On Thu, 25 Oct 2001 09:10:08 +0100, "P&G" <
petegray@...>
wrote:
>> In reality,
>> a number always qualifies some other noun unless you're talking
>> algebra.
>
>Alas, no. What you say is true of course at a deep level, but if you are
>talking human languages, then the surface construct could be different. I
>have already pointed out that Latin milia (= 1000's) is a noun, and cannot
>qualify another noun.
In Chinese, numerals can almost never qualify another noun (you need a
numerical classifier). In Slavic, 1 is an adjective, 2, 3 & 4
sort-a-kinda, and from 5 up the numerals are basically nouns (the noun
counted goes in the gen. pl.). In Arabic, 1 (waaHid) is an adjective
[?aHad is a pronoun], 2 sort-a-kinda (it agrees in gender with the
noun qualified), 3-10 are like nouns (counted noun in gen.pl., but
numeral disagrees in gender with counted noun [*]). 11-19 govern the
noun in the acc. sg. (13-19 disagree in gender).
[*] i.e. with fem. nouns, the numeral has its base form, with masc.
nouns it adds the "feminine" ending -atu(n) [which is also a plural
morpheme, at least in Akkadian].