LINGUIST List 8.230

Tue Feb 18 1997

Disc: Low vowels in PIE

Editor for this issue: Ann Dizdar <>


  1. Patrick C. Ryan, Re: 8.137, Disc: Low vowels in PIE

Message 1: Re: 8.137, Disc: Low vowels in PIE

Date: Sun, 02 Feb 1997 19:54:16 -0600
From: Patrick C. Ryan <protol9IDT.NET>
Subject: Re: 8.137, Disc: Low vowels in PIE

Another short comment on Steven Schaufele's summary:

I am amazed that the argument about monovocality in IE still rages on.

That Late Nostratic (and earliest IE and AA) HAD to be monovocalic is
easy to prove.

Language A: C(1)eC(2) = "come"; C(1)aC(2) = "stay"; C(1)oC(2) =

In Language A, Ablaut and vowel gradation are incompatible with
maintaining semantic integrity. For Ablaut, if we altered the <e> to
<o> of C(1)eC(2) to create a nominal, "arrival" (C(1)oC(2)), it would
be indistinguishable from C(1)oC(2), "laugh". This would not be
acceptable on the wholesale basis which would be inevitable.

Language B: C(1)[y]aC(2) = "come"; C(1)aC(2) = "stay"; C(1)[w]aC(2) =

In Language B:, Ablaut and vowel gradation can be seen to be easily
possible without impairing semantic integrity. C(1)[y]eC(2) can be
"coming"; C(1)[y]aC(2) can be "presence"; and C(1)[y]oC(2) can be
"arrival" without being confused with C(1)[w]aC(2) or C(1)aC(2).

The brackets indicate palatal and velar glides which both contrast with
NO glide.

In languages like Kabardian, which is either monovocal or very close
to it, the quantity of simple roots is expanded by precisely the
mechanism I have suggested: consonantal glides (an no glide).

Since IE is thought to have originated in the geographical area of
Caucasian languages, is it such a large leap of logic to suggest that
IE (just prior to Ablaut), had a phonological system similar to many
Caucasian languages, i.e. characterized by vowel glides (and no
glide), and non-phonemic vowels (areal phenomenon?).

Monovocality was convincingly demonstrated for IE by W.P. Lehmann,
who, called the phenomenon "syllabicity". It is also likely for
Nostratic based on the arguments of I. Diakonoff for AA.

The commonest vowel of the earth's languages is low central "a", a
vowel that requires the absolute minimum of tongue movement. This is
the vowel that is combined with syllables with consonantal glides
since the glides (contrasting with no glide) provide the
differentiation necessary for simple root formation.

PRIOR to Ablaut or vowel gradation, there would be no particular
difficulty in envisaging a V which accomodated itself to its phonetic
environment: e.g. Labial + V = rounded V but AFTER Ablaut or vowel
gradation, allophonic phonetic environmental variations of V would
conflict with the grammatical information conveyed by vowel
alteration, and would be impermissible.

This leads to the conclusion that POST-Ablaut IE had one V which
contrasted with no other vowel to provide SEMANTIC differences but had
allophones which conveyed GRAMMATICAL differences: realized roughly as
/e/ (imperfective), /a/ (stative), and /o/ (perfective).

If /e/ and /o/ (and /a/) WERE phonemic, we should be able to find
entries in an IE dictionary where C(1)eC(2) contrasts semantically
with C(1)oC(2) on a regular basis. Everyone knows that this is not
possible so how can anyone maintain the phonemicity of IE /e/ and /o/?

After additional formants were added to simple roots characterized by
consonants with glides (and no glide), the semantic differentiation
previously provided by them became unnecessary. E.g. let us suppose
that m[y]Vn means "speak" and mVn means "remain" and m[w]Vn means
"crush". If -H was added to m[y]Vn (m[y]VnVH) the glide would have
become redundant to distinguish mVnVH from mVn and m[w]Vn UNLESS the
identical formant was added to them. By this process, glides became
redundant and disappeared.

If one objects to glides for IE, one has only to look at the Baltic
and Slavic languages for evidence of palatal glides, and at
Thurneysen's treatment of Old Irish for a residual system based on
palatal/velar and no glide.

Another systemic advantage of a system characterized by glides rather
than vocalic contrasts is that stress-accent can be introduced to
provide other nuances more easily since stress-unaccented syllables
maintain their vocalic integrity with difficulty.

This, of course, leaves us with the problem of /i/ and /u/.
Very competent scholars have argued the question of whether IE
diphthongs are phonemic, and whether the -i and -u of such diphthongs
is a phonetic reduction from avocalic /j/ and /w/ OR whether /j/ and
/w/ are modifications of original /i/ and /u/ FROM IE EVIDENCE.

While the IE evidence can be interpreted in different ways (perhaps),
the facts of Nostratic (which, at the very least, is a common ancestor
for IE and AA) conclusively show that IE /i/ and /u/ are reductions
from /j/ and /w/ since they reflect AA /j/ (but also /$/ain) and
/w/. IF IE /i/ and /u/ were, at origin, vowels (so that Vi/u was an
analogical expansion of i/u; and j/w were "consonantalized" i/u) then
they would NOT be recoverable in AA since there is NO correspondence
among vowels in the two proto-languages. But, because they are (IE i/j
= AA y/$; IE u/w = AA w), they MUST HAVE BEEN consonantal in IE (and
in Nostratic).

Some IE scholars, arguing from internal IE evidence, have suggesting
that long vowels or diphthongs were phonemic in IE.
IE inherited a reflex of a laryngal/pharyngal spirant from Nostratic
which had the effect of usually lengthening any vowel so we cannot say
that V: is phonemic at the earliest stage of IE since it only
represents V+H.

How can we be so certain? Pairs like IE bha:r- and Arabic bahara are
irrefutable proof that vocalic length in IE is attributable to a
"laryngeal" /H/, that is cognate with AA h and H.

Many of the scholarly disagreements in IE could be avoided if some
scholars would eschew Straussvogelpolitik and acknowledge the
relationship of IE and AA through Nostratic.

Sadly, however, even some Nostraticists cannot see that Ablaut and
vowel gradation NECESSITATE a monovocal system; and a monovocal system
practically necessitates consonantal glides.

It will be interesting to see if the linguists on this list
(non-IEists and AAists: those who have no axes to grind) will agree
with the simple logic of this argument.
Pat Ryan

 PATRICK C. RYAN {} * (501) 227-9947
 9115 W. 34th St. * Little Rock, AR 72204-4441 * USA 
 Veit ek, at ek hekk, vindga meidhi, naetr allar niu, 
 geiri undadhr... a theim meidhi er mangi veit
 hvers hann af rotum renn.' * (Havamal 138)
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue