Greek question

Patrick C. Ryan proto-language at
Sun Feb 7 21:00:17 UTC 1999

[ moderator re-formatted ]

Dear Peter and IEists:

-----Original Message-----
From: Peter &/or Graham <petegray at>
Date: Sunday, February 07, 1999 11:54 AM

>Pat said:
>. Sturtevant did mention the
>>Greek forms with /kh/, <*gh.

>That leaves three questions unanswered:
>(a) Was Sturtevant prioritising the Greek evidence over the Sanskrit?

In my opinion, no. I do not think, and I think he did not think that the
Sanskrit and Greek evidence conflicted with regard to the final element <gh>
of the earliest, non-compounded root(s): *negh- / *neugh-.

>Or was he suggesting two forms (*nokt- and *negh)?

My reading of Sturtevant is that he was clearly supposing a root of *negh-
for Hittite nekuz but, upon consideration, he may have been supporting
*negw-, viz. *negh-w-.

>Neither is satisfactory.

I agree that *nokt- is not satisfactory.  But, on the basis of *neuk-,
'dark', I believe it likeliest that there were two basically equivalent
roots: *negh- and *neugh-.

However, it also would not greatly trouble me if we reconstructed *negh-
with an optional -w- extension, *negh-w-, which, in the cases of Greek and
the root *neuk-, was metathesized into the root. However, on the basis of
Egyptian nHzj, 'Nubian', which I believe can be analyzed as nH, 'black' +
zj, 'man', I reject a reconstruction of *negw- or *negwh-  since these would
have to correlate with Egyptian nS (hook-s) or nX (bar-h).

[ Moderator's interjection:
  There is a problem of ASCII transcription here:  The <w> which has been
  written in this form by a number of people should be understood as a super-
  script.  The stem in question contains a labiovelar, and there is no question
  of "an optional extension".

  Cowgill's Law:  In Greek, *o > u in the environment of a labiovelar.  Compare
  for example _kuklos_.
  --rma ]

>(b) How do we explain the Greek vocalism if the PIE form is *negh?   At
>least *negwh would give us a mechanism for the /y/ vowel (< /u/)

See above.

What is a "/y/" vowel? Do you mean the Greek /u/?

[ Moderator's comment:
  Yes.  The IPA transcription, including ASCII IPA, for a high front rounded
  vowel is <y>.  Greek /u/ > /y/ in historical times.
  --rma ]

>(c) Can we reconcile the Skt -kt-, which argues against the voiced aspirate,
>with the Gk -kh -, which implies it?

In my opinion, yes. I believe that Sanskrit <k> can represent either IE <*k>
or <*kh>.

[ Moderator's opinion:
  There is very little good evidence for voiceless aspirates in Indo-European.
  --rma ]

>The suggestion that **negh-t- > *nekt- before Barthomolae's law is up and
>running in I-I crashes into the problem that the past participle forms in -tos
>appear widely in attested IE, and so must also be fairly old forms.

The Egyptian evidence (*nH) suggests to me that the Nostratic form did not
include the -t- (or the -w-).

>We would have to argue that the past participles were still perceived as
>root + -tos, at a time when *nekt- was no longer perceived as *negh+t-.

Not sure I follow this.


[ Moderator's comment:
  In Sanskrit, past participles of roots ending in voiced aspirates are the
  very best evidence for Bartholomae's Law:  budh+ta- -> buddha-.  But if the
  voiced aspirates live that long in the history of Indic, a form **negh^wt-
  should give *nagdh-, not the attested nakt-.  The alternative is to assume
  that past participles are still a live formation, which assumption is not
  borne by the evidence of re-formulated past participles whose relation to
  their roots has been obscured by phonological developments.
  --rma ]

More information about the Indo-european mailing list