Races and Peoples: Lectures on the Science of Ethnography

by Daniel G. Brinton

1. Isolating	Chinese, Thibetan, Sifan, Tai.
	Siamese, Annamite, Burmese, Assamese.
2. Agglutinative	1. By reduplication and prefixes	Polynesian, Papuan, Bantu.
	2. By suffixes	Sibiric tongues, (Ural-altaic), Basque.
		Japanese, Korean, Dravidian.
3. Incorporative	1. With synthetic tendency	Algonkin, Nahuatl.
		Quichua, Guarani.
	2. With analytic tendency.	Otomi, Maya, Sahaptin.
4. Inflectional	1. By annexing grammatical elements.	Egyptian.
	2. By inner changes of stem.	Libyan, Semitic.
	3. By addition of suffixes.	Aryac tongues.

The principles on which languages should be compared are frequently misunderstood, and this is one of the reasons why the value of linguistics to ethnography has so often been underrated.

The first rule which should be observed is to rank grammatical structure far above verbal coincidences. The neglect of this rule will condemn any effort at comparison. For example, there have been writers who have sought to derive the Polynesian, an agglutinative, from the Sanscrit, an inflected tongue; or an American from a Semitic stock. Such attempts reveal an ignorance of the nature of language.

A second rule is that in tracing the etymology of words, the phonetic laws of the special group to which they belong must be followed. This is an even more frequent source of error than the former. Writers of high reputation will trace variations in African or American or Semitic names by the phonetic laws of the Aryac dialects—an absurd error, as the phonetic changes are not at all the same in different linguistic stocks.

Yet a third rule is to appraise correctly the value of verbal identities. Generally, it is placed too high. All developed tongues include many “loan words,” borrowed from a variety of sources. They are not prima facie evidence of ethnic relation; they have frequently been transmitted through other nations, as is the case with thousands of English words.

An absolute verbal identity is always suspicious; or rather it is of no ethnic value. There must be a series of words in the languages compared of the same or similar meanings, but whose forms have been altered by the phonetic laws peculiar to the group, for such lists of words to merit the attention of a scientific linguist.

The question how far languages can be accepted as indicating the relationships of peoples has been a bone of contention. One principle we may lay down, with unimportant exceptions—No nation has ever willingly adopted a foreign tongue. Whenever such a change has taken place, it has been under stress of sovereignty, vi coactum, as the lawyers say. Hence in the savage state, where prolonged domination of one tribe by another rarely occurs, language is an excellent ethnic guide, as in America and ancient Europe.

Another principle is that in a conflict of tongues, as after conquest, that tongue prevails which belongs to the more cultured people, whether this be conqueror or conquered. This is well illustrated by the survival of the Romance languages after the inroads of the Teutonic hordes at the Fall of the Western Empire.

A third maxim in linguistic ethnography is that mixture of languages, especially in grammatical structure, indicates mixture of blood. When, for instance, we find the Maltese a dialect partly Arabic, partly Romance, we may correctly infer that the people of the island are descended from both these stocks. This holds good even of loan words, when they are numerous; for though such have no influence on the grammatical structure of a tongue, they testify to some relations between nations, which we may be sure corresponded to others of a sexual nature.

The “American citizens of African descent” speak English only; and though they have been in contact with the white race for but three or four generations, the majority of those now living are related to it by blood, that is, are mulattoes.