The Classification of Languages 213
be as many as twenty, as in Bantu dialects, or it may be as few as four,,
as m one of the dialects of the Australian aborigines. The classificatory
mark is not necessarily a prefix. In the Papuan language cited by
Capell, it is a suffix like the gender-terminal of an Aryan adjective
Thus the distinction between the classificatory and the flexional type
is not so sharp as it first seems to be The trade-mark of the Indo-
European adjective as a separate entity is that it carries the suffix
determined by one of the three gender-classes to which a noun is
assigned We know that what are called adjectives in Aryan languages
were once indistinguishable from nouns, and the example of Finnish
(p 197) shows us how easily the ending of the noun gets attached to
an accompanying epithet In each of the three Aryan gender-classes
we meet with a greater or less proportion of nouns with characteristic
affixes limited to one of them, and the notion of sex which an
American or an Englishman associates with gender has a very flimsy
relation to the classification of Indo-European nouns in their respective
gender-classes.
Though we have no first-hand knowledge about the origin of gender,
we know enough to dismiss the likelihood that it had any essential
connexion with sex The most plausible view is that the distinction of
gender in the Indo-European family is all that is left of a system of
suffixes essentially like the Bantu prefixes If so, the former luxuriance
of such a system has been corroded in turn by nomadic habits and
civilized hving as primitive Aryan-speaking tribes successively came
into contact with new objects which did not fit into the framework of a
classification suited to the limited experience of settled life at a low
level of technical equipment
PHONETIC PATTERN OF LANGUAGE FAMILIES
Just as we recognize grammatical processes such as isolation, aggluti-
nation, amalgamation, root-inflexion, we can also recognize sound-
patterns which predominate in one or other group Such phonetic
patterns furnish us with an additional clue to linguistic affinities,
albeit a clue which too few philologists have followed up Our last sec-
tion illustrates one phonetic type which is distributed over a large part of
the world In a multitude of unrelated languages, including Japanese,
Malayo-Polynesian, and Bantu dialects, agglutinative regularity coexists
with a sound-pattern quite unlike that of our own language or of any
languages related to it Jespersen (Growth and Structure of the English
Language) illustrates the contrast by the following passage from the