15 What can we conclude? 359Adjectives that are exclusively attributive provide the best evidence in favour
of a fully independent adjective class in languages where this may otherwise be
in doubt, as appears to be the case for Chinese for instance. If adjectives cannot
function directly as NP heads (noun-like), nor as predicates (verb-like) or cop-
ula complements (noun-like), and have exclusively attributive function (adjective-
like), then they are plausibly assigned to an independent adjective class, as they are
in languages such as Hua (see Chapter i and Wetzer 1992), As already discussed in
§2 above, the existence of what appears to be a large class of attributive-only adjec-
tives (some 200-400 members) has recently been reported in Chinese.^3
Evidence of what appear to be adjective-like attributive-only forms is found in a
number of languages in this volume, including Japanese, Korean, Mam, Manange,
North-East Ambae, and Qiang. In most cases, only rare examples can be found. In
Manange a small set of negated verb-like adjectives, e.g. a-kye 'unpleasant/ugly',
are attributive-only. No such restrictions apply to their positive counterparts, e.g.
kye 'pretty/nice' (see Chapter 3). In Wolof, the common borrowed formpH'r 'pure'
(< French pur) is unlike all other adjectives in showing no verb-like behaviour: it is
exclusively attributive and directly modifies the noun without relativization.
The existence of a small set of forms that are attributive-only with clear adjectival
function has long been recognized in Japanese and Korean grammatical descrip-
tion. Backhouse (Chapter 2) notes that such adjectives, e.g. roku (no) '(no) good',
(and demonstratives) have been called 'adnouns' in the past (e.g. Martin 1975) and
behave very much like noun-like uninflected adjectives in the noun phrase.
Sohn also refers briefly (fn. 3, Chapter 9) to the existence of a small Specifier or
pre-noun class of adjective-like terms (e.g. mat 'oldest', say 'new'), demonstratives,
and numerals. He does not wish to treat any of these items as adjectives since they
do not share the verbal properties of inflection and predicative function that the
large independent class of verb-like adjectives otherwise has in that language. Lee
and Ramsey (2000) also refer to this class as pre-noun, but, significantly, Lee (1989)
places adjective-like members in their own small adjective class. There is good
syntactic evidence in favour of Lee's (1989) decision to distinguish these uninflect-
ed adjectives from demonstratives and other members of his suggested specifier/
pre-noun class. The word order of adjectives is rigid and they always appear dir-
ectly before the modified noun. The order of demonstratives is not fixed and can
be scrambled with other pre-nominal modifiers such as possessives and relative
clauses. None of these elements can ever, however, intervene between pre-nominal
adjective and noun. Hence ce nay chayk (lit. that my book)and nay ce chayk (lit. my
that book) 'that book of mine' are acceptable, but there is no alternative to ce say
cheyk (lit. that new book) 'that new book (Sohn 1994: 224-5).
LaPolla and Huang in their description of adverbial phrases in Chapter 13 briefly
discuss the existence of an unnamed class of words used mainly with adverbial func-
(^3) Thai also appears to have attributive-only adjectival elements, albeit apparently few in number.
Sookgasem (1996) treats these as adjectives, although more information is required. See Prasi-
thrathsint (2000) for an alternative view of adjectives in Thai.