Formal and functional linguists give competing explanations for lexical
categories (nouns, verbs, adjectives). The latter explicate them in
terms of
meaning and use while the former do so in terms of syntax. I will show
that
both approaches capture only part of the story, because lexical
categories
are primitives, constituting a unique category that is not reducible to
either syntax alone or meaning/use alone.
1. In formal accounts like Baker 2003, verbs are defined as the category
that requires a specifier (subject). However, there are verbs such as
ada in
Malay that lack subjects in certain constructions. Subject-less
ada-sentences correspond to there-sentences in English.
2. Baker also defines adjectives as an elsewhere category resulting from
syntactic necessity, i.e. it fills positions in a syntactic structure
that
exclude both nouns and verbs. However, there are syntactic positions
where
both nouns and adjectives occur, such as the complement of be-verbs.
They
also occur there with different meanings, contrary to Baker's claim that
cognate nouns and adjectives differ syntactically rather than
semantically.
(2) The water is ice.
(1)
Ada beberapa jenis buah yang dijual di kedai itu.
be/have several kind fruit rel pro sold prep shop dem
"(There) are several kinds of fruit that are sold in that shop."
(3) The water is icy.
3. In functional accounts like Hopper and Thompson 1984 and Givon 1984,
for
example, lexical categories have a prototype-structure: nouns typically
denote things, adjectives typically denote states or properties, and
verbs
typically denote events. This characterization corresponds to a
continuum of
permanence, with nouns denoting the most permanence and verbs the least.
However, the generalizations are not useful in defining the categories.
A
sneeze is neither permanent nor thing-like, and the verb exist could
denote
a permanent state but is not an event. These categories do not exhibit
true
prototype-organization. Nouns for example (honesty, firm, corporation,
quark, dog, destruction) are not related in a network of similarities.
4. Languages have words that are ambiguous with respect to categories,
such
as fun in English, which shows properties of being both a noun and an
adjective.
(4) The party seems fun.
(5) They had fun at the party.
Such cases suggest that words are matched to non-derivable categories
rather
than categories being matched to either syntactic structure or human
perception and the external world.
5. There are well-known crosslinguistic differences with respect to
lexical
categories. For example, the concept of intelligence can be expressed by
an
adjective in English but only by a noun, nzeru, in Chichewa, and this
suggests that lexical categories are not reducible to syntax or
semantics.