Language: An Introduction to the Study of Speech

by Edward Sapir

When it comes to the actual task of classification, we find that we have no easy road to travel. Various classifications have been suggested, and they all contain elements of value. Yet none proves satisfactory. They do not so much enfold the known languages in their embrace as force them down into narrow, straight-backed seats. The difficulties have been of various kinds. First and foremost, it has been difficult to choose a point of view. On what basis shall we classify? A language shows us so many facets that we may well be puzzled. And is one point of view sufficient? Secondly, it is dangerous to generalize from a small number of selected languages. To take, as the sum total of our material, Latin, Arabic, Turkish, Chinese, and perhaps Eskimo or Sioux as an afterthought, is to court disaster. We have no right to assume that a sprinkling of exotic types will do to supplement the few languages nearer home that we are more immediately interested in. Thirdly, the strong craving for a simple formula[92] has been the undoing of linguists. There is something irresistible about a method of classification that starts with two poles, exemplified, say, by Chinese and Latin, clusters what it conveniently can about these poles, and throws everything else into a “transitional type.” Hence has arisen the still popular classification of languages into an “isolating” group, an “agglutinative” group, and an “inflective” group. Sometimes the languages of the American Indians are made to straggle along as an uncomfortable “polysynthetic” rear-guard to the agglutinative languages. There is justification for the use of all of these terms, though not perhaps in quite the spirit in which they are commonly employed. In any case it is very difficult to assign all known languages to one or other of these groups, the more so as they are not mutually exclusive. A language may be both agglutinative and inflective, or inflective and polysynthetic, or even polysynthetic and isolating, as we shall see a little later on.

There is a fourth reason why the classification of languages has generally proved a fruitless undertaking. It is probably the most powerful deterrent of all to clear thinking. This is the evolutionary prejudice which instilled itself into the social sciences towards the middle of the last century and which is only now beginning to abate its tyrannical hold on our mind. Intermingled with this scientific prejudice and largely anticipating it was another, a more human one. The vast majority of linguistic theorists themselves spoke languages of a certain type, of which the most fully developed varieties were the Latin and Greek that they had learned in their childhood. It was not difficult for them to be persuaded that these familiar languages represented the “highest” development that speech had yet attained and that all other types were but steps on the way to this beloved “inflective” type. Whatever conformed to the pattern of Sanskrit and Greek and Latin and German was accepted as expressive of the “highest,” whatever departed from it was frowned upon as a shortcoming or was at best an interesting aberration.[93] Now any classification that starts with preconceived values or that works up to sentimental satisfactions is self-condemned as unscientific. A linguist that insists on talking about the Latin type of morphology as though it were necessarily the high-water mark of linguistic development is like the zoölogist that sees in the organic world a huge conspiracy to evolve the race-horse or the Jersey cow. Language in its fundamental forms is the symbolic expression of human intuitions. These may shape themselves in a hundred ways, regardless of the material advancement or backwardness of the people that handle the forms, of which, it need hardly be said, they are in the main unconscious. If, therefore, we wish to understand language in its true inwardness we must disabuse our minds of preferred “values”[94] and accustom ourselves to look upon English and Hottentot with the same cool, yet interested, detachment.

We come back to our first difficulty. What point of view shall we adopt for our classification? After all that we have said about grammatical form in the preceding chapter, it is clear that we cannot now make the distinction between form languages and formless languages that used to appeal to some of the older writers. Every language can and must express the fundamental syntactic relations even though there is not a single affix to be found in its vocabulary. We conclude that every language is a form language. Aside from the expression of pure relation a language may, of course, be “formless”—formless, that is, in the mechanical and rather superficial sense that it is not encumbered by the use of non-radical elements. The attempt has sometimes been made to formulate a distinction on the basis of “inner form.” Chinese, for instance, has no formal elements pure and simple, no “outer form,” but it evidences a keen sense of relations, of the difference between subject and object, attribute and predicate, and so on. In other words, it has an “inner form” in the same sense in which Latin possesses it, though it is outwardly “formless” where Latin is outwardly “formal.” On the other hand, there are supposed to be languages[95] which have no true grasp of the fundamental relations but content themselves with the more or less minute expression of material ideas, sometimes with an exuberant display of “outer form,” leaving the pure relations to be merely inferred from the context. I am strongly inclined to believe that this supposed “inner formlessness” of certain languages is an illusion. It may well be that in these languages the relations are not expressed in as immaterial a way as in Chinese or even as in Latin,[96] or that the principle of order is subject to greater fluctuations than in Chinese, or that a tendency to complex derivations relieves the language of the necessity of expressing certain relations as explicitly as a more analytic language would have them expressed.[97] All this does not mean that the languages in question have not a true feeling for the fundamental relations. We shall therefore not be able to use the notion of “inner formlessness,” except in the greatly modified sense that syntactic relations may be fused with notions of another order. To this criterion of classification we shall have to return a little later.

More justifiable would be a classification according to the formal processes[98] most typically developed in the language. Those languages that always identify the word with the radical element would be set off as an “isolating” group against such as either affix modifying elements (affixing languages) or possess the power to change the significance of the radical element by internal changes (reduplication; vocalic and consonantal change; changes in quantity, stress, and pitch). The latter type might be not inaptly termed “symbolic” languages.[99] The affixing languages would naturally subdivide themselves into such as are prevailingly prefixing, like Bantu or Tlingit, and such as are mainly or entirely suffixing, like Eskimo or Algonkin or Latin. There are two serious difficulties with this fourfold classification (isolating, prefixing, suffixing, symbolic). In the first place, most languages fall into more than one of these groups. The Semitic languages, for instance, are prefixing, suffixing, and symbolic at one and the same time. In the second place, the classification in its bare form is superficial. It would throw together languages that differ utterly in spirit merely because of a certain external formal resemblance. There is clearly a world of difference between a prefixing language like Cambodgian, which limits itself, so far as its prefixes (and infixes) are concerned, to the expression of derivational concepts, and the Bantu languages, in which the prefixed elements have a far-reaching significance as symbols of syntactic relations. The classification has much greater value if it is taken to refer to the expression of relational concepts[100] alone. In this modified form we shall return to it as a subsidiary criterion. We shall find that the terms “isolating,” “affixing,” and “symbolic” have a real value. But instead of distinguishing between prefixing and suffixing languages, we shall find that it is of superior interest to make another distinction, one that is based on the relative firmness with which the affixed elements are united with the core of the word.[101]

There is another very useful set of distinctions that can be made, but these too must not be applied exclusively, or our classification will again be superficial. I refer to the notions of “analytic,” “synthetic,” and “polysynthetic.” The terms explain themselves. An analytic language is one that either does not combine concepts into single words at all (Chinese) or does so economically (English, French). In an analytic language the sentence is always of prime importance, the word is of minor interest. In a synthetic language (Latin, Arabic, Finnish) the concepts cluster more thickly, the words are more richly chambered, but there is a tendency, on the whole, to keep the range of concrete significance in the single word down to a moderate compass. A polysynthetic language, as its name implies, is more than ordinarily synthetic. The elaboration of the word is extreme. Concepts which we should never dream of treating in a subordinate fashion are symbolized by derivational affixes or “symbolic” changes in the radical element, while the more abstract notions, including the syntactic relations, may also be conveyed by the word. A polysynthetic language illustrates no principles that are not already exemplified in the more familiar synthetic languages. It is related to them very much as a synthetic language is related to our own analytic English.[102] The three terms are purely quantitative—and relative, that is, a language may be “analytic” from one standpoint, “synthetic” from another. I believe the terms are more useful in defining certain drifts than as absolute counters. It is often illuminating to point out that a language has been becoming more and more analytic in the course of its history or that it shows signs of having crystallized from a simple analytic base into a highly synthetic form.[103]

We now come to the difference between an “inflective” and an “agglutinative” language. As I have already remarked, the distinction is a useful, even a necessary, one, but it has been generally obscured by a number of irrelevancies and by the unavailing effort to make the terms cover all languages that are not, like Chinese, of a definitely isolating cast. The meaning that we had best assign to the term “inflective” can be gained by considering very briefly what are some of the basic features of Latin and Greek that have been looked upon as peculiar to the inflective languages. First of all, they are synthetic rather than analytic. This does not help us much. Relatively to many another language that resembles them in broad structural respects, Latin and Greek are not notably synthetic; on the other hand, their modern descendants, Italian and Modern Greek, while far more analytic[104] than they, have not departed so widely in structural outlines as to warrant their being put in a distinct major group. An inflective language, we must insist, may be analytic, synthetic, or polysynthetic.

Latin and Greek are mainly affixing in their method, with the emphasis heavily on suffixing. The agglutinative languages are just as typically affixing as they, some among them favoring prefixes, others running to the use of suffixes. Affixing alone does not define inflection. Possibly everything depends on just what kind of affixing we have to deal with. If we compare our English words farmer and goodness with such words as height and depth, we cannot fail to be struck by a notable difference in the affixing technique of the two sets. The -er and -ness are affixed quite mechanically to radical elements which are at the same time independent words (farm, good). They are in no sense independently significant elements, but they convey their meaning (agentive, abstract quality) with unfailing directness. Their use is simple and regular and we should have no difficulty in appending them to any verb or to any adjective, however recent in origin. From a verb to camouflage we may form the noun camouflager “one who camouflages,” from an adjective jazzy proceeds with perfect case the noun jazziness. It is different with height and depth. Functionally they are related to high and deep precisely as is goodness to good, but the degree of coalescence between radical element and affix is greater. Radical element and affix, while measurably distinct, cannot be torn apart quite so readily as could the good and -ness of goodness. The -t of height is not the typical form of the affix (compare strength, length, filth, breadth, youth), while dep- is not identical with deep. We may designate the two types of affixing as “fusing” and “juxtaposing.” The juxtaposing technique we may call an “agglutinative” one, if we like.

Is the fusing technique thereby set off as the essence of inflection? I am afraid that we have not yet reached our goal. If our language were crammed full of coalescences of the type of depth, but if, on the other hand, it used the plural independently of verb concord (e.g., the books falls like the book falls, or the book fall like the books fall), the personal endings independently of tense (e.g., the book fells like the book falls, or the book fall like the book fell), and the pronouns independently of case (e.g., I see he like he sees me, or him see the man like the man sees him), we should hesitate to describe it as inflective. The mere fact of fusion does not seem to satisfy us as a clear indication of the inflective process. There are, indeed, a large number of languages that fuse radical element and affix in as complete and intricate a fashion as one could hope to find anywhere without thereby giving signs of that particular kind of formalism that marks off such languages as Latin and Greek as inflective.

What is true of fusion is equally true of the “symbolic” processes.[105] There are linguists that speak of alternations like drink and drank as though they represented the high-water mark of inflection, a kind of spiritualized essence of pure inflective form. In such Greek forms, nevertheless, as pepomph-a “I have sent,” as contrasted with pemp-o “I send,” with its trebly symbolic change of the radical element (reduplicating pe-, change of e to o, change of p to ph), it is rather the peculiar alternation of the first person singular -a of the perfect with the -o of the present that gives them their inflective cast. Nothing could be more erroneous than to imagine that symbolic changes of the radical element, even for the expression of such abstract concepts as those of number and tense, is always associated with the syntactic peculiarities of an inflective language. If by an “agglutinative” language we mean one that affixes according to the juxtaposing technique, then we can only say that there are hundreds of fusing and symbolic languages—non-agglutinative by definition—that are, for all that, quite alien in spirit to the inflective type of Latin and Greek. We can call such languages inflective, if we like, but we must then be prepared to revise radically our notion of inflective form.

It is necessary to understand that fusion of the radical element and the affix may be taken in a broader psychological sense than I have yet indicated. If every noun plural in English were of the type of book: books, if there were not such conflicting patterns as deer: deer, ox: oxen, goose: geese to complicate the general form picture of plurality, there is little doubt that the fusion of the elements book and -s into the unified word books would be felt as a little less complete than it actually is. One reasons, or feels, unconsciously about the matter somewhat as follows:—If the form pattern represented by the word books is identical, as far as use is concerned, with that of the word oxen, the pluralizing elements -s and -en cannot have quite so definite, quite so autonomous, a value as we might at first be inclined to suppose. They are plural elements only in so far as plurality is predicated of certain selected concepts. The words books and oxen are therefore a little other than mechanical combinations of the symbol of a thing (book, ox) and a clear symbol of plurality. There is a slight psychological uncertainty or haze about the juncture in book-s and ox-en. A little of the force of -s and -en is anticipated by, or appropriated by, the words book and ox themselves, just as the conceptual force of -th in dep-th is appreciably weaker than that of -ness in good-ness in spite of the functional parallelism between depth and goodness. Where there is uncertainty about the juncture, where the affixed element cannot rightly claim to possess its full share of significance, the unity of the complete word is more strongly emphasized. The mind must rest on something. If it cannot linger on the constituent elements, it hastens all the more eagerly to the acceptance of the word as a whole. A word like goodness illustrates “agglutination,” books “regular fusion,” depth “irregular fusion,” geese “symbolic fusion” or “symbolism.”[106]