A language family consists of two or more tongues that are distinct and yet related historically in that they are all descended from a single ancestor language, either known or assumed to have existed. The languages of a family are closely related in phonetics, grammar, and vocabulary. The attempts made to classify Native American languages into such families have encountered various obstacles. One is the absence of written records of these languages except in the case of Aztec and Maya. Even there the texts are comparatively few in number; the Spanish conquerors destroyed almost all the texts they found. Another problem is that most records of any linguistic value were made after 1850. Also, there are at present insufficient numbers of trained persons able to record many of the indigenous American languages and collect data, especially in Mexico and Central and South America. The absence of grammars handed down from the past, owing to either the dearth of writing or the destruction of written texts, has further hampered the study of the Native American tongues. Linguistic scholars, therefore, have to turn to native informants to gain material for the analysis of these languages.
Native American languages cannot be differentiated as a linguistic unit from other languages of the world but are grouped into a number of separate linguistic stocks having significantly different phonetics, vocabularies, and grammars. Asia is generally accepted as the original home of the Native Americans, although linguistic investigations have not yet established any definite link between the Native American languages and those spoken in Asia or elsewhere in the Eastern Hemisphere. Some scholars postulate a connection between the Eskimo-Aleut family and several other families or subfamilies (among them Altaic, Paleosiberian, Finno-Ugric, and Sino-Tibetan). Others see a relationship between members of the Nadene stock (to which Navajo and Apache belong) and Sino-Tibetan, to which Chinese belongs; however, such theories remain unproved.