Re: [tied] Re: Numerals query again + Ge'ez forms

Sorry for the delay - kind of got forgotten in my drafts folder.

> On Thu, 27 Nov 2003 14:05:42 +0100 (MET), Harald Hammarstrom
> <haha2581@...> wrote:
>
> >I had to go look for my old papers and you are dead right. It's my memory
> >that was wrong. I quote unpublished Jan Retsö "Introduction to Sabaean",
> >2000. p. 4:
> >"The nominal declension
> >Most Sabaean nouns (substantives, adjectives and numerals) are declined
> >according to states. The two basic ones are the construct and the
> >absolute. The construct has no visible marking. The absolute is formed by
> >adding the suffix <m>:
> ><yd> - <ydm>
> ><hgr> - <hgrm>
> >
> >There is no semantic difference between the two states. They serve as
> >markers of syntactic relations between nouns. The absolute state can be
> >seen as the lexical form of a noun (its quotation form) and is the normal
> >form of a noun (substantive or adjective) which is not restricted by
> >genitival attributes in a sentence. For the use of the construct state,
> >see below, lesson 5 and 6.
> >
> >Sabaean distinguishes definiteness by adding a suffix <n> to the noun
> >which thus functions as a definite article. The article replaces the
> >absolute suffix. Most nouns thus have two independent forms:
> >
> ><hgrm> 'a town' - <hgrn> 'the town'
> >
> >The absolute-suffix is, however, not an indefinite article which is seen
> >from its frequent use in proper names which are always definite: <mlkm>
> >'Malik'.
> >
> >There's no certain indication that Sabaean had a case-inflection like the
> >Arabiyya or Akkadian. It has been assumed that the vocalisation of the
> >singular definite state was -a:n but it seems more likely that there was
> >an alteration between short and long vowels: -a:n/-an perhaps governed by
> >accent. In Arabic there are cases of isolated lexemes ending in a vowel
> >+ -n like bilGHa:n-un 'story-teller, gossiper', 3ljan-un 'robust, rough',
> >firsin-un 'camel's foot', burTun-un 'claw'. The absolute was -am thus:
> >baytam. With this should be compared the Arabic words like KHiDrim-un
> >'courageous', Duj3um-un 'strong', all probably loanwords from a dialect or
> >language (Himyaritic) close to Sabaean(1). Also forms like 'ibnum, 'ibnim,
> >'ibnam 'son' are documented in the Arabiyya(2)."
>
> Thanks. If I can summarize, the Sabaean noun had three "states":
>
> contruct absolute definite
> -0 -m -n
> /-a/? /-am/ -/an/ or /-a:n/

I don't think we have the right to suppose the constr. sg was /-a/ or
anything else.

> Now I would be very interested in knowing what the plural and dual
> equivalents were.

Either <-t> plural, probably pornounced /a:t/ like in Ge'ez and
Arabic (which looks exactly like the fem. sg) or broken pl. declined like
a regular singular noun. There's a also a meagerly attested:

-w/y -n -n

The cstr. pl. is attested only with the noun <bn> (which also has other
pl. formations). The only certain cases with this kind of pl. formation
in Sabaean is <bnn> 'sons', <ywmn> 'days' and the adjective <hn'n>
'healthy, sound'.

The dual is however a freq. category in Sabaean inscriptions:

cstr: abs: def:
-y -n -nhn

(Suffixed after the -t in fem. words). The Dual construct must have
been /-ay/ or /e:/. The dual definite may have sounded -ayna:n/-e:na:n and
the absolute -ayn/e:n. The def. pl. may have been either i:n (as in
Arabic) or -a:n (as in Ethiopic). In present-day Yemen there are plenty of
place-names ending in -a:n (Kawkaba:n, 3amra:n etc.) which may originally
be clan-names. In that case the suffix is a plural which is an argument in
favour of the -a:n.

All according to Jan Retsö _Introduction to Sabaean_ 2000.

> We've been discussing this on sci.lang, where I got side-tracked from
> *sab`atu(m)

I read an article by Hetzron in JSS 1977 which seconded my opinion that
the -ú in the numerals are from analogy with pronoun citing the
important piece of evidence of 'ahadú fem. 'ahattí just like the pronoun.

> into a discussion about Semitic nunation and mimation in
> general. I can give a short resumé of my own views here (although it's a
> bit off-topic [a lot, really]):
>
> The way I see it, proto-Semitic had a construct state:
>
> sg. masc. fem.
> nom -0 -at-0
> acc -0 -at-0

Why no -a here? If not, why does in appear when -n in added?

> gen -i -at-i
> pl.
> nom -u: -a:t-0/-a:n-0
> obl -i: -a:t-i
> du.
> nom -a: -at-a:
> obl -a-i: -at-a-i:

Why the long i:? Just to explain the Aramaic?

> There was a definite state, constructed by adding a old pronominal article
> *V, pl./du. *Vn):
>
> sg. masc. fem.
> nom -u -at-u
> acc -a -at-a
> gen -i -at-i
> pl.
> nom -u:-n -a:t-u
> obl -i:-n -a:t-i
> du.
> nom -a:-n -at-a:-n
> obl -a-i:-n -at-a-i:-n
>
> Note that the fem.pl., which already contains the plural marker -a:t- (~
> -a:n-) does not add the plural form of the "article".
>
> Finally, there was an "indefinite" state, with *-n (perhaps *-m?) added to
> all forms, whether singular, plural and dual:
>
> sg. masc. fem.
> nom -u-n -at-u-n
> acc -a-n -at-a-n
> gen -i-n -at-i-n
> pl.
> nom -u:-n -a:t-u-n
> obl -i:-n -a:t-i-n
> du.
> nom -a:-n -at-a:-n
> obl -a-i:-n -at-a-i:-n
>
> Akkadian has a construct and an absolute state, where the absolute state
> continues the old "indefinite". The final sequence *-u(:)n becomes
> *-u(:)m, and the *-m is analogically extended to the oblique case(s).
> After a long vowel, *-m (but not *-n) is dropped. This results in
> Akkadian:
>
> sg. masc. fem.
> nom -u-m -at-u-m
> acc -a-m -at-a-m
> gen -i-m -at-i-m
> pl.
> nom -u: -a:t-u-m
> obl -i: -a:t-i-m
> du.
> nom -a:-n -at-a:-n
> obl -i:-n -at-i:-n
>
> Canaanite (Ugaritic, Phoenician, Hebrew...) also has a construct and an
> absolute state, but here the absolute continues the old definite state.
> The *-m from *-um/*-u:m has here been extended to the dual as well:
>
> Ugaritic Hebrew
> nom -u -at-u
> acc -a -at-a -0 -a:(h) < *-ata
> gen -i -at-i
> pl.
> nom -u:-m(a) -a:t-u
> obl -i:-m(a) -a:t-i -i:m -o:t
> du.
> nom -a:-m -at-a:-m
> obl -i:-m -at-i:-m -ayim -a:tayim
>
> Aramaic is essentially the same, but has generalized nunation:
>
> sg. -0 -a:(h) < *-ata
> pl. -i:-n -a:n
> du. -a-yi-n -&t-a-yi-n
>
> Note the fem.pl. in -a:n for expected -a:t (Hebrew -o:t).
>
> Arabic maintains all three states:
>
> construct definite indefinite
> nom -u -at-u -u -at-u -u-n -at-u-n
> acc -a -at-a -a -at-a -a-n -at-a-n
> gen -i -at-i -i -at-i -i-n -at-i-n
> pl.
> nom -u: -a:t-u -u:-n(a) -a:t-u -u:-n(a) -a:t-u-n
> obl -i: -a:t-i -i:-n(a) -a:t-i -i:-n(a) -a:t-i-n
> du.
> nom -a: -at-a: -a:-n(i) -at-a:-n(i) -a:-n(i) -at-a:-n(i)
> obl -a-y -at-a-y -a-y-n(i) -at-a-y-n(i) -a-y-n(i) -at-a-y-n(i)
>
> As in NW Semitic, the singular of the construct state has been equated with
> the singular of the definite stae.
>
> So, I would like to know how Sabaean fits into this.

I don't know more than the above but if you have a mailing address I can
send you photocopies of Retsö (+ dictionary and corpus of
inscriptions) and/or Lambdin and you can double check yourself. I can
copy and send from work so it does not cost me a nickle.

all the best,

Harald