The gift of speech and a well ordered language are characteristic of every known group of human beings. No tribe has ever been found which is without language and all statements to the contrary may be dismissed as mere folklore. There seems to be no warrant whatever for the statement which is sometimes made that there are certain peoples whose vocabulary is so limited that they cannot get on without the supplementary use of gesture, so that intelligible communication between members of such a group becomes impossible in the dark. The truth of the matter is that language is an essentially perfect means of expression and communication among every known people. Of all aspects of culture it is a fair guess that language was the first to receive a highly developed form and that its essential perfection is a prerequisite to the development of culture as a whole.
There are some general characteristics which apply to all languages, living or extinct, written or unwritten. In the first place language is primarily a system of phonetic symbols for the expression of communicable thought and feeling. In other words, the symbols of language are differentiated products of the vocal behavior which is associated with the larynx of the higher mammals. As a mere matter of theory it is conceivable that something like a linguistic structure could have been evolved out of gesture or other forms of bodily behavior. The fact that at an advanced stage in the history of the human race writing emerged in close imitation of the patterns of spoken language proves that language as a purely instrumental and logical device is not dependent on the use of articulate sounds. Nevertheless, the actual history of man and a wealth of anthropological evidence indicate with overwhelming certainty that phonetic language takes precedence over all other kinds of communicative symbolism, which are by comparison either substitutive, like writing, or merely supplementary, like the gesture accompanying speech. The speech apparatus which is used in the articulation of language is the same for all known peoples. It consists of the larynx, with its delicately adjustable glottal chords, the nose, the tongue, the hard and soft palate, the teeth and the lips. While the original impulses leading to speech may be thought of as localized in the larynx, the finer phonetic articulations are due chiefly to the muscular activity of the tongue, an organ whose primary function has of course nothing whatever to do with sound production but which in actual speech behavior is indispensable for the development of emotionally expressive sound into what we call language It is so indispensable in fact that one of the most common terms for language or speech is "tongue." Language is thus not a simple biological function even as regards the simple matter of sound production, for primary laryngeal patterns of behavior have had to be completely overhauled by the interference of lingual, labial and nasal modifications before a "speech organ" was ready for work. Perhaps it is because this speech organ is a diffused and secondary network of physiological activities which do not correspond to the primary functions of the organs involved that language has been enabled to free itself from direct bodily expressiveness.
Not only are all languages phonetic in character; they are also "phonemic." Between the articulation of the voice into the phonetic sequence, which is immediately audible as a mere sensation, and the complicated patterning of phonetic sequences into such symbolically significant entities as words, phrases and sentences there is a very interesting process of phonetic selection and generalization which is easily overlooked but which is crucial for the development of the specifically symbolic aspect of language. Language is not merely articulated sound; its significant structure is dependent upon the unconscious selection of a fixed number of "phonetic stations," or sound units. These are in actual behavior individually modifiable; but the essential point is that through the unconscious selection of sounds as phonemes definite psychological barriers are erected between various phonetic stations, so that speech ,ceases to be an expressive flow of sound and becomes a symbolic composition with limited materials or units. The analogy with musical theory seems quite fair. Even the most resplendent and dynamic symphony is built up of tangibly distinct musical entities or notes which in the physical world flow into each other in an indefinite continuum but which in the world of aesthetic composition and appreciation are definitely bounded off against each other, so that they may enter into an intricate mathematics of significant relationships. The phonemes of a language are in principle a distinct system peculiar to the given language, and its words must be made up, in unconscious theory if not always in actualized behavior, of these phonemes. Languages differ very widely in their phonemic structure. But whatever the details of these structures may be,
(156) the important fact remains that there is no known language which has not a perfectly definite phonemic system. The difference between a sound and a phoneme can be illustrated by a simple example in English. If the word matter is pronounced in a slovenly fashion, as in the phrase "What's the matter?" the t sound, not being pronounced with the full energy required to bring out its proper physical characteristics, tends to slip into a d. Nevertheless, this phonetic d will not be felt as a functional d but as a variety of t of a particular type of expressiveness. Obviously the functional relation between the proper t sound of such a word as matter and its d variant is quite other than the relation of the t of such a word as town and the d of down. In every known language it is possible to distinguish merely phonetic variations, whether expressive or not, from symbolically functional ones of a phonemic order.
In all known languages phonemes are built up into distinct and arbitrary sequences which are at once recognized by the speakers as meaningful symbols of reference. In English, for instance, the sequence g plus o in the word go is an unanalyzable unit and the meaning attaching to the symbol cannot be derived by relating to each other values which might be imputed to the g and to the o independently. In other words, while the mechanical functional units of language are phonemes, the true units of language as symbolism are conventional groupings of such phonemes. The size of these units and the laws of their mechanical structure vary widely in the different languages and their limiting conditions may be said to constitute the phonemic mechanics, or phonology, of a particular language. But the fundamental theory of sound symbolism remains the same everywhere. The formal behavior of the irreducible symbol also varies within wide limits in the languages of the world. Such a unit may be either a complete word, as in the English example already given, or a significant element, like the suffix ness of goodness. Between the meaningful and unanalyzable word or word element and the integrated meaning of continuous discourse lies the whole complicated field of the formal procedures which are intuitively employed by the speakers of a language in order to build up aesthetically and functionally satisfying symbol sequences out of the theoretically isolable units. These procedures constitute grammar, which may be defined as the sum total of formal economies intuitively recognized by the speakers of a language. There seem to be no types of cultural patterns which vary more surprisingly and with a greater exuberance of detail than the morphologies of the known languages. In spite of endless differences of detail, however, it may justly be said that all grammars have the same degree of fixity. One language may be more complex or difficult grammatically than another, but the-re is no meaning whatever in the statement which is sometimes made that one language is more grammatical, or form bound, than another. Our rationalizations of the structure of our own language lead to a self-consciousness of speech and of academic discipline which are of course interesting psychological and social phenomena in themselves but have very little to do with the question of form in language.
Besides these general formal characteristics language has certain psychological qualities which make it peculiarly important for the student of social science. In the first place, language is felt to be a perfect symbolic system, in a perfectly homogeneous medium, for the handling of all references and meanings that a given culture is capable of, whether these be in the form of actual communications or in that of such ideal substitutes of communication as thinking. The content of every culture is expressible in its language and there are no linguistic materials whether as to content or form which are not felt to symbolize actual meanings, whatever may be the attitude of those who belong to other cultures. New cultural experiences frequently make it necessary to enlarge the resources of a language, but such enlargement is never an arbitrary addition to the materials and forms already present; it is merely a further application of principles already in use and in many cases little more than a metaphorical extension of old terms and meanings. It is highly important to realize that once the form of a language is established it can discover meanings for its speakers which are not simply traceable to the given quality of experience itself but must be explained to a large extent as the projection of potential meanings into the, raw material of experience. If a man who has never seen more than a single elephant in the course of his life nevertheless speaks without the slightest hesitation of ten elephants or a million elephants or a herd of elephants or of elephants walking two by two or three by three or of generations of elephants, it is obvious that language has the power to analyze experience into theoretically dissociable elements: and to create that world of the potential intergrading
( 157) with the actual which enables human beings to transcend the immediately given in their individual experiences and to join in a larger common understanding. This common understanding constitutes culture, which cannot be adequately defined by a description of those more colorful patterns of behavior in society which lie open to observation. Language is heuristic, not merely in the simple sense which this example suggests but in the much more far reaching sense that its forms predetermine for us certain modes of observation and interpretation. This means of course that as our scientific experience grows we must learn to fight the implications of language. "The grass waves in the wind" is shown by its linguistic form to be a member of the same relational class of experiences as "The man works in the house." As an interim solution of the problem of expressing the experience referred to in this sentence it is clear that the language has proved useful, for it has made significant use of certain symbols of conceptual relation, such as agency and location. If we feel the sentence to be poetic or metaphorical, it is largely because other more complex types of experience with their appropriate symbolisms of reference enable us to reinterpret the situation and to say, for instance, "The grass is waved by the wind" or "The wind causes the grass to wave." The point is that no matter how sophisticated our modes of interpretation become, we never really get beyond the projection and continuous transfer of relations suggested by the forms of our speech. After all, to say that "Friction causes such and such a result" is not very different from saying that "The grass waves in the wind." Language is at one and the same time helping and retarding us in our exploration of experience, and the details of these processes of help and hindrance are deposited in the subtler meanings of different cultures.
A further psychological characteristic of language is the fact that while it may be looked upon as a symbolic system which reports or refers to or otherwise substitutes for direct experience, it does not as a matter of actual behavior stand apart from or run parallel to direct experience but completely interpenetrates with it. This is indicated by the widespread feeling, particularly among primitive people, of that virtual identity or close correspondence of word and thing which leads to the magic of spells. On our own level it is generally difficult to make a complete divorce between objective reality and our linguistic symbols of reference to it; and things, qualities and events are on the whole felt to be what they are called. For the normal person every experience, real or potential, is saturated with verbalism. This explains why so many lovers of nature, for instance, do not feel that they are truly in touch with it until they have mastered the names of a great many flowers and trees, as though the primary world of reality were a verbal one and as though one could not get close to nature unless one first mastered the terminology which somehow magically expresses it. It is this constant interplay between language and experience which removes language from the cold status of such purely and simply symbolic systems as mathematical symbolism or flag signaling. This interpenetration is not only an intimate associative fact; it is also a contextual one. It is important to realize that language may not only refer to experience or even mold, interpret and discover experience but that it also substitutes for it in the sense that in those sequences of interpersonal behavior which form the greater part of our daily lives speech and action supplement each other and do each other's work in a web of unbroken, pattern. If one, says to me "Lend me a dollar," I may hand over the money without a word or I may give it with an accompanying "Here it is" or I may say "I haven't got it. I'll give it to you tomorrow." Each of these responses is structurally equivalent, if one thinks of the larger behavior pattern. It is clear that if language is in its analyzed form a symbolic system of reference it is far from being merely that if we consider the psychological part that it plays in continuous behavior. The reason for this almost unique position of intimacy which language holds among all known symbolisms is probably the fact that it is learned in the earliest years of childhood.
It is because it is learned early and piecemeal, in constant association with the color and the requirements of actual contexts, that language in spite of its quasi-mathematical form is rarely a purely referential organization. It tends to be so only in scientific discourse, and even there it may be seriously doubted whether the ideal of pure reference is ever attained by, language. Ordinary speech is directly expressive and the purely formal patterns of sounds, words, grammatical forms, phrases and sentences are always to be thought of as compounded with intended or unintended symbolisms of expression, if they are to be understood fully from the standpoint of behavior. The choice of words in a particular context may convey, the opposite of what they
(158) mean on the surface. The same external message is differently interpreted according to whether the speaker has this or that psychological status in his personal relations, or whether such primary expressions as those of affection or anger or fear may inform the spoken words with a significance which completely transcends their normal value. On the whole, however, there is no danger that the expressive character of language will be overlooked. It is too obvious a fact to call for much emphasis. What is often overlooked and is, as a matter of fact, not altogether easy to understand is that the quasi-mathematical patterns, as we have called them, of the grammarian's language, unreal as these are in a contextual sense, have nevertheless a tremendous intuitional vitality; and that these patterns, never divorced in experience from the expressive ones, are nevertheless easily separated from them by the normal individual. The fact that almost any word or phrase can be made to take on an infinite variety of meanings seems to indicate that in all language behavior there are intertwined in enormously complex patterns isolable patterns of two distinct orders. These may be roughly defined as patterns of reference and patterns of expression.
That language is a perfect symbolism of experience, that in the actual contexts of behavior it cannot be divorced from action and that it is the carrier of an infinitely nuanced expressiveness are universally valid psychological facts. There is a fourth general psychological peculiarity which applies more particularly to the languages of sophisticated peoples. This is the fact that the referential form systems which are actualized in language behavior do not need speech in its literal sense in order to preserve their substantial integrity. The history of writing is in essence the long attempt to develop an independent symbolism on the basis of graphic representation, followed by the slow and begrudging realization that spoken language is a more powerful symbolism than any graphic one can possibly be and that true progress in the art of writing lay in the virtual abandonment of the principle with which it originally started. Effective systems of writing, whether alphabetic or not, are more or less exact transfers of speech. The original language system may maintain itself in other and remoter transfers, one of the best examples of these being the Morse telegraph code. It is a very interesting fact that the principle of linguistic transfer is not entirely absent even among the unlettered peoples of the world. Some at least of the drum signal and horn signal systems of the west African natives are in principle transfers of the organizations of speech, often in minute phonetic detail.
Many attempts have been made to unravel the origin of language but most of these are hardly more than exercises of the speculative imagination. Linguists as a whole have lost interest in the problem and this for two reasons. In the first place, it has come to be realized that there exist no truly primitive languages in a psychological sense, that modern researches in archaeology have indefinitely extended the time of man's cultural past and that it is therefore vain to go much beyond the perspective opened up by the study of actual languages. In the second place, our knowledge of psychology, particularly of the symbolic processes in general, is not felt to be sound enough or far reaching enough to help materially with the problem of the emergence of speech. It is probable that the origin of language is not a problem that can ,be solved out of the resources of linguistics alone but that it is essentially a particular case of a much wider problem of the genesis of symbolic behavior and of the specialization of such behavior in the laryngeal region, which may be presumed to have had only expressive functions to begin with. Perhaps a close study of the behavior of very young children under controlled conditions may provide some valuable hints, but it seems dangerous to reason from such experiments to the behavior of precultural man. It is more likely that the kinds of studies which are now in progress of the behavior of the higher apes will help supply some idea of the genesis of speech.
The most popular earlier theories were the interjectional and onomatopoetic theories. The former derived speech from involuntary cries of an expressive nature, while the latter maintained that the words of actual language are conventionalized forms of imitation of the sounds of nature. Both of these theories suffer from two fatal defects. While it is true that both interjectional and onomatopoetic elements are found in most languages, they are always relatively unimportant and tend to contrast somewhat with the more normal materials of language. The very fact that they are constantly being formed anew seems to indicate that they belong rather to the directly expressive layer of speech which intercrosses with the main level of referential symbolism. The second difficulty is even more serious. The essential problem of the origin of
( 159) speech is not to attempt to discover the kinds of vocal elements which constitute the historical nucleus of language. It is rather to point out how vocal articulations of any sort could become dissociated from their original expressive value. About all that can be said at present is that while speech as a finished organization is a distinctly human achievement, its roots probably lie in the power of the higher apes to solve specific problems by abstracting general forms or schemata from the details of given situations; that the habit of interpreting certain selected elements in a situation as signs of a desired total one gradually led in early man to a dim feeling for symbolism; and that in the long run and for reasons which can hardly be guessed at the elements of experience which were most often interpreted in a symbolic sense came to be the largely useless or supplementary vocal behavior that must have often attended significant action. According to this point of view language is not so much directly developed out of vocal expression as it is an actualization in terms of vocal expression of the tendency to master reality, not by direct and ad hoc handling of its elements but by the reduction of experience to familiar forms. Vocal expression is only superficially the same as language. The tendency to derive speech from emotional expression has not led to anything tangible in the way of scientific theory and the attempt must now be made to see in language the slowly evolved product of a peculiar technique or tendency which may be called the symbolic one, and to see the relatively meaningless or incomplete part as a sign of the whole. Language then is what it is essentially not because of its admirable expressive power but in spite of it. Speech as behavior is a wonderfully complex blend of two pattern systems, the symbolic and the expressive, neither of which could have developed to its present perfection without the interference of the other.
It is difficult to see adequately the functions of language, because it is so deeply rooted in the whole of human behavior that it may be suspected that there is little in the functional side of our conscious behavior in which language does not play its part. The primary function of language is generally said to be communication. There can be no quarrel with this so long as it is distinctly understood that there may be effective communication without overt speech and that language is highly relevant to situations which are not obviously of a communicative sort. To say that thought, which is hardly possible in any sustained sense without the symbolic organization brought by language, is that form of communication in which the speaker and the person addressed are identified in one person is not far from begging the question. The autistic speech of children seems to show that the purely communicative aspect of language has been exaggerated. It is best to admit that language is primarily a vocal actualization of the tendency to see reality symbolically, that it is precisely this quality which renders it a fit instrument for communication and that it is in the actual give and take of social Intercourse that it has been complicated and refined into the form in which it is known today. Besides the very general function which language fulfils in the spheres of thought, communication and expression which are implicit in its very nature there may be pointed out a number of special derivatives of these which are of particular interest to students of society.
Language is a great force of socialization, probably the greatest that exists. By this is meant not merely the obvious fact that significant social intercourse is hardly possible without language but that the mere fact of a common speech serves as a peculiarly potent symbol o f the social solidarity of those who speak the language. The psychological significance of this gees far beyond the association of particular languages with nationalities, political entities or smaller local groups. In between the recognized dialect or language as a whole and the individualized speech of a given individual lies a kind of linguistic unit which is not often discussed by the linguist but which is of the greatest importance to social psychology. This is the subform of a language which is current among a group of people who are held together by ties of common interest. Such a group may be a family, the undergraduates of a college, a labor union, the underworld in a large city, the members of a club, a group of four or five friends who hold together through life in spite of differences of professional interest, and untold thousands of other kinds of groups. Each of these tends to develop peculiarities of speech which have the symbolic function of somehow distinguishing the group from the larger group into which its members might be too completely absorbed. The complete absence of linguistic indices of such small groups is obscurely felt as a defect or sign of emotional poverty. Within the confines of a particular family, for instance, the name Georgy, having once been mispronounced
( 160) Doody in childhood, may take on the latter form forever after; and this unofficial pronunciation of a familiar name as applied to a particular person becomes a very important symbol indeed of the solidarity of a particular family and of the continuance of the sentiment that keeps its members together. A stranger cannot lightly take on the privilege of saying Doody if the members of the family feel that he is not entitled to go beyond the degree of familiarity symbolized by the use of Georgy or George. Again, no one is entitled to say "trig" or "math" who has not gone through certain familiar and painful experiences as a high school or undergraduate student. The use of such words at once declares the speaker a member of an unorganized but psychologically real group. A self-made mathematician has hardly the right to use the word "math" in referring to his own interests because the student overtones of the word do not properly apply to him. The extraordinary importance of minute linguistic differences for the symbolization of psychologically real as contrasted with politically or sociologically official groups is intuitively felt by most people. "He talks like us" is equivalent to saying "He is one of us."
There is another important sense in which language is a socializer beyond its literal use as a means of communication. This is in the establishment of rapport between the members of a physical group, such as a house party. It is not what is said that matters so much as that something is said. Particularly where cultural understandings of an intimate sort are somewhat lacking among the members of a physical group it is felt to be important that the lack be made good by a constant supply of small talk. This caressing or reassuring quality of speech in general, even where no one has anything of moment to communicate, reminds us how much more language is than a mere technique of communication. Nothing better shows how completely the life of man as an animal made over by culture is dominated by the verbal substitutes for the physical world.
The use of language in cultural accumulation and historical transmission is obvious and important. This applies not only to sophisticated levels but to primitive ones as well. A great deal of the cultural stock in trade of a primitive society is presented in a more or less well defined linguistic form. Proverbs, medicine formulae, standardized prayers, folk tales, standardized speeches, song texts, genealogies, are some of the more overt forms which language takes as a culture preserving instrument. The pragmatic ideal of education, which aims to reduce the influence of standardized lore to a minimum and to get the individual to educate himself through as direct a contact as possible with the facts of his environment, is certainly not realized among the primitive-s, who are often as word bound as the humanistic tradition itself. Few cultures perhaps have gone to the length of the classical Chinese culture or of rabbinical Jewish culture in making the word do duty for the thing or the personal experience as the ultimate unit of reality. Modern civilization, as a whole, with its schools, its libraries and its endless stores of knowledge, opinion and sentiment stored up in verbalized form, would be unthinkable without language made eternal as document. On the whole, we probably tend to exaggerate the differences between "high" and "low" cultures or saturated and emergent cultures in the matter of traditionally conserved verbal authority. The enormous differences that seem to exist are rather differences in the outward form and content of the cultures themselves than in the psychological relation which obtains between the individual and his culture.
In spite of the fact that language acts as a socializing and uniformizing force it is at the same time the most potent single known factor for the growth of individuality. The fundamental quality of one's voice, the phonetic patterns of speech, the speed and relative smoothness of articulation, the length and build of the sentences, the character and range of the vocabulary, the stylistic consistency o f the words used, the readiness with which words respond to the requirements of the social environment, in particular the suitability of one's language to the language habits of the person addressed-all these are so many complex indicators of the personality. "Actions speak louder than words" may be an excellent maxim from the pragmatic point of view but betrays little insight into the nature of speech. The language habits of people are by no means irrelevant as unconscious indicators of the more important traits of their personalities, and the folk is psychologically wiser than the adage in paying a great deal of attention willingly or not to the psychological significance of a man's language. The normal person is never convinced by the mere content of speech but is very sensitive to many of the implications of language behavior, however feebly (if at all) these may have been consciously analyzed. All
(161) in all, it is not too much to say that one of the really important functions of language is to be constantly declaring to society the psychological place held by all of its members. Besides this more general type of personality expression or fulfilment there is to be kept in mind the important role which language plays as a substitutive means of expression for those individuals who have a greater than normal difficulty in adjusting themselves to the environment in terms of primary action patterns. Even in the most primitive cultures the strategic word is likely to be more powerful than the direct blow. It is unwise to speak too blithely of "mere" words, for to do so may be to imperil the value and perhaps the very existence of civilization and personality.
The languages of the world may be classified either structurally or genetically. An adequate structural analysis is an intricate matter and no classification seems to have been suggested which does justice to the bewildering variety of known forms. It is useful to recognize three distinct criteria of classification: the relative degree of synthesis or elaboration of the words of the language; the degree to which the various parts of a word are welded together; and the extent to which the fundamental relational concepts of the language are directly expressed as such. As regards synthesis languages range all the way from the isolating type, in which the single word is essentially unanalyzable, to the type represented by many American Indian languages, in which the single word is functionally often the equivalent of a sentence with many concrete references that would in most languages require the use of a number of words. Four stages of synthesis may be conveniently recognized; the isolating type, the weakly synthetic type, the fully synthetic type and the polysynthetic type. The classical example of the first type is Chinese, which does not allow the words of the language to be modified by internal changes or the addition of prefixed or suffixed elements to express such concepts as those of number, tense, mode, case relation and the like. This seems to be one of the more uncommon types of language and is best represented by a number of languages in eastern Asia. Besides Chinese itself Siamese, Burmese, modern Tibetan, Annamite and Khmer, or Cambodian, may be given as examples. The older view, which regarded such languages as representing a peculiarly primitive stage in the evolution of language, may now be dismissed as antiquated. All evidence points to the contrary hypothesis that such languages are the logically extreme analytic developments of more synthetic languages which because of processes of phonetic disintegration have had to reexpress by analytical means combinations of ideas originally expressed within the framework of the single a word. The weakly synthetic type of language is best represented by the most familiar modern languages of Europe, such as English, French, Spanish, Italian, German, Dutch and Danish. Such languages modify words to some extent but have only a moderate formal elaboration of the word. The plural formations of English and French, for instance, are relatively simple and the tense and modal systems of all the languages of this type tend to use analytic methods as supplementary to the older synthetic one. The third group of languages is represented by such languages as Arabic and earlier Indo-European languages, like Sanskrit, Latin and Greek. These are all languages of great formal complexity, in which classificatory ideas, such as sex gender, number, case relations, tense and mood, are expressed with considerable nicety and in a great variety of ways. Because of the rich formal implications of the single word the sentence tends not to be so highly energized and ordered as in the first mentioned types. Lastly, the polysynthetic languages add to the formal complexity of the treatment of fundamental relational ideas the power to arrange a number of logically distinct, concrete ideas into an ordered whole within the confines of a single word. Eskimo and Algonquin are classical examples of this type.
From the standpoint of the mechanical cohesiveness with which the elements of words are united languages may be conveniently grouped into four types. The first of these, in which there is no such process of combination, is the isolating type already referred to. To the second group of languages belong all those in which the word can be adequately analyzed into a mechanical sum of elements, each of which has its more or less clearly established meaning and each of which is regularly used in all other words into which the associated notion enters. These are the so-called agglutinative languages. The majority of languages seem to use the agglutinative technique, which has the great advantage of combining logical analysis with economy of means. The Altaic languages, of which Turkish is a good example, and the Bantu languages of Africa are agglutinative in form. In the third type, the so-called inflective languages, the degree of union between the radical element or
( 162) stem of the word and the modifying prefixes or suffixes is greater than in the agglutinative languages, so that it becomes difficult in many cases to isolate the stem and set it off against the accreted elements. More important than this, however, is the fact that there is less of a one to one correspondence between .the linguistic element and the notion referred to than in the agglutinative languages. In Latin, for instance, the notion of plurality is expressed in a great variety of ways which seem to have little phonetic connection with each other. For example, the final vowel or diphthong of equi (horses), dona (gifts), mensae (tables) and the final vowel and consonant of hostes (enemies) are functionally equivalent elements the distribution of which is dependent on purely formal and historical factors that have no logical relevance. Furthermore in the verb the notion of plurality is quite differently expressed, as in the last two consonants of amant (they love). It used to be fashionable to contrast in a favorable sense the "chemical" qualities of such inflective languages as Latin and Greek with the soberly mechanical quality of such languages as Turkish. But these evaluations may now be dismissed as antiquated and subjective. They were obviously due to the fact that scholars who wrote in English, French and German were not above rationalizing the linguistic structures with which they were most familiar into a position of ideal advantage. As an offshoot of the inflective languages may be considered a fourth group, those in which the processes of welding, due to the operation of complex phonetic laws, have gone so far as to result in the creation of patterns of internal change of the nuclear elements of speech. Such familiar English examples as the words sing, sang, sung, song will serve to give some idea of the nature of these structures, which may be termed symbolistic. The kinds of internal change which may be recognized are changes in vocalic quality, changes in consonants, changes in quantity, various types of reduplication or repetition, changes in stress accent and, as in Chinese and many African languages, changes in pitch. The classical example of this type of language is Arabic, in which as in the other Semitic languages nuclear meanings are expressed by sequences of consonants, which have, however, to be connected by significant vowels whose sequence patterns establish fixed functions independent of the meanings conveyed by the consonantal framework.
Elaboration and technique of word analysis are perhaps of less logical and psychological significance than the selection and treatment of fundamental relational concepts for grammatical treatment. It would be very difficult, however, to devise a satisfactory conceptual classification of languages because of the extraordinary diversity of the concepts and classifications of ideas which are illustrated in linguistic form. In the Indo-European and Semitic languages, for instance, noun classification on the basis of gender is a vital principle of structure; but in most of the other languages of the world this principle is absent, although other methods of noun classification are found. Again, tense or case relations may be formally important in one language, for example, Latin, but of relatively little grammatical importance in another, although the logical references implied by such forms must naturally be taken care of in the economy of the language, as, for instance, by the use of specific words within the framework of due sentence. Perhaps the most fundamental conceptual basis of classification is that of the expression of fundamental syntactic relations as such versus their expression in necessary combination with notions of a concrete order. In Latin, for example, the notion of the subject of a predicate is never purely expressed in a formal sense., because there is no distinctive symbol for this relation. It is impossible to render it without at the same time defining the number and gender of the subject of the sentence. There are languages, however, in which syntactic relations are expressed purely, without admixture of implications of a nonrelational sort. We may speak therefore of pure relational languages as contrasted with mixed relational languages. Most of the languages with which we are familiar belong to the latter category. It goes without saving that such a conceptual classification has no direct relation to the other two types of classification which we have mentioned.
The genetic classification of languages is one which attempts to arrange the languages of the world in groups and subgroups in accordance with the main lines of historical connection, which can be worked out o n the basis either of documentary evidence or of a careful comparison of the languages studied. Because of the far reaching effect of slow phonetic changes and of other causes languages which were originally nothing but dialects of the same form of speech have diverged so widely that it is not apparent that they are but specialized developments of a single prototype. An enormous amount of work
( 163) has been done in the genetic classification and subclassification of the languages of the world, but very many problems still await research and solution. At the present time it is known definitely that there are certain very large linguistic groups, or families, as they are often called, the members of which may, roughly speaking, be looked upon as lineally descended from languages which can be theoretically reconstructed in their main phonetic and structural outlines. It is obvious, however, that languages may so diverge as to leave little trace of their original relationship. It is therefore very dangerous to assume that languages are not at last analysis divergent members of a single genetic group merely because the evidence is negative. The only contrast that is legitimate is between languages known to be historically related and languages not known to be so related. Languages known to be related cannot be legitimately contrasted with languages known not to be related.
Because of the fact that languages have differentiated at different rates and because of the important effects of cultural diffusion, which have brought it about that strategically placed languages, such as Arabic, Latin and English, have spread over large parts of the earth at the expense of others, very varied conditions are found to prevail in regard to the distribution of linguistic families. In Europe, for instance, there are only two linguistic families of importance represented today, the Indo-European languages and the Ugro-Finnic languages, of which Finnish and Hungarian are examples. The Basque dialects of southern France and northern Spain are the survivors of another and apparently isolated group. On the other hand, in aboriginal America the linguistic differentiation is extreme and a surprisingly large number of essentially unrelated linguistic families must be recognized. Some of the families occupy very small areas, while others, such as the Algonquin and the Athabaskan languages of North America, are spread over a large territory. The technique of establishing linguistic families and of working out the precise relationship of the languages included in these families is too difficult to be gone into here. It suffices to say that random word comparisons are of little importance. Experience shows that very precise phonetic relations can be worked out between the languages of a group and that on the whole fundamental morphological features tend to preserve themselves over exceedingly long periods of time. Thus modern Lithuanian is in structure, vocabulary and, to a large extent, even phonemic pattern very much the kind of a language which must be assumed as the prototype for the Indo-European languages as a whole. In spite of the fact that structural classifications are in theory unrelated to genetic ones and in spite of the fact that languages can be shown to have influenced each other, not only in phonetics and vocabulary but also to an appreciable extent in. structure, it is not often found that the languages of a genetic group exhibit utterly irreconcilable structures. Thus even English, which is one of the least conservative of Indo-European languages, has many far reaching points of structure in common with as remote a language as Sanskrit in contrast, say, to Basque or Finnish. Again, different as are Assyrian, modern -Arabic and the Semitic languages of Abyssinia they exhibit numerous points of resemblance in phonetics, vocabulary and structure which set them off at once from, say, Turkish or the Negro languages of the Nile headwaters.
The complete rationale of linguistic change, involving as it does many of the most complex processes of psychology and sociology, has not yet been satisfactorily worked out, but there are a number of general processes that emerge with sufficient clarity. For practical purposes inherent changes may be distinguished from changes due to contact with other linguistic communities. There can be no hard line of division between these two groups of changes because every individual's language is a distinct psychological entity in itself, so that all inherent changes are likely at last analysis to be peculiarly remote or subtle forms of change due to contact. The distinction, however, has great practical value, all the more so as there is a tendency among anthropologists and sociologists to operate far too hastily with wholesale linguistic changes due to external ethnic and cultural influences. The enormous amount of study that has been lavished on the history of particular languages and groups of languages shows very clearly that the most powerful differentiating factors are not outside influences, as ordinarily understood, but rather the very slow but powerful unconscious changes in certain directions which seem to be implicit in the phonemic systems and. morphologies of the languages themselves. These "drifts" are powerfully conditioned by unconscious formal feelings and are made necessary by the inability of human beings to actualize ideal patterns in a permanently set fashion.
Linguistic changes may be analyzed into pho-
( 164) -netic changes, changes in form and changes in vocabulary. Of these the phonetic changes seem to be the most important and the most removed from direct observation. The factors which lead to these phonetic changes are probably exceedingly complex and no doubt include the operation of obscure symbolisms which define the relation of various age groups to one another. Not all phonetic changes, however, can be explained in terms of social symbolism. It seems that many of them are due to the operation of unconscious economies in actualizing sounds or combinations of sounds. The most impressive thing about internal phonetic change is its high degree of regularity. It is this regularity, whatever its ultimate cause, that is more responsible than any other single factor for the enviable degree of exactness which linguistics has attained as a historical discipline. Changes in grammatical form often follow in the wake of destructive phonetic changes. In many cases it can be seen how irregularities produced by the disintegrating effect of phonetic change are ironed out by the analogical spread of more regular forms. The cumulative effect of these corrective changes is quite sensibly to modify the structure of the language in many details and sometimes even in its fundamental features. Changes in vocabulary are due to a great variety of causes, most of which are of a cultural rather than of a strictly linguistic nature. The too frequent use of a word, for instance, may reduce it to a commonplace term, so that it needs to be replaced by a new word. On the other hand, changes of attitude may make certain words with their traditional overtones of meaning unacceptable to the younger generation, so that they tend to become obsolete. Probably the most important single source of change in vocabulary is the creation of new words on analogies which have spread from a few specific words.
Of the linguistic changes due to the more obvious types of contact the one which seems to have played the most important part in the history of language is the "borrowing" of words across linguistic frontiers. This borrowing naturally goes hand in hand with cultural diffusion. An analysis of the provenience of the words of a given language is frequently an important index of the direction of cultural influence. Our English vocabulary, for instance, is very richly stratified in a cultural sense. The various layers of early Latin, mediaeval French, humanistic Latin and Greek and modern French borrowings constitute a fairly accurate gauge of the time, extent and nature of the various foreign cultural influences which have helped to mold English civilization. The notable lack of German loan words in English until a very recent period, as contrasted with the large number of Italian words which were adopted at the time of the Renaissance and later, is again a historically significant fact. By the diffusion of culturally important words, such as those referring to art, literature, the church, military affairs, sport and business, there have grown up important transnational vocabularies which do something to combat the isolating effect of the large number of languages which are still spoken in the modern world. Such borrowings have taken place in all directions, but the number of truly important source languages is surprisingly, small. Among the more important of them are Chinese, which has saturated the vocabularies of Korean, Japanese and Annamite; Sanskrit, whose influence on the cultural vocabulary of central Asia, India and Indo-China has been enormous; Arabic, Greek, Latin and French. English, Spanish and Italian have also been of great importance as agencies of cultural transmission, but their influence seems less far reaching than that of the languages mentioned above. The cultural influence of a language is not always in direct proportion to its intrinsic literary interest or to the cultural place which its speakers have held in the history of the world. For example, while Hebrew is the carrier of a peculiarly significant culture, actually it has not had as important an influence on other languages of Asia as Aramaic, a sister language of the Semitic stock.
The phonetic influence exerted by a foreign language may be very considerable, and there is a great deal of evidence to show that dialectic peculiarities have often originated as a result of the unconscious transfer of phonetic habits from the language in which one was brought up to that which has been adopted later in life. Apart, however, from such complete changes in speech is the remarkable fact that distinctive phonetic features tend to be distributed over wide areas regardless of the vocabularies acid structures of the languages involved. One of the most striking examples of this type of distribution is found among the Indian languages of the Pacific coast of California, Oregon, Washington, British Columbia and southern Alaska. Here are a large number of absolutely distinct languages, belonging to a number of genetically unrelated stocks, so far as we are able to tell, which nevertheless have many important and distinctive
( 165) phonetic features in common. An analogous fact is the distribution of certain peculiar phonetic features in both the Slavic languages and the Ugro-Finnic languages, which are unrelated to them. Such processes of phonetic diffusion must be due to the influence exerted by bilingual speakers, who act as unconscious agents for the spread of phonetic habits over wide areas. Primitive man is not isolated, and bilingualism is probably as important a factor in the contact of primitive groups as it is on more sophisticated levels.
Opinions differ as to the importance of the purely morphological influence exerted by one language on another in contrast with the more external types of phonetic and lexical influence. Undoubtedly such influences must be taken into account, but so far they have not been shown to operate on any great scale. In spite of the centuries of contact, for instance, between Semitic and Indo-European languages we know of no language which is definitely a blend of the structures of these two stocks. Similarly, while Japanese is flooded with Chinese loan words, there seems to be no structural influence of the latter on the former. A type of influence which is neither one of vocabulary nor of linguistic form, in the ordinary sense of the word, and to which insufficient attention has so far been called, is that of meaning pattern. It is a remarkable fact of modern European culture, for instance, that while the actual terms used for certain ideas vary enormously from language to language, the range of significance of these equivalent terms tends to be very similar, so that to a large extent the vocabulary of one language tends to be a psychological and cultural translation of the vocabulary of another. A simple example of this sort would be the translation of such terms as Your Excellency to equivalent but etymologically unrelated terms in Russian. Another instance of this kind would be the interesting parallelism in nomenclature between the kinship terms of affinity in English, French and German. Such terms as mother-in-law, belle-mère and Schwiegermutter are not, strictly speaking, equivalent either as to etymology or literal meaning but they are patterned in exactly the same manner. Thus mother-in-law and father-in-law are parallel in nomenclature to belle-mère and beau-père. and to Schwiegermutter and Schwiegervater. These terms clearly illustrate the diffusion of a lexical pattern which in turn probably expresses a growing feeling of the sentimental equivalence of blood relatives and relatives by marriage.
The importance of language as a whole for the definition, expression and transmission of culture is undoubted. The relevance of linguistic details, in both content and form, for the profounder understanding of culture is also clear. It does not follow, however, that there is a simple correspondence between the form of a language and the form of the culture of those who speak it. The tendency to sere linguistic categories as directly expressive of overt cultural outlines, which seems to have come into fashion among certain sociologists and anthropologists, should be resisted as in no way warranted by the actual facts. There is no general correlation between cultural type and linguistic structure. So far as can be seen, isolating or agglutinative or inflective types of speech are possible on any level of civilization. Nor does the presence or absence of grammatical gender, for example, seem to have any relevance for the understanding of the social organization or religion or folklore of the associated peoples. If there were any such parallelism as has sometimes been maintained, it would be quite impossible to understand the rapidity with which culture diffuses in spite of profound linguistic differences between the borrowing and giving communities. The cultural significance of linguistic form, in other words, lies on a much more submerged level than on the overt one of definite cultural pattern. It is only very rarely, as a matter of fact, that it can be pointed out how a cultural trait has had some influence on the fundamental structure of a language. To a certain extent this lack of correspondence may be due to the fact that linguistic changes do not proceed pat the same rate as most cultural changes, which are on the whole far more rapid. Short of yielding to another language which takes its place, linguistic organization, largely because it is unconscious, tends to maintain itself indefinitely and does not allow its fundamental formal categories to be seriously influenced by changing cultural needs. If the forms of culture and language were then in complete correspondence with one another, the nature of the processes making for linguistic and cultural changes respectively .would soon bring about a lack of necessary correspondence. This is exactly what is found to b-e the case. Logically it is, indefensible that the masculine, feminine and neuter genders of German and Russian should be allowed to continue their sway in the modem world; but any intellectualist attempt to weed out these unnecessary genders would obviously be fruitless, for the
( 166) normal speaker does not actually feel the clash which the logician requires.
It is another matter when we pass from general form to the detailed content of a language. Vocabulary is a very sensitive index of the culture of a people and changes of meaning, loss of old words, the creation and borrowing of new ones are all dependent on the history of culture itself. Languages differ widely in the nature of their vocabularies. Distinctions which seem inevitable to us may be utterly ignored in languages which reflect an entirely different type of culture, while these in turn insist on distinctions which are all but unintelligible to us. Such differences of vocabulary go far beyond the names of cultural objects, such as arrow point, coat of armor or gunboat. They apply just as well to the mental world. It would be difficult in some languages, for instance, to express the distinction which we feel between "to kill" and "to murder" for the simple reason that the underlying legal philosophy which determines our use of these words does not seem natural to all societies. Abstract terms, which are so necessary to our thinking, may be infrequent in a language whose speakers formulate their behavior on more pragmatic lines. On the other hand, the question of the presence or absence of abstract nouns may be bound up with the fundamental form of the language; and there exist a large number of primitive languages whose structure allows of the very ready creation and use of abstract nouns of quality or action.
There are many language patterns of a special sort which are of interest to the social scientist. One of these is the tendency to create tabus for certain words or names. A very widespread custom among primitive peoples, for instance, is the tabu which is placed not only on the use of the name of a person recently deceased but of any word that is etymologically connected in the feeling of the speakers with such a name. This means that ideas have often to be expressed by circumlocutions or that terms must be borrowed from neighboring dialects. Sometimes certain names or words are too holy to be pronounced except under very special conditions, and curious patterns of behavior develop which are designed to prevent one from making use of such interdicted terms. An example of this is the Jewish custom of pronouncing the Hebrew name for God, not as Yahwe or Jehovah but as Adonai, My Lord. Such customs seem strange to us but equally strange to many primitive communities would be our extraordinary reluctance to pronounce obscene words under normal :social conditions. Another class of special linguistic phenomena is the use of esoteric language devices, such as passwords or technical terminologies for ceremonial attitudes or practises. Among the Eskimo, for example, the medicine man has a peculiar vocabulary which is not understood by those who are not members of his guild. Special dialectic forms or otherwise peculiar linguistic patterns are common among primitive peoples for the texts of songs. Sometimes, :as in Melanesia, such song texts are due to the influence of neighboring dialects. This is strangely analogous to the practise among ourselves of singing songs in Italian, French or German rather than in English, and it is likely that the historical processes which have led to the parallel custom are of a similar nature. Thieves' jargons and secret languages of children may also be mentioned. These lead over into special sign and gesture languages, many of which are based directly on spoken or written speech; they seem to exist on many levels of culture. The sign language of the Plains Indians of North America arose in response to the need for some medium of communication between tribes speaking mutually unintelligible languages. Within the Christian church may be noted the elaboration of gesture languages by orders of monks vowed to silence. Not only a language or a terminology but the mere external form in which it is. written may become important as a symbol of sentimental or social distinction. Thus Croatian and Serbian are essentially the same language but they are presented in very different outward forms, the former being written in Latin characters, the latter in the Cyrillic character of the Greek Orthodox church. This external difference, associated with a difference of religion, has of course the important function of preventing people who speak closely related languages or dialects but who wish for reasons of sentiment not to confound themselves in a larger unity from becoming too keenly aware of how much they actually resemble each other.
The relation of language to rationalism and internationalism presents a number of interesting sociological problems. Anthropology makes a rigid distinction between ethnic units based on race, on culture and on language. It points out that these do not need to coincide in the least --that they do not, as a matter of fact, often coincide in reality. But with the increased emphasis on nationalism in modern times the question of the symbolic meaning of race and lan-
( 167) -guage has taken on a new significance and, whatever the scientist may say, the layman is ever inclined to see culture, language and race as but different facets of a single social unity, which he tends in turn to identify with such a political entity as England or France or Germany. To point out, as the anthropologist easily can, that cultural distributions and nationalities override language and race groups does not end the matter for the sociologist, because he feels that the concept of nation or nationality must be integrally imaged by the non-analytical person as carrying with it the connotation, real or supposed, of both race and language. From this standpoint it really makes little difference whether history and anthropology support the popular identification of nationality, language and race. The important thing to hold on to is that a particular language tends to become the fitting expression of a self-conscious nationality and that such a group will construct for itself in spite of all that the physical anthropologist can do a race to which is to be attributed the mystic power of creating a language and a culture as twin expressions of its psychic peculiarities.
So far as language and race are concerned, it is true that the major races of man have tended in the past to be set off against each other by important differences of language. There is less point to this, however, than might be imagined, because the linguistic differentiations within any given race are just as far reaching as those which can be pointed out across racial lines, yet they do not at all correspond to subracial units. Even the major races are not always clearly sundered by language. This is notably the case with the Malayo-Polynesian languages, which are spoken by peoples as racially distinct as the Malays, the Polynesians and the Negroes of Melanesia. Not one of the great languages of modern man follows racial lines. French, for example, is spoken by a highly mixed population, which is largely Nordic in the north, Alpine in the center and Mediterranean in the south, each of these subraces being liberally represented in the rest of Europe.
While language differences have always been important symbols of cultural difference, it is only in comparatively recent times, with the exaggerated development of the ideal of the sovereign nation and with the resulting eagerness to discover linguistic symbols for this ideal of sovereignty, that language differences have taken on an implication of antagonism. In ancient Rome and all through mediaeval Europe there were plenty of cultural differences running side by side with linguistic ones, and the political status of Roman citizen or the fact of adherence to the Roman Catholic church was of vastly greater significance as a symbol of the individual's place in the world than the language or dialect which he happened to speak. It is probably altogether incorrect to maintain that language differences are responsible for national antagonisms. It would seem to be much more reasonable to suppose that a political and national unit, once definitely formed, uses a prevailing language as a symbol of it-, identity, whence gradually emerges the peculiarly modern feeling that every language should properly be the expression of a distinctive nationality. In earlier times there seems to have been little systematic attempt to impose the language of a conquering people on the subject people, although it happened frequently as a result of the processes implicit in the spread of culture that such a conqueror's language was gradually taken over by the dispossessed population. Witness the spread of the Romance languages and of the modern Arabic dialects. On the other hand, it seems to have happened about as frequently that the conquering group was culturally and linguistically absorbed and that their own language disappeared without necessary danger to their privileged status. Thus foreign dynasties in China have always submitted to the superior culture of the Chinese and have taken on their language. In the same way the Moslem Moguls of India, while true to their religion, which was adopted by millions in northern India, made one of the Hindu vernaculars the basis of the great literary language of Moslem India, Hindustani. Definitely repressive attitudes toward the languages and dialects of subject peoples seem to be distinctive only of European political policy in comparatively recent times. The attempt of czarist Russia to stamp oust Polish by forbidding its teaching in the schools and the similarly repressive policy of contemporary Italy in its attempt to wipe out German from the territory recently acquired from Austria are illuminating examples of the heightened emphasis on language as a symbol of political allegiance in the modern world.
To match these repressive measures there is the oft repeated attempt of minority groups to erect their language into the status of a fully accredited medium of cultural and literary expression. Many of these restored or semimanu-
(168) -factured languages have come in on the wave of resistance to exterior political or cultural hostility. Such are the Gaelic of Ireland, the Lithuanian of a recently created republic and the Hebrew of the Zionists. In other cases such languages have come in more peacefully because of a sentimental interest in local culture. Such are the modern Provençal of southern France, the Plattdeutsch of northern Germany, Frisian and the Norwegian landsmaal. It is doubtful whether these persistent attempts to make true culture languages of local dialects that have long ceased to be of primary literary importance can succeed in the long run. The failure of modern Provençal to hold its own and the very dubious success of Gaelic make it seem probable that following the recent tendency to resurrect minor languages will come a renewed leveling of speech more suitably expressing the internationalism which is slowly emerging.
The logical necessity of an international language in modern times is in strange contrast to the indifference and even opposition with which most people consider its possibility. The attempts so far made to solve this problem, of which Esperanto has probably had the greatest measure of practical success, have not affected more than a very small proportion of the people whose international interests and needs might have led to a desire for a simple and uniform means of international expression, at least for certain purposes. It is in the less important countries of Europe, such as Czechoslovakia, that Esperanto has been moderately successful, and for obvious reasons. The opposition to an international language has little logic or psychology in its favor. The supposed artificiality of such a language as Esperanto or of any of the equivalent languages that have been proposed has been absurdly exaggerated, for in sober truth there is practically nothing in these languages that is not taken from the common stock of words and forms which have gradually developed in Europe. Such an international language could of course have only the status of a secondary form of speech for distinctly limited purposes. Thus considered the learning of a constructed international language offers no further psychological problem than the learning of any other language which is acquired after childhood through the medium of books and with the conscious application of grammatical rules. The lack of interest in the international language problem in spite of the manifest need for one is an excellent example of how little logic or intellectual necessity has to do with the acquirement of language habits. Even the acquiring of the barest smattering of a foreign national language is imaginatively equivalent to some measure of identification with a people or a culture. The purely instrumental value of such knowledge is frequently nil. Any consciously constructed international language has to deal with the great difficulty of not being felt to represent a distinctive people or culture. Hence the learning of it is of very little symbolic significance for the average person, who remains blind to the fact that such a language., easy and regular as it inevitably must be, would solve many of his educational and practical difficulties at a single blow. The future alone will tell whether the logical advantages and theoretical necessity of an international language can overcome the largely symbolic opposition which it has to meet. In any event it is at least conceivable that one of the great national languages of modern times, such as English or Spanish or Russian, may in due course find itself in the position of a de facto international language without any conscious attempt having been made to put it there.