Besides the division of languages into families bearing traces of a common origin. there is a division into three orders, as they may be called, depending upon a radical differ-, ence of structure. Speech, as the expression of thought, contains two elements; ideas or conceptions, which constitute the substance or material part; and the relations of these ideas to one another, which constitute the formal part; and the nature of a language depends upon the particular way in which the vocal expression of these two elements is combined. At the foundation of all words lie roots (q.v.), or simple sounds expressive of meaning. Now, some languages, as the Chinese (q.v.), use these roots in their naked form as words, the same syllable, according to its position, serving as noun, adjective, verb, ete.—e. g., to mean m great, greatness, to lie or to make great, greatly or very. The relational part of the thought, for the most part, gets no vocal expression, it is only indicated by position, as when min, people, and power, are simply put together (min to signify the people's power. Relations not readily indicated by position are expressed in a round-about way by using additional significant words: thus, Ockung (mass or multitude) fin (man) = men; au (woman) (child) = daughter; y nun (employ people power) = with the people's power. Even in such cases each root p:e serves its independence, and i, felt to express its own radical meaning. Languages like the Chinese, whose development has been arrested at this rudimentary stage, arc called monosyllable or isolating.
The next stage of development is that of the agglutinate languages, wh id) arc by far the most numerous, including tne Turanian and American families.' In these the relational part of thought obtains prominent vocal expression by separate roots joined Or glued on to the significaut roots as terininations. These termtnations were originally themselves signi11 .ant roots, and many of them are still used as separate significant words, although the greater part have sunk dawn to mere signs of cases and other relations. The com pound expression thus formed never,• however. attains perfect unity; the significant root al\l'a vs remains rigid, unobscured in its sense and unehaeged in form, and the termina tion is felt ;is somthing distinct from the body of the word.
Thus, the Finnish declension exhibits a structure of the most mechanical and trans parent kind—c. g.. ka•lm, bear; Icarhu-n, of the bear; karhut-tit, without bear; karltu sta, out of the bear; and so on through fifteen cases, The insertion of the plural suffix, i, gives of the bears; Ica rhu-i-ta, without bears; ka•hu-i-sta, out of the beats; etc. But this composite mechanical structure reaches its climax—remaining all the while veleetly transparent—in the Turkish verb. Thus, the root see has the indefinite .
meaning of loving, and the inf, is sec-mek, to love; which then, by the insertion of cer tain sn.fixes, can take on as many as forty forms or voices—e. g., see-ins-Mel% not to love; x.e:-.--ine-llick, not to he able to love; see-db.-Inch, to cause to love; sev-air-ish-mck, to cause one another to love; see-il-mek, to be loved; .set-il-e-anc-m eh, not to be able to he loved; etc. Eicii of these forms, then, runs through a large round of tenses and moods, with their persons and nunthers.
The languages of the American Indians are all of this agglutinating type, although they have also got the name incorporative, or intercalative, because they run a whole phrase or sentence into one word—e. g., &Toni, to wash; hopocuni, to wash hands; le9p.•,a.luai, to wash feet; ninacaqua, I (ni) eat (qua) flesh (wen). The Basque language partakes of this character.
It is only in the third or inflectional stage that perfect unity of the two elements is attained. In the Aryan and Semitic tongues, which alone have reached this Itighest state of chwelopment, the significant root and the termination have become blended into one both in effect and form, and phonetic changes have for the most part obliterated the traces of composition. Vet no doubt is felt by philologists that the most highly organ ized of the inflecting or amalgamating languages began with the radical stage, and pas ted througli the agglutinate. The analytic powers of comparative grammar have sec:eye:ling in tracing back the formal elements of the Aryan tongues to original inde pendent word.!, :agglutinated to other words to modify them. See INFIECTION. Against this theory it has been urged, that there is no historical instance of a language so chang ing its type, and passing from one stage to another. But a sufficient account of this phenomenon may lie found in the different mental habits and political positions of the peoples (see Max ?biller, Lectures on the SCie 17 CC of Language, first series, page 316). Besides, the languages of the lower types do show a tendency, under favorable circum stances, to produce grammatical forms of the higher kind, Even in Chinese, in some of its ma legal dialects, something like cases is to be seen; and Finnish and Turkish, in contact with the inflected languages of Europe, are making approaches to the inflec tional type.
On the the other hand the inflectional languages lad, before the earliest times of which we any written monuments, entered on die reverse phase—tlie analytic. By the peocess of phonetic change and decay, the grammatical forms have been gradually becoming, obliterated and losing their power, end their place has been supplied by sepa rate words, in the shape of prepositions and auxiliary verbs. See INFLEcTrox.