Re: a:-stems

From: Miguel Carrasquer Vidal
Message: 58368
Date: 2008-05-05

On Sat, 3 May 2008 19:26:11 -0500, "Patrick Ryan"
<proto-language@...> wrote:

>Miguel, in our discussion of the 'Vocalic Theory', you mentioned an example
>of a development that you thought the VT could not explain:
>specifically, the instrumental singular of -*a-stems, which you set forth as
> That's hardly satisfactory. We have three morphemes here:
> the thematic vowel (*-o-), the feminine marker *-yeh2- ~
> *-ih2-, and the instrumental ending *-eh1. The combine as
> *-o-yh2-ah1, with colouring of the instrumental ending by
> the *h2 of the feminine.
>I would like to clarify that is what you meant to write because it does not
>correlate with other information I have on this inflection.
>We are talking, I presume, about the -*ia: variant of the *a-stems.
>First, I see no trace of a "thematic vowel". I can find no reference to
>*a-stems having thematic vowels in the materials I have.
>My sources list the instrumental singular of *a:-stems as *Ø.
>You called it -*yH2eH1, which amounts to -*a:.
>Can you clarify this at all?

Of course.

The declension of the a:-stems is usually reconstructed more
or less as follows:

Nom *-a: (= *-eh2)
Skt. -a:, Lith. -à/-ó-, Slav. -a, Goth. -a/-o:, Grk.
-e:/-a:, Arm. -0, OIr. -0. Latin is irregular with short -a
(as in the

Acc *-a:m (= *-eh2-m)
Skt. -a:m, Lith. -aN, Slav. -oN, Goth. -a/-o:, Grk.
-e:n/-a:n, Lat. -am, Arm. -0. Old Irish is irregular with
-YN (< *-en). Note that *-m is consonantal here and in the
Apl. (we do not have *-eh2-m., *-eh2-n.s)

Voc *-a (= *-e[h2])
Lith. -a, Slav. -o, Grk. -a. The laryngeal coloured the
vowel before it was dropped. Sankrit -e (*-ai) is irregular
here, more below.

Gen *-a:s (= *-eh2-os)
Skt. (root nouns) -a:s, Lith. -õs(-), Slav. -y, Goth. -o:s,
OLat. -a:s, Grk. -e:s/-a:s. See below for unexpected Skt.
-a:ya:s (and perhaps OIr. -Ye (*-ia:s)).

Dat *-a:i (= *-eh2-ei)
Skt. (root nouns) -a:i, Lith. -ai, Slav. -ê, Goth. -ai, OIr.
-Y, Lat. -ae, Grk. -e:i/-a:i, Arm. -i. See below for Skt.

Loc *-a:[i] (= *-eh2[-i])
Lith. -õj-e, Slav. -ê, Lat. -ae, ?Grk. -ai, Arm. -i.
Sanskrit root nouns in -a: would probably have had -a:m
(*-a: + -m) if any forms had been attested. For Skt. -a:ya:m
see below.

Ins *-a: (= *-eh2-eh1)
Skt. -a:, Lith. -à, Arm. -aw (*-a:-bhi). For Skt. -aya:,
Slav. -ojoN, Grk. -ã, Arm. -oj^, see below.


Nom *-a:s (= *-eh2-es)
Skt. -a:s, Lith. -os, Goth. -o:s, OIr. -a, OLat. -a:s, Arm.

Acc *-a:ns (= *-eh2-ns)
Skt. -a:s, Lith. -às, Slav. -y, Goth. -o:s, OIr. -a, Lat.
-a:s, Grk. -a:s, -ans, Arm. -s.

Gen *-a:o:m, *-a:Co:m (= *-eh2-õm)
Skt. -a:n.a:m/-a:sa:m, Lith. -u~N, Slav. -U, Goth. -o:, OIr.
-N, Lat. -a:rum, Grk. -õn.

Dat *-a:Mos (= *-eh2-Mos)
Skt. -a:bhyas, Lith. -óms, Slav. -amU, Goth. -o:m.

Loc *-a:su (= *-eh2-su)
Skt. -a:su, OLith. -osù, Slav. -axU

Ins *-a:Mis (= *-eh2-Mi[:]s)
Skt. -a:bhis, Lith. -omìs, Slav. -ami, Arm. -awk`

NA *-ai (= *-eh2-ih1 ?)
Skt. e:, Lith. -ì, Slav. -ê, Oir. -Y

GL *-a:ous (= *-eh2-[H]ou[s] ?)
Slav. -u. For Skt. -ayo:s, see below.

DI *-a:Mo: (= *-eh2-Moh3)
Skt. -a:bhya:m, Lith. -óm/-om~, Slav. -ama.

The *ih2-stems of the de:vi: type are consonant stems, so
they originally showed stress alternation between the root
and the endings as follows:

Nom *-i: (= *'-ih2)
Skt. de:vi:

Acc *-i:m (= *'-ih2-m)
Skt. de:vi:m

Voc *-i (= *'-i[h2])
Skt. de:vi

Gen *-ya:s (= *-yéh2-os)
Skt. de:vya:s

Dat *-ya:i (= *-yéh2-ei)
Skt. de:vya:i

Loc *-ya:[i] (= *-yéh2-[i])
Skt. de:vya:m

Ins *-ya: (= *-yh2-éh1)
Skt. de:vya:

Nom *-i:[e]s (= *'-ih2-es)
Skt. de:vi:s

Acc *-i:ns (= *'-ih2-ns)
Skt. de:vi:s

Gen *-i:[C]o:m (= *-ih2-ó:m)
Skt. de:vi:na:m

Dat *-i:Mos (= *-ih2-Mós)
Skt. de:vi:bhyas

Loc *-i:su (= *-ih2-sú)
Skt. de:vi:s.u

Ins *-i:Mis (= *-ih2-Mí[:]s)
Skt. de:vi:bhis

NA *-i: (= *-ih2-íh1)
Skt. de:vi:

GL *-yous (= *-yh2-[H]óu[s])
Skt. de:vyo:s

DI *-i:Mo: (= *-ih2-Móh3)
Skt. de:vi:bha:m

The stems in *-ih2 are used make feminines of both consonant
stems (e.g. participles in -a(n)ti:, -us.i:) and thematic
stems, such as de:vi: (f. of de:vas < *deiw-o-s). The stems
in *-eh2 are used to make thematic feminines (e.g. the
adjectives in *-os ~ *-eh2). It would make sense, then, if
the a:-stems could be analyzed as thematic stems (*-e-) with
the (athematic) feminine suffix *-ih2 added. The problem of
the missing *-i- can be resolved by positing a soundlaw that
deletes the *i after a stressed thematic vowel *-é-. That
would give:

Nom *-a: (= *-é-[i]h2)
Skt. -a:, Lith. -à/-ó-, Slav. -a, Goth. -a/-o:, Grk.
-e:/-a:, Arm. -0, OIr. -0.

Acc *-a:m (= *-é-[i]h2-m)
Skt. -a:m, Lith. -aN, Slav. -oN, Goth. -a/-o:, Grk.
-e:n/-a:n, Lat. -am, Arm. -0.

Voc *-oi (= *'-o-i[h2])
Skt. -e:

Gen *-oya:s (= *-o-yéh2-os)
Skt. -a:ya:s, Slav. (pron.) -ojeN, OIr. -Ye (< *-ia:s).

Dat *-oya:i (= *-o-yéh2-ei)
Skt. -a:ya:i, Slav. (pron.) -oji.

Loc *-oya:[i] (= *-o-yéh2[-i])
Skt. -a:ya:m, Slav. (pron) -oji.

Ins *-oy[y]a: (= *-o-yh2-éh1)
Skt. -aya:, Slav. -ojoN, Grk. -ã, Arm. -oj^.
The -a- in Skt. is short because of the laryngeal in
*-oyh2ah1. The development -ojja: > -oj^ is regular in
Armenian (the oblique -oj^ is found in kin, knoj^ "woman",
and in ija-stems such as t`eli, t`elwoj^). Slavic has added
-m (*-oj?a: + -m > -ojoN) [for which cf. the Skt. Lsg. and
DIdu. above, the Grk. dual oblique -oîn mentioned below, and
the Lith. adjectival Isg. -áN-]. The Greek fossil
instrumental in -ã (e.g. <kruphã>) cannot represent *-ah2ah1
> *-a: (which gives an acute in Lithuanian -à), and is
better explained as contracted from *-oy[h2]ah1.


Nom *-a:s (= *'-e-[i]h2-es)
Skt. -a:s, Lith. -os, Goth. -o:s, OIr. -a, OLat. -a:s, Arm.

Acc *-a:ns (= *'-e-[i]h2-ns)
Skt. -a:s, Lith. -às, Slav. -y, Goth. -o:s, OIr. -a, Lat.
-a:s, Grk. -a:s, -ans, Arm. -s.

Gen *-oi:sõm (= *-o-yh2-s-õ'm)
Slav. (pron.) -êxU. The forms of the plural are identical
to those of the masculine. Any intonational differences
(*-oj-sõm vs. *-o-jh2-sõm) would have been eliminated by
Meillet's law.

Dat *-oi:Mos (= *-o-yh2-Mós)
Slav. (pron.) -êmU

Loc *-oi:su (= *-o-yh2-sú)
Slav. (pron.) -êxU

Ins *-oi:Mi:s (= *-o-yh2-Mi:s)
Slav. (pron.) -êmi

NA *-oi: (= *-o-yh2-íh1)
Skt. e:, Lith. -ì, Slav. -ê, Oir. -Y

GL *-oyyous (= *-o-yh2-[H]óu[s])
Skt. -ayo:s, Slav. (pron.) -oju. Cf. also Grk. -oîn (masc. &
fem.) < *-oj-h3u or *-o-jh2-h3u + -m. The merger of masc.
pl./du. oblique *-oy- with fem. *-o-yh2- would explain why
no language distinguishes o- and a:-stems in the dual.

DI *-oi:Mo: (= *-o-yh2-Móh3)
Slav. (pron.) -êma.

We see that Indo-European shows traces of two a:-stem
paradigms: an archaic one with alternation *-é-h2 ~
*-o-yéh2- ~ *-o-yh2-', best preserved in Sanskrit and Slavic
[but with loss of any alternation in stress position], and
an innovated one with *-eh2- throughout [and columnal
accentuation on the root or on the thematic *-éh2-], as
reflected by the traditional reconstruction given at the top
of this post.

Miguel Carrasquer Vidal