Three-Way Contrast of Secondary Articulations in PIE

Stanley Friesen sarima at
Sat Mar 24 14:32:28 UTC 2001

At 10:19 AM 3/22/01 -0600, proto-language wrote:

>I do not doubt that what you have described is natural for you and for
>speakers of many languages in which contiguous vowels are present.

>However, IE as it is normally reconstruted, did not have contiguous vowels so
>the real process you describe seems to me to be irrelevant.

That is not strictly so.  Look at one form of the genitive of u-stems:
*-uos.  Since the /u/ is clearly a vowel in the other case forms, this
almost certainly was a vowel-vowel contact in the pre-proto stage at least.

>Furthermore, whether /i/ and /u/ derive from earlier /y/ and /w/ in avocalic
>environments or not; and, if not for some instances at least, have separate
>phonemic status in earliest IE (which is what Rich is suggesting), no one
>to my knowledge has ever challenged the phonemic status of /y/ and /w/.

I suppose I would not either, though I see few independent occurrences of
these sounds - where they do not seem derived from older vocalic /u/ and
/i/.  I can think of *yeug.  (Still that one alone is probably enough to
establish at least /y/ as a phoneme rather than an allophone of /i/).

>I have no axe to grind by denying IE /i/ and /u/. It is just that, where I can
>find AA cognates for IE roots, a medial or final /i/ or /u/ in IE shows up as
>a consonantal /y/ or /w/ in AA. I have to conclude that all IE /i/ and /u/
>derive from avocalic /y/ and /w/.

Does proto-AA as currently reconstructed have vocalic /i/ and /u/?  If not,
I would suspect that the same error has been made here as in traditional
PIE reconstructions.  Given the 'triliteral root' tradition in the Semitic
languages, I would think there would be a considerable bias in this direction.

>Finally, if /i/ and /u/ had separate phonemic status in earliest IE, one might
>expect them to be part of the Ablaut process; e.g. raising /e/ to /i/ for some
>grammatical nuance. Or, if /i/ and /u/ are phonemic vowels, what would be the
>zero-grade form of a Ci/u(C) syllable?

The same as the zero grade of a CeC syllable - CC.

>Must we say that /i/ and /u/ are entirely outside the system?

No, but e/o ablaut is rather odd and special, so there is no reason to
suppose it would have applied to /i/ and /u/.  And other than that the
earliest layer really only shows full and zero grade.

May the peace of God be with you.         sarima at

More information about the Indo-european mailing list