Re: Paleoasiatic (was: Favourite Language Group?)

John Cowan writes:
 > Tim May scripsit:
 >
 > > at least 26 numeral classes,
 >
 > Probably the degenerate descendants of a classifier+numeral+noun system.
 >

Possibly.  They sound rather unusual to me, although I don't know much
about numeral classes/classifiers beyond Japanese.  I can copy _that_
part, anyway.  Here it is.

|Numerals
|
|Nivkh has specific sets of numerals for the enumeration of various
|categories of objects - boats, people, animals, fishing-nets, small
|round obj.cts, etc.  Panfilov (1962) lists 26 such sets.  For example,
|`three' with reference to people is _t'aqr_; to animals, _t'3m_; to
|sweep-nets, _t'for_; etc.  It is noteworthy that the palatalised t'
|remains stable as initial in all variants, an the same stability
|applies in other numbers.  Thus, _n'_ is the initial for the 26
|variants of _n'im_ 1, and m is the initial for all variants of _mim_ 2.

Replies