Languages, Classification of

(1) The genetic classification of languages is based on relationship—that is, common origin.

The relationship of some languages is considered proved if a common origin is revealed for a significant portion of the morphemes of the languages, all the grammatical affixes (if they exist), and many of the roots (including the areas of the vocabulary that are usually distinguished by a special stability—that is, pronouns, names for certain parts of the body, and words meaning “water,” “fire,” “sun,” “to be,” “to give,” “to eat,” “to drink,” and so on). A common origin for roots and affixes is confirmed by the presence of regular interlingual phonetic correspondences. If it is possible to create a comparative historical phonetics that allows the roots of the parent language to be approximately reconstructed and their transformation into the roots of the daughter language to be traced (according to strict rules), the relationship of the latter is established. In this sense the existence of several language families in the Old World is indisputable: Indo-European, Uralic (with Finno-Ugric and Samoyed subfamilies), Turkic, Mongolian, Manchu-Tungusic, Dravidian, Kartvelian, and Semito-Hamitic (Afro-Asiatic).

In the 1960’s an attempt was made to prove strictly the ancient relationship between the above-mentioned eight language families, which were combined into the Nostratic linguistic family. A comparative phonetics of the languages, tracing the regular phonetic correspondences in more than 600 roots and affixes, was successfully constructed. In addition, there is reason to assume the Nostratic origin of Yukaghir (possibly belonging to the Uralic group), Chukchi-Kamchadal, Nivkh (Gilyak), Korean, Japanese, perhaps Eskimo-Aleutian, and possibly also Elamite and Etruscan.

The position of the Abkhazo-Adygeian and Nakhian-Dagestanian language families of the Caucasus is unclear. Many linguists combine them with the Kartvelian languages into the Ibero-Caucasian linguistic family; however, the existence of the latter has not yet been proved (that is, regular sound correspondences have not been established, and few common Caucasian roots have been found). The genetic position of the Hurrian-Urartaean language family, which some linguists link with the Nostratic and others with the Nakhian-Dagestanian languages, remains debatable. Some researchers believe in the existence of the Altaic language family, uniting the Turkic, Mongolian, and Manchu-Tungusic languages and, in the opinion of many, the Korean and Japanese languages as well. The presence of a great number of common roots and regular sound correspondences in these languages is indisputable. However, for a definitive determination of the nature of the relationships between the Altaic languages, it is necessary to ascertain how great is the number of common Altaic roots and affixes in which the interlingual coincidences can be explained neither by borrowing nor by common Nostratic relationship.

The most important linguistic families of South and Southeast Asia are the Austroasiatic, Tai-Kadai, Miao-Yao, and Austrone-sian (Malayo-Polynesian) families, combined by some linguists into the hypothetical family of Austric languages. Among the Eurasian languages that do not fall into these hypothetical groupings are the Sino-Tibetan language family, the Yenisei and Andamanese families, and isolated languages, such as Basque, Burushaski, and Ainu, and ancient languages, such as Sumerian, Kassite, and Hattian. In eastern Indonesia there is the isolated North Halmahera linguistic family. Among the non-Austrone-sian languages of New Guinea and neighboring islands, collectively called the Papuan languages, there are, according to the latest data, 13 linguistic families and, moreover, many isolated languages. The Australian languages form a special family. The genetic placing of the almost unstudied extinct Tasmanian languages is unknown. All of the numerous African linguistic groups except for the Semito-Hamitic languages were combined by the American linguist J. Greenberg into three hypothetical families—Niger-Kordofan, including Bantu; Nilo-Saharan; and Khoisan. However, unification into these families, argued by lexical parallels, is still merely a plausible working hypothesis, since regular sound correspondences are not established.

(2) Typological classification of languages is based on morphological data that were collected independently of genetic or spatial proximity, relying exclusively on the properties of linguistic structure. Typological language classification seeks to include material from all the languages of the world, to reflect their similarities and differences, and to find the possible language types and specifics of each language or group of typologically similar languages. Modern typological language classification uses not only morphological but also phonological, syntactic, and semantic data.

The language type—that is, the characteristics of the basic properties of its structure—is the basis for including a language in a typological classification. The type, however, is not absolutely actualized in a language; in reality several types are represented in each language—that is, every language is polytypological. Therefore it is appropriate to say to what degree some type is present in the structure of a given language; on this basis attempts are made to give a quantitative interpretation of the typological characteristics of the language. The main problem in typological language classification is the formulation of (1) language descriptions that are consistent in terminology and based on a common concept of linguistic structure and (2) a system of nonconflicting and sufficient criteria for typological description.

The following typological language classification is the most widely accepted: the amorphous language type—invariable words with grammatical significance of word order, weak opposition of significant and syntactic roots (for example, ancient Chinese, Vietnamese, and Yoruba); the agglutinating (agglutinative) language type—a developed system of monosemous affixes, the absence of grammatical alternations in the root, the same type of inflection for all words belonging to one part of speech, a weak bond (the presence of distinct boundaries) between morphs (such as many Finno-Ugric languages, the Turkic languages, and Bantu); the inflectional type—including languages with internal inflection, that is, those with grammatically significant alternation in the root (the Semitic languages), and languages with inflectional endings and fusion, that is, those with simultaneous expression of several grammatical meanings by one affix (such as Russian rukami, “with the hands,” instrumental case, plural), a strong bond (the absence of distinct boundaries) between the morphs, and different types of declensions and conjugations (to some degree, Somali, Estonian, and the Nakhian [Chechen-Batsbi] languages); in the ancient and some modern Indo-European languages, internal inflection and fusion are combined.

A number of typologists also distinguish incorporating (poly-synthetic) languages, which have “word-sentences” and compound complexes—that is, the verbal form includes, sometimes in a clipped form, the nominal stems corresponding to the object, the circumstances, and the subject, as well as some grammatical markers (such as some American Indian languages, some Paleo-Asiatic, and the Caucasian languages). This typological classification of languages, which is basically a morphological classification, should not be considered definitive, primarily because it cannot reflect all the specifics of a given language by considering its structure. But it contains, in a form that is not immediately apparent, the possibility of making the classification as such more precise through analysis of other areas of language. For example, in amorphous languages such as classical Chinese, Vietnamese, and the languages of Guinea, monosyllabism of a word equivalent to a morpheme, the presence of polytonism, and a number of other related characteristics are observed.