Re: Evolution of Plural Markers

--- In conlang@yahoogroups.com, Rob Haden <magwich78@Y...> wrote:
> What are the realistic ways by which plural markers develop?
  In Hungarian the adjectival suffix <-s> [S] has plural
(collective, to be precise) connotation in a subset of words. E.g.
from the word <nád> [na:d] 'reed, Phragmites communis', we can form
an adjective <nádas> [na:dOS] with the meaning 'with reed, of reed,
resembling to reed etc.'. When this word is used as noun, it
denotes 'reed plot, place where reed grow' with a strong collective
sense like 'place full of reeds, many reeds'.

  This suffix <-s> [S] is a possible candidate to become a plural
marker in the "future Hungarian". There are evidences that similar
processes already took place in the past, because the present
Hungarian plural marker <-k> seems to be etymologically related to
the Finnish collective suffix <-kko> in words like <kuusi> 'fir' >
<kuusikko> 'fir plantation, many firs', <kivi> 'stone' > <kivikko>
'place full of stones'. In the case of the Finnish word <kuusikko>,
we have the exact parallelism between the Finnish collective suffix
<-kko> and the Hungarian adjectival suffix <-s> [S], cf. Hungarian
<fenyo"> [fEJ9:] 'fir, pine' + <-s> [S] > <fenyves> [fEJvES] 'fir
plantation, pine forest'.

  Even one of the reconstucted plural markers in the
Proto-Finno-Ugrian *<-jA> is usually compared to another Hungarian
adjectival suffix <-i> [i]. This suggests that this plural marker
evolved from an adjectival suffix in the protolanguage, too.

-----

  Historically, the Common Turkic plural marker <-lAr> is a
compound suffix. It's composed of an <-l> element and the <-Ar>
collective suffix.

  This <-l> element is considered as an older plural marker, but
it's worthy of note that Turkic languages have an adjectival suffix
<-lI>, cf. <Osman> proper name > <osmanli> '(Empire/people) of
Osman'.

-----

  Some of the Finnish and Esthonian dialectal plural markers, like
<-lOi>, <-se>, <-a> were abstracted by restructuring the morphemic
chain. E.g. <-lOi> is a combination of an original <-i> plural
marker and a derivational suffix that is semantically dummy in word
pairs like <peukku> ~ <peukalo> 'thumb'. <-se> and <-a> were
originally root alternations before inflexional suffixes in certain
paradigms.

-----

  In Sumerian there were half a dozen plural categories. As far as
I know, animate determined (demonstrative) plural marker <-ene>,
eg. <lú> 'man', <lú-ene> 'certain men', seems to be connected with
the demonstrative suffix (enclitic pronoun) <ne(-e)> 'this (one)'.
  The determinative plural marker <-dili-dili> (and its contracted
form <-didli>) is a reduplication of the numeral <dili> 'one,
single' with a meaning of 'one by one', cf. <mu-dili-dili> 'entries
one by one, individual entries (in a lexical series)', <lú-dili-
dili> 'men particularly, individual men'.
  The late Sumerian plural marker <-me-ec> has a literal meaning
of 'they are' (<me> 'to be' + <-ec> intransitive 3rd personal
plural suffix) and it was abstracted from a subordinate phrase.

-----
Notes:
- In Sumerian data ETCSL display conventions were used
- In Finno-Ugrian and Turkic suffixes capitals represent vowels
  affected by vowel harmony, eg. <-lAr> = <-lar> + <-lär>
- In Hungarian <o"> stands for latin small letter o with double acute