numerals and number markers

Marcel Erdal erdal at EM.UNI-FRANKFURT.DE
Mon Feb 16 11:58:21 UTC 1998


I have a few oblique remarks on the subject, which involve Hebrew,
Turkish and Greek.
Firstly, with inital stress,
(1) $lo$-t-enu
can also mean 'ALL three of us', although these forms can also be expanded
by kol 'all':
(2)   kol $lo$-t-enu 'all three of us'.
The forms must be numeral-pronoun amalgams, since, if the speakers are
(say) cats, the expressions stand for
(3)   (kol) $lo$-et ha-xatul-im.
This, again, can be replaced by
(4)   (kol) kvuca-t/xavura-t ha-xatul-im 'the (whole) group of cats'
, suggesting that $lo$-et is not a counting word here (i.e. not a numeral)
but a noun referring to a group of three.
This is corroboratated by the fact that, in (Modern) Greek, we have
(5)   (ke) i tria to pir-ame   "(and) MASC.PL.ARTICLE three
PRON.NEUTR.SG.
				take-2.PL.PAST"   'we (all) three took it'
where the numeral has to be accompanied by the definite article.
Similarly, in (3), the definite article cannot be deleted without taking the
word 'three' out of its construct state. (3), that is, must mean something
like
(3)' 'the threehood' of cats.
Turkish is unlike Hebrew in not having a definite article but like Hebrew
in having possessive suffixes. (It is like Greek in that it can use the word
'and, also' to mean 'all' when going with numerals.) In Turkish, we have
(6) UC-UmUz "three-1.PL" 'three of us'
, perhaps out of a bigger group,
(7) UC-UmUz de "three-1.PL and,also" 'the three of us = we/us three'
and
(8) her UC-UmUz (de) "all three-1.PL and,also" 'all three of us',
the element de being optional in (7). How would one express (6) in
Hebrew? Not by (1), but by
(9) $lo$a me-it-anu   'three from-?-1.PL' 'three among us'.
All this suggests that (1) refers to a definite group of three ('THE three of
us'). This last statement, again, is corroborated by the fact that there are,
I think, limitations on the magnitude of the group in Hebrew and, possibly
also Turkish: I don't think it would be all too well accepted beyond 'five',
acceptability probably decreasing as one goes up, in both languages. This,
again, would suggest that what is meant is a RECOGNIZABLE group.
Other Turkic languages have, by the way, collectives, where the suffix
-a(g)u, -e(g)U is added to the number word. In those languages, it is
always this form which is used in the construction I was describing.
Summing up, I believe that we do not have numerals here, but collective
nouns coming from numerals, and referring to groups meant to be
recognized as such by the addressee.
Regards, Marcel Erdal
--
Prof. Dr. Marcel Erdal
J.W. Goethe Universitaet, FB 11, Turkologie,
Postfach 11 19 32,
D-60054 Frankfurt a.M.
Tel. +49-69-79 82 28 58, Fax. +49-69-79 82 49 74



More information about the Lingtyp mailing list