34 R. M. W. Dixon
I and basically H:
- Many of the languages of North America (including most languages in Na-dene,
Algonquian-Ritwan, Salish, Siouan, Iroquoian, Muskogean, Tsimshian, Zuni). - Some languages from South America (including the Arawak family)
- Most Austronesian languages (excluding those in the Philippines).
- Ainu.
I and neither H nor D
- Most languages from south-east and east Asia (including Sinitic, Tibeto-Bur-
man, Tai-Kadai, and some Austroasiatic).
It should be emphasized that this is very much a first-run-through of the data.
Detailed study of the adjective classes in individual languages is required. There
may, indeed, be languages of more than one type within a single genetic or areal
grouping. Surveying Nilo-Saharan languages, Dimmendaal (2000: 218-19) notes
that—in accordance with (33)—'adjectives tend to pattern with nouns in depend-
ent-marking languages and with verbs in head-marking languages'.
There are a number of exceptions to the generalization in (33), some of these
being of particular interest.
(i) It is clear that, at an earlier stage, Australian languages were entirely dependent
marking; in keeping with this, adjective classes are almost all 'non-verb-like' and
also 'noun-like'. (In fact, fairly subtle criteria have to be applied, in most languages,
to distinguish between adjectives and nouns; see the discussion of Alpher's crite-
ria in §5.)
In recent times, bound pronouns have evolved over a good deal of the conti-
nent. In most of the languages in which they occur, these are clitics attached to the
verb or a verbal auxiliary, and they are not always obligatory. However, languages
over a continuous area in the central north have developed obligatory pronominal
prefixes to verbs, a clear head-marking strategy. As a consequence they have lost
or are in the process of losing dependent marking from NPs (see Dixon 2002 for
full details).
Interestingly, a couple of these head-marking languages appear to be assign-
ing more verb-like properties to their adjective class. It was mentioned, in §6.3,
that in Emmi adjectives are negated like verbs, differently from nouns. And that in
Nunggubuyu an adjective may take subject pronominal prefixes, like an intransi-
tive verb, showing that it is functioning as head of an intransitive predicate.
The shift of a language from a dependent-marking to a head-marking profile is
well-attested. Bound pronouns develop from what were free forms, and are obliga-
torily included in each predicate, with the old dependent marking on NPs drop-
ping out of use. It maybe that the shift from a 'non-verb-like' to a 'verb-like' adjec-
tive class—in order to re-establish the correlation in (33)—tends to follow the shift
from dependent to head marking, but operating at a slower pace.