have started from some spontaneous animal activity, not been invented for a purpose; for only human beings invent instruments for a purpose preconceived. Before speech there is no conception; there is only perception, and a characteristic repertoire of actions, and a readiness to act according to the enticements of the perceived world. In speech as we know it, however, there seems to be one flowing, articulate symbolic act in which conventional signs are strung together in conventional ways without much trouble, and similar processes evoked in other persons, all as nicely timed as a rally of pingpong. Nothing seems more integral and self-contained than the outpouring of language in conversation. How is one ever to break it down into primitive acts?

It was from the psychiatric literature on language—on aphasia, paraphasia, agrammatism, alexia, and kindred subjects—that something like a guiding principle emerged. The most baffling thing about the cerebral disturbances of speech is, what strange losses people can sustain: loss of grammatical form without any loss or confusion of words, so the patient can speak only in “telegraph style,” or contrariwise, loss or confusion of words without loss of sentence structure, so speech flows in easy sentence-like utterances, but only the prepositions, connectives, and vocal punctuations are recognizable; the informative words are all garbled or senselessly.¹⁵ Lewis Carroll’s

illustrates this separation of sentence form and verbal content. There may be inability to understand spoken language, but not inability to understand printed or written language,¹⁶ yet without any defect of hearing; or the other way about—inability to read, but not to understand speech—without any ocular trouble.¹⁷There are cases of alexia for words but not for letters,¹⁸ and the recognition, naming and using of numbers is often intact where neither letters nor words can be recognized.¹⁹ Furthermore, some brain injuries leave the victim able to repeat words spoken for him, but not to speak spontaneously, and others make him unable to repeat words just heard, but not unable to utter them in spontaneous speech. There are even several cases on record of persons in whom a cerebral lesion caused inability to name any inanimate object, but not inability to name living things, and call people by their proper names, and, conversely, cases of inability to name persons, animals, or any parts of them, but not to find the words for inanimate objects like watches and slipppers.²⁰

In the face of these peculiar, sometimes really bizarre exhibits, it occurred to me that what can be separately lost from the integral phenomenon of speech may have been separately developed in the prehistoric, prehuman brain. Here is at least a working notion of a new way to break down the verbal process, that might yield a new conception of what has gone into it.

In singling out such elements, and trying to trace them back to some plausible—though of course hypothetical—prehuman proclivities, one meets with the surprising fact that some of these habits, that may be supposed to have prepared speech, actually exist in the animal kingdom, and are even quite highly developed, sometimes in relatively low animals. But they are far from any kind of speech. They are raw, unassembled materials, that would be needed in conjunction, as a foundation, if speech were to arise. In the pre-human primate they must have coincided at some time to provide that foundation.

This principle of analysis takes us much further back into preparatory phases of mental development than the usual anthropological approach to the problem of speech, which reaches back only to the supposed archaic forms of genuine language. Not only mental activities, but some grosser somatic conditions that made them possible, must have met in the animal stock that produced the human race. For instance, the continuity of language requires a bodily mechanism that can sustain a long process of vocalization. Not all animals can do that; it is interesting that the chimpanzee, which is nearest to man in mental capacity, cannot sustain a vowel sound; also it rarely produces a pure and simple sound. Its larynx is too complicated, and it has more than one source of air supply for it, and no fine control of a single set of bellows to mete out its vocal power.²¹ The gibbon has a simpler larynx, more like ours, and also the requisite propensity to utter long, chant-like ululations in chorus; that is, it has the physical powers of vocalization, and the habit of using them in a gathered company—two prerequisites for speech.²² But its brain is too inferior to endow its joyful noise with anything but self-expression and mutual stimulation to keep it up.

Another condition of speech is the epieritical ear, that distinguishes one sound from another, beyond the usual distinction of noises according to their sources—that is, beyond distinguishing them as calls of other creatures, as footsteps, perhaps as the splash of water, and for the rest either as meaningless rumbles and creaks, or not at all. The epicritical power of hearing requires a highly specialized cochlea and a distribution of the auditory nerve in the brain that is not found in all the higher animals, but occurs in several birds—an anomalous development in a relatively low type of brain. Those birds that imitate the whistles of other birds and the sounds of human speech, whereby we know they have a highly analytic hearing (which anatomical findings bear out),²³ have something more that is relevant to our own powers: the control of the vocal apparatus by the ear, which seems to be rudimentary in most animals, although the mechanisms of hearing and sound-making are always associated—even in the cricket, that has its peripheral organs of hearing in the thighs.²⁴The kind of feedback that molds an utterance according to sounds heard, and makes formal imitation possible, is another specialization beyond the epicritical receptor organ. Dogs have the fine receptor, the ear that discriminates articulate sounds within a general category, for they can respond selectively to quite a gamut of verbal signals, and Pavlov found their discrimination of tonal pitch superior to man’s; but dogs never show the slightest impulse or ability to imitate foreign sounds.

So we find several prerequisites for speech—sustained and variable vocalization, the tendency to responsive utterance, the epicritical hearing and fine control of vocalization by the ear that implement imitation—prefigured in the behavior patterns of widely different animals. Yet none of those animals use language. These traits are only some of its conditions, and even they do not coincide in anyone species. In the proto-human primate they must have coincided—not only with each other, but with some further ones as well, that may or may not occur in other creatures.

The decisive function in the making of language comes, I think, from quite another quarter than the vocal-auditory complexes that serve its normal expression. That other quarter is the visual system, in which the visual image—the paradigm of what, therefore, we call “imagination”—almost certainly is produced.

How a visual image is engendered and what nervous mechanisms participate in its creation no one has yet described; I have gathered a few ideas on the subject, but they need not detain us here. The important thing is that images are the things that naturally take on the character of symbols. Images are “such stuff as dreams are made on”; dreams have the tendency to assume symbolic value, apparently very early in our lives; and the peculiar involutions of meaning in their imagery, the vagueness of connections, the spontaneity of their presentations and the emotional excitement of any very vivid dream, may well reflect the nature of primitive symbolic experience.

The old problem, how words became attached to objects as their distinctive names, and how they became generalized so they denoted kinds of things rather than individuals, may find its solution if we can give up the notion that primitive man invented speech, and agreed on names for things and other basic conventions. I do not believe names were originally assigned to things at all; naming is a process that presupposes speech. Now that we have language, we can give names to new comets, new gadgets, and constantly to new babies. But in the making of speech, I think it more likely that definite phonetic structures were already at hand, developed in another context, and that meanings accrued to them—vaguely and variably at first, but by natural processes that tended to specify and fix them. Such meanings were not pragmatic signal values of specific sounds for specific things; several eminent psychiatrists to the contrary notwithstanding,²⁵ primitive denotation was not like using a proper name. When words took shape, they were general in intent, from the beginning; their connotations inhered in them, and their denotations were whatever fitted this inherent sense.

Now that I have thus pontificated on what happened, let me explain why I think something like this must have happened, and how it would account for the greatest of all mysteries of language—the fact that language is symbolic, when no animal utterance shows any tendency that way. The biological factors that caused this great shift in the vocal function were, I believe, the development of visual imagery in the humanoid brain, and the part it came to play in a highly exciting, elating experience, the festal dance. (How pre-human beings advanced from animal behavior to formalized tribal dance is another relevant subject I cannot broach here.) The mental image was, I think, the catalyst that precipitated the conceptual import of speech.

As I remarked before, images are more prone than anything else we know to become symbols; they have several attributes that work together to make them symbolic. So it was another of the evolutionary coincidences that the Calibans who preceded us suffered a peculiar specialization in their visual systems, so that we produce mental images without even trying—most successfully, in fact, while we sleep.

There is a reason, of course, why this should be a hominid specialty, and we can at least guess what caused our odd and rather impractical habit of visualizing, with and without stimulation from the end-organs, the eyes. The human brain presumably developed, like any animal brain we know, as a mediating organ between afferent impulses and their efferent completion, that is, their spending themselves in action. In animals, typically, every stimulation that takes effect at all is spent in some overt act, which may be anything from are flex twitch of the skin to a directed act of the whole aroused creature. But the messages which come into our brains are so many and various that it would be impossible and exhausting to spend each afferent impulse in overt action. So a great many, especially the countless visual impressions we take in, have to be finished within the brain; the cerebral response is the formation of an image. This automatic process may occur in animals, too, but sporadically and at a lower intensity, and therefore without further consequences. If animals have images, I don’t think they are bothered by them or use them; such passing visions may be like our after-images, automatic products of sensory stimulation.²⁶

In human beings, however, image making has become a normal conclusion for acts of focussed gazing. Since, in the waking state, it is easier to look at things than not to, image-production is generally effortless and unintentional, and in the normal course of development soon becomes so rich that there is a constant play of imagery. Every impression is apt to produce an image, however briefly and incompletely, and out of this welter a few more definite visualizations emerge at intervals.

The several characteristics that make the mental image prone to become symbolic are, in the first place, this spontaneous, quasi-automatic production; secondly, a tendency of image-making processes to mesh, and pool their results; then, their origin in actual perception, which gives images an obvious relation to the sources of perception—things perceived—a relation we call “representation”; furthermore, the very important fact that an image, once formed, can be reactivated in many ways, by all sorts of external and internal stimulations; and finally, its involvement with emotion. Let us consider what each of these traits has to do with the making of the primitive symbol, and with the enlistment of the vocal organs for its projection.

A biological mechanism that is about to assume a new function is usually developed at least somewhat beyond the needs of its original function; that is, its activity has a certain amount of play, sometimes called “excess energy,” which allows unpredictable developments. A new departure is not likely to be based on rare occurrences, for to become established it has to survive many miscarriages, and that means that it has to begin over and over again—that is, the conditions for it have to be generous. So, in a brain where imagination was to take on a new and momentous function—symbolization—the pro-duction of images had to be a vigorous business, generating images all the time, so that most of them could be wasted, and the symbolic activity could still begin again and again, and proceed to various degrees, without interfering with the normal functions of the brain in the whole organic economy. So the normality and ease of image producing met one of the first requisites for the rise of a higher function.²⁷

The second important feature of mental images for symbol making is the fact that the processes of imagination seem particularly prone to affect each other, to mingle and mesh and share their paths of activity, inhibiting or reinforcing nervous impulses in progress, and especially inducing all sorts of neighboring reactions. Consequently their products tend to fuse; images that share some features fuse into one image with emphasis on those features, which thereby are stressed, and dominate the welter of other characters that, for their part, are weakened by fusion. Images, therefore, modify each other, some dominate others, and all tend to become simplified. Emphasis is what gives contours and gradients and other structural elements to images. Emphasis is the natural process of abstraction, whereby our visual representations are made to differ from the direct perceptions that started them. Rudolph Arnheim, in his book, Art and Visual Perception,²⁸ has gone quite deeply into the distinctions between the laws of perception and those of representation. The point of interest here is that the power of abstract symbolic thinking, which plays such a great part in later human mentality, rests on a relatively primitive talent of abstractive seeing that comes with the nature of the visual image.²⁹

The third major condition is simply the fact that images stem from percepts, and the process of their derivation is an original continuity of a peripheral event, the effect of a visible object on the eye, with the further nervous events that terminate in the formation of an image in the brain. The eye is the end-organ of the visual apparatus; what goes on behind the retina, and especially, perhaps, beyond the chiasma, is the rest of our seeing, with all its reverberations and complications and their astounding effects. The recognition of an image as something connected with the external world is intuitive,³⁰ as the response to external things in direct visual perception, which all seeing animals exhibit, is instinctive. This recognition of images as representations of visible things is the basis on which the whole public importance of symbols is built—their use for reference. But there must have been another coincidence to make that happen.

This crucial fourth factor is really part of that lability of imagination, and openness to influence, that we have already remarked; but more precisely, it is the fact that the occurrence of an image may be induced by a great many different kinds of stimulation, either from outside the organism or from within.³¹ Often one cannot tell what evokes a mental image; sometimes a whole situation that often recurs will always do it; for instance, whenever you step out on a pier and smell salt water you may have an image of your first sail boat. Even the salt smell alone may invoke it. So may the mention of the boat’s name. Those are more specific stimuli, but there can be all kinds. This readiness to occur in a total context, but also to be touched off by small fragments of that context encountered in other settings, is the trait that frees the mental image from its original connection with peripheral vision, that is, from the thing it first represented. Add to this the tendency of images with traits in common to fuse and make a simplified image—that is, to become schematic—and you see how much of our image-making would become casual acts of ideation, without any specific memory bonds to perceptual experiences. Not only the images themselves that share a schematic character, but also their representational functions fuse; anyone of them can represent the original percept of any other; that is, as representations whole families of them can stand proxy one for another. Any image of a grasshopper can represent any grasshopper we have actually seen, that was not so distinctive that it created an image too different to fit the schema. If such an oddity appears we form an image of a special kind of grasshopper. With its liberation from perception the image becomes general; and as soon as it can represent something else than its own original stimulus, it becomes a symbol. Schematic similarities in otherwise distinct images make it possible to recall one object through the image of another. Thus, for instance, the outline of the new moon is like that of a small curved boat. We can see the moon as a canoe, or a canoe as a moon. Either assimilation reinforces the perception of shape. This is the natural process of abstraction. We speak of the sickle, the bowl, the disc of the moon in its various phases. In developed thinking we know whether we are talking about the moon or about a boat—that is, we know which image is standing proxy for the other; but studies in the symbolic functions occurring in dream and myth and some psychoses give support to the belief that this is a sober insight which was probably not very early.³² At the level of prehuman image-mongering, the question is rather how one image, even without sensory support, becomes dominant over others, so that they are its symbolic representatives in imagination.

Here, the mechanism seems to be the connection of imagery with emotion. In the complex of images, the one most charged with emotion becomes the dominant image which all the others repeat, reinforce and represent within the brain itself, even below the level of awareness—in the limbo of what Freud called “the dream work,” whereby the significant images, the symbols for conception, are made.

These are, I think, the main physical and behavioral factors that must have existed conjointly in the one animal species that has developed speech: the power of elaborate vocalization, the discriminative ear that heard patterns of sounds, the nervous mechanisms that controlled utterance by hearing of inner and outer sounds, and the tendency to utter long passages of sound in gatherings of many individuals—that is, the habit of joint ululation—with considerable articulation that recurred at about the same point within every such occasion; and, in these same beings, the high mental activity that issued in visual image-making. The gatherings were probably communal rituals, or rather, awesome aesthetic precursors of genuine ritual, the ululations the vocal elements in primitive dance. This idea was propounded long ago by J. Donovan,³³ but no one seems to have paid much attention to it. I adopted it in an early book, Philosophy in a New Key, and the more I reflect on it the more I think it is sound. It was Donovan’s idea that words were not primitive elements in human utterance when it became symbolic, but that meaning first accrued to longer passages, which were gradually broken or condensed into separate bits, each with its own fixed sense. But what he did not say—and I did not see, twenty years ago—was how conceptual meaning accrued to any vocal products at all. I certainly never realized what part the private mental image played in preparing the way for symbolic language—that the whole mechanism of symbolization was probably worked out in the visual system before its power could be transferred to the vocal-auditory realm. Now, with that helpful surmise, let us see how the transfer would be possible, and not too improbable.

In the elaborate development of tribal dance all individuals of the primitive horde became familiar with the vocal sounds that belonged to various sequences of steps and gestures, some perhaps mimetic, others simply athletic, but working up to climaxes of excitement. The “song,” or vocal part of the dance, became more and more differentiated with the evolution of the gestic patterns. At high points there were undoubtedly special shouts and elaborate halloos. In the over-stimulated brains of the celebrants, images must have been evoked at these points of action and special vocalization-images that tended to recur in that context, until for each individual his own symbolic images were built into the familiar patterns of tribal rituals. A dance passage takes time and energy and usually several persons to produce, but the vocal ingredient can be produced with little effort and a minimum of time by any individual. To remember the dance would bring the vocal element to his throat; as the memory of playing “London Bridge” will usually cause a child to hum the tune, with no thought of a bridge or a fair lady, but of the game. So people could reactivate their emotional symbolic images by a snatch of the festal songs. If the dance-action is, say, swinging a club, or even feels like that familiar and expansive act, the various images evoked will be of a club, or clubs, or raising or swinging clubs, or cracking them against each other. It is the image that symbolizes the activity and the objects involved in it. The image is the magical effect of the sound pattern when it is intoned apart from the dance.

The image is a pure conception; it does not signalize or demand its object, but denotes it. Of course, this denotative symbol, the image, begets no communication, for it is purely private. But the things imaged are public, and the sounds that activate images are public; they affect everybody by evoking images at roughly the same moments of dance-action. Within a fairly wide range it does not matter how different the private images are. They are equivalent symbols for the act or the objects that mark those stations in the ritual where the vocal bits belong, which may be uttered out of context by some individual; and suddenly meaning accrues to the phrase, other beings understand, especially if a connoted object is physically at hand, apart from its ritual context.

I suspect that the first meanings of such secularized vocalization were very vague; swing a club, hit a man with a club, kill man and beast, whirl and hit, get hit, wave a club at the moon—may all have belonged by turns to one long utterance, in which the separate articulate parts need not have had any separable meanings.³⁴ But once such passages were used to evoke ideas, their vocalization would quickly become modified; especially by reduction to the speaking voice, which can utter its sounds with more speed and less effort than any singing voice. This everyday utterance would tend to emphasize vowels and consonants—that is, mouth articulations—to replace distinctions of pitch. Some languages have kept tonal distinctions, without precise pitch, as semantic devices. But in most human speech tones serve only for punctuation and emotional coloring.

The great step from anthropoid to anthropos, animal to man, was taken when the vocal organs were moved to register the occurrence of an image, and stirred an equivalent occurrence in another brain, and the two creatures referred to the same thing. At that point, the vocal habit that had long served for communion assumed the function of communication. To evoke ideas in each other’s minds, not in the course of action, but of emotion and memory—that is, in reflection—is to communicate about something; and that is what no animals do.

From then on, speech probably advanced with headlong speed; the vaguely articulated phrases of the gathered horde contracted around their cores of meaning and made long, rich, omnibus words, and broke up into more specifically denotative words, until practically the whole phonetic repertoire was formalized into separable bits, and language entered the synthetic stage of making sentences out of words—the reverse of its pristine articulate process. The new motive of communication must have driven it like wildfire. At this stage if not before, the actual evocation of images became dispensable. We do not need vision to learn speech. The symbolic function has passed to the act of speech itself, and from there finally to the word itself, so that even hearing may be prosthetically replaced. For when verbalization is complete, people have not only speech, but language.

I think there were other uses of speech-like utterance, too—the principle of tracking down the elements of language that may be separately lost by cerebral impairment even today, leads in many directions. Proper names may not have had the same origin as genuine nouns, and numerals are something different again; onomatopoetic words, too, seem to have had their own genesis, apart from the main source of language. But under the influence of language all utterances tended to become words. This is still the case. For instance, our expletives, that have no real verbal meaning in present-day language, always fall under its influence: only a German says “ach”—most Americans cannot even pronounce it; he says “au” where an American says “ouch”; and who but a Frenchwoman would say “ou-la-la”?

Once communication got started, the rise of human mentality may have been cataclysmic, a matter of a few generations wherever it began at all. It must have been an exciting and disconcerting phase of our history. We have traces of it even to this day in the holy fear in which many people hold divine names, blessings, curses, magic formulae—all verbal fragments, imbued with the mystic power of thought that came with speech.

In looking back over all these processes that must have come together to beget language, I am struck by a few outstanding facts: in the first place, the depth to which the foundations go on which this highest of all creature attainments is built; secondly, the complexity of all living functions—for every one of those preparatory traits was itself a highly integrated complex of many nervous processes; thirdly, the fact that not one of the constituents in the new and fateful talent was a mode of animal communication. It seems most likely that the office of communication was taken over by speech, from entirely different activities, when speech was well started; but undoubtedly communication was what henceforth made its history. Finally, it is a notable fact that the two senses which hold the greatest places in the human cortex, sight and hearing, were both needed to produce language; neither a sightless nor a deaf race could have evolved it. If man could either hear no evil or see no evil, he could speak no evil; nor yet any good.

¹ Benjamin Lee Whorf, “Languages and Logic,” The Technology Review, XLIII (1941), 270.

² Note, however, H. J. Pos. “Réflexions sur le problème de l’origine du langage.” Acta Psychologica (1950), who maintains that the primary forms of language were imperative and vocative.

³ John Dewey, in Experience and Nature (Chicago, 1925), says that primitive signs “become language only when used within a context of mutual assistance and direction. The latter are alone of prime importance in considering the transformation of organic gestures and cries into names, things with significance, or the origin of language” (p. 175).

⁴ See, for example, J. B. S. Haldane, “Animal Communication and the Origin of Human Language,” Science Progress, CLXXI (1955), 385-401; and especially, Julian Huxley and Ludwig Koch, Animal Language (London: Country Life, Ltd., 1938).

⁵ K. v. Frisch, Bees: their Vision, Chemical Senses, and Language (Ithaca, New York, 1950). Also, The Dancing Bees (New York, 1955).

⁶ This is the view expressed by Charles Morris in The Nature of Mind (Houston, Texas, 1929) and in Signs, Language and Behavior (New York, 1946); also by J. Dewey, op. cit., and elsewhere.

⁷ As L. Bontan remarked in his article on the vocal habits of gibbons, “les animaux n’ont pas un langage rudimentaire. Leur langage n’est pas un langage. . . .” (“Le pseudo-langage. Observations effectés sur un anthropoide, le gibbon [Hylobates Leucogenvs-Obilby),” Actes de la Soc. Linnénne de Bordeaux, LCVII (1913), 5-77.)

⁸ See R. M. Yerkes and H. W. Nissen, “Prelinguistic Sign Behavior in the Chimpanzee,” Science, N.S. LXXXIX (1939), 585-587. The upshot of the reported experiments is “that delayed response, in the absence of spatial cues or with misleading cues, is either extremely difficult or impossible for most chimpanzees . . . . There is abundant evidence that various other types of sign process than the symbolic are of frequent occurrence and function effectively in the chimpanzee” (p. 587). Perhaps the title: “Nonlinguistic Sign Behavior . . .” would have been more accurate.

Despite such observations, the authors of Animal Language do not hesitate to attribute conversation to monkeys, and even to animals below the primates, nor to refer to their repertoire of sounds as a vocabulary having direct affinities with human speech. “The gregarious baboons,” writes Mr. Huxley, who composed the text, “are very conversational animals. Most of its communications, both in the pack and in its component family groups, are effected by voice” (Huxley and Koch. op. cit., p. 55). And more remarkable still: “The sea-lions, . . . as befits their social and intelligent nature, are noisy animals, and possess a considerable vocabulary, although the different sounds are all variations on one theme—the familiar, rather raucous bark. Mr. Koch believes that sea-lions also express different meanings (as do the Chinese) by merely changing the pitch of their note” (ibid., p. 49).

⁹ Cf. Leo Loeb, The Biological Basis of Individuality (Springfield, Illinois: Charles C. Thomas, 1945).

¹⁰ Cf. W. H. Marshall and S. A. Talbot, “Recent Evidence for Neural Mechanisms in Vision Leading to a General Theory of Sensory Activity.” In H. Kluver’s Visual Mechanisms (1942). pp. 117-164. “In the cat, optic tract endings in the geniculate divide into several branches and as many as forty ring-shaped boutons have been seen on single radiation cells which may come from as many as ten optic tract fibers. Each fiber also divides to form synapses with several radiation cells. In addition to bouton contacts the radiation cells have numerous dendritic processes, with which the optic tract endings make apparently more numerous synapses . . . than with the radiation cells themselves” (p. 122).

Cf. R. Lorente de Nó, “Vestibulo-Ocular Reflex Arc,” Arch. Neurol. and Psychiat., XXX (1933), 245-291. “On each cell in the nervous system numerous synapses, sometimes several thousand, are found. The synapses are always of different kinds, occasionally of ten or more” (p. 279).

¹¹ A belief which has, indeed, been challenged a good many times; but it seems to be ingrained.

¹² C. J. Herrick & G. E. Coghill, “The Development of Reflex Mechanism in Amblystoma,” J. of Compo Neurol., xxv (1915).

¹³ Op. cit., p. 247. Here the simplification serves for economy; but Gerhardt v. Bonin, in his essay “Types and Similitudes,” Philosophy of Science, XIII (1946), 196-202 observes that “the paleontological evidence has presented cases, such as the ammonites, where evolution produced at first more and more complicated, and later simpler and simpler forms” (p. 198).

¹⁴ E.g., Lord Haldane, op. cit., says, “A Pithecanthropus child which gave the danger call or the food discovery call without due cause was probably punished” (p. 398). But animals do not punish their young for mischief done; the “cuffing” a cub may receive from its mother is always interference with its momentary annoying act, to stop it. The concept of a deed, and hence praise and punishment, belong to human life.

¹⁵ See esp. M. Isserlin, “Über Agrammatismus,” Ztschr. f. d. ges. Neurol. u. Psychiat., LXXV (1922), 332-410.

¹⁶ H. Kogerer, “Worttaubheit, Melodientaubbheit, Gebardeagnosie,” Ztschr. f. d. ges. Neurol. u. Psychiat., XCLL (1924),469-483. Also H. Liepmann and M. Pappenheim, “Über einen Fall von sogenannter Leitungsaphasie mit anatomischem Befund,” Ztschr. f. d. ges. Neurol. u. Psychiat., XXVII (1915), 1-41.

¹⁷ All these special forms are listed in J. M. Nielsen’s Agnosia, Apraxia, Aphasia (New York, 1936; 2nd ed., 1946).

¹⁸ Goodhart & Savitsky, “Alexia following Injuries of the Head,” Archives of Neurol. & Psychiatry, XXX (1933), 223-224.

²⁰ J. M. Nielsen, “Visual Agnosia for Animate Objects. Report of a Case with Autopsy.” Trans. Amer. Neurol. Assoc. (1942), pp. 128-130.

²¹ See G. Kelemen, “Structure and Performance in Animal Language,” Archives of Otolaryngology, L (1949), 740-744.

²³ Otto Kalischer. “Das Grosshirn der Papageien in anatomischer und physiologischer Beziehung,” Abhandlungen der königl.-Preuss. Akademie der Wissenschaften, IV (1905), 1-105. A study based on ten years’ work of training, operating, retraining, finally autopsying some 60 talking parrots.

²⁴ Louis Guggenheim, Phylogenesis of the Ear (Culver City, California, 1948). See p. 78.

²⁵ For instance, Sylvano Arieti, with whose views of symbol formation I agree in some respects (as will shortly be apparent), holds that in a primordial family a baby might babble “ma-ma” and associate the utterance “with the mother or with the image of the mother”; and that “if a second sibling understands that the sound ma-ma refers to mother, language is originated . . . . But at this level the sound ma-ma refers only to a particular mother . . . and not to any mother. In other words, the symbol ma-ma denotes, but does not possess much connotation power.” “The Possibility of Psychosomatic Involvement of the Central Nervous System in Schizophrenia,” J. of Nervous & Mental Disease, CXXIII (1956), 324-333. See esp. p. 32. Also J. S. Kasanin, “The Disturbance of Conceptual Thinking in Schizophrenia,” in Language and Thought in Schizophrenia, ed. by J .S. Kasanin and N. D. C. Lewis (U. of Calif. Press, 1944): “—when the child says ‘table’ or ‘chair’ he does not mean tables or chairs in general, but the table or chair which is in his house or which belongs to him.”

²⁶ This difference in the frequency, intensity, and clarity of images in human and animal brains, is strikingly corroborated and anatomically explained in Niessl v. Mayendorf’s article, “Über den vasomotorischen Mechanisms der Halluzinationenem,” Ztschr. f. d. ges. Neurol. u. Psychiat., CXIV (1928), 311-322.

²⁷ This fact is mentioned by P. L. Short in his paper, “The Objective Study of Mental Imagery,” Brit. J. of Psychol., XLIV (1953), 38-51, where he writes: “. . . in thinking, it is the images that occur most readily and habitually that are important, not the ones thought to be most ‘intense’ or ‘vivid’ at a given moment. The mere emergence of very vivid images may not be associated at all with the tendency to have and to use images” (p. 38). He also notes the importance of the connection between percepts and centrally produced images.

²⁹ Some interesting comments on abstractive seeing may also be found in Leo Steinberg’s paper, “The Eye is a Part of the Mind,” Partisan Review, XX (1953), 194-212. (Reprinted in Reflections on Art, Johns Hopkins University Press, 1958.) There are also various studies of the neural processes involved in such sensory abstraction, e.g., D. M. Purdy’s “The Structure of the Visual World.” Psychol. Rev., XLIII (1936), 59-82, esp., Part III; Fred Attneave’s technological essay, “Some Informational Aspects of Visual Perception,” Psychol. Rev., LXI (1954), 183-193; Norbert Wiener’s Cybernetics (New York, 1948); and esp. in a study by W. H. Marshall and S. A. Talbot. “Recent Evidence for Neural Mechanisms in Vision Leading to a General Theory of Sensory Acuity,” in H. Kluver’s Visual Mechanisms (1942). pp. 117-164.

³⁰ Cf. D. Forsyth, “The Infantile Psyche, with Special Reference to Visual Projection,” Brit. J. of Psychol., XI (1920-21), 263-276.

³¹ D. Forsyth, op. cit. (p. 265): “The visual organ . . . transmits a centripetal wave of excitement which is registered in the mind as a memorative impression of the excitation. This visual memory becomes associated with inner (somatic) excitations, and can subsequently be activated from either of the two directions in which it has established excitatory connections. . . .”

³² The sources substantiating this proposition are too scattered and numerous to quote. One of the first explicit statements of it is found in an article which has become a classic—Herbert Silberer’s “Über die Symbolbildung,” Jahrbuch für psychoanalytische u. psychopathologische Forschungen, III (1912), 661-723, republished in English translation, unfortunately with some deletions, in David Rapaport’s anthology, Organization and Pathology of Thought (New York, 1951). Silberer wrote: “A people which speaks in metaphors does not experience what it says as metaphoric; the symbols it uses are regarded by it not as symbols, but rather as realities . . . .” (Rapaport, p. 212). They certainly all contradict the claim of J. P. Sartre in L’Imagination (Paris, 1948), p. 104, that one never mistakes a phantasy image for a percept: “Aucune image, jamais, ne vient se méler aux choses réelles” (p. 109). And further: “. . . il m’est impossible de former une image sans savoir en même temps que je forme une image . . .” (p. 110)

³³ “The Festal Origin of Human Speech:’ Mind, 0.S. XVI (1891), 498-506, and XVII (1892), 325-339.

³⁴ In the 1957 ed. of the Encyclopaedia Britannica, s.v. “Language,” Otto Jespersen voices the same opinion.

The Origins of Speech and Its Communicative Function

Langer main page