From: Miguel Carrasquer Vidal
Message: 56284
Date: 2008-03-30
>----- Original Message -----As I said, I don't have **o at that stage.
>From: Miguel Carrasquer Vidal
>
>>I have some questions about your theory.
>>
>>Starting with *a *i *o *u
>>then assuming for example
>>stress lengthening,
>>we reach your pre-PIE system
>>
>>*a *a:
>>*i *i:
>>*o *o:
>>*u *u:
>
>That would be:
>
>*a *a:
>*i *i:
>*u *u:
>
>>Then you suppose
>>*a > PIE *e (ok)
>
>**a, **i, **u > *e, under the stress.
>Miguel.
>==========
>If we start with
>PIE stage 1
>*a *i *o *u plus stress
>Rule 1 : stressed becomes longNo. It's lexical. Some roots have long vowels, some short.
>PIE stage 2
>*a *a: *u *o *o: *i *i: *u *u:
>stress irrelevant because
>length and stress are the same
>Relevant examples :Something like that, except that the alternations are
>a: podos ~ pedis
>u: nuktos~ noctis
>i: ke:rd ~ sir
>What is the example for o ~ o: ?None.
>Rule 2 *a: and *o: fused ; *i: > *e:Except there is no short *o. That explains the asymmetry *o:
>*a splits into *aH2 and *eH1
>PIE stage 3
>*a/e *o: *u *i *e: o
> *o but not *e: > *e.Rule 4 does not apply anywhere. Celtic and Italic _have_
>(Maybe Tokharian steps out now)
>This might be the last really common stage.
>
>Rule 3 *o: and o fused.
>PIE stage 4
>*a/e *o *u *i *e:
>
>Rule 4 all inherited vowels become short.
>This is Western PIE (Celtic + Italic)
>Rule 4 does not apply to Anatolic and Central PIE.
>I think you mean that. Right ?*i and *u have a stronger tendency to get reduced. This
>
>Short vowels were unstressed in PIE1 and PIE2
>When did the re-stress of short vowels happen ?
>I can't figure out what's going on during PIE3 and PIE4 ?
>I don't really understand what happens with *i and *u
>after PIE 2 ?
>
>ARnaud
>
>===========
>
>Except in morpheme initial and final position:
>**-a > *e
>**-i > *ye ~ *i
>**-u > *we ~ *u
>
>========
>I need to first understand *i and *u after PIE2.
>Arnaud
>======
>
>>*a: > PIE *o(:) (pod/ped ?)
>>*i: > PIE *e: (Cf. H2i:kwr.)
>>*u: > PIE *o(:) (Example ?)
>
>>nukt / nokt "night" ?
>
>**nú:gt- > *nókWt- under the stress; the oblique e.g. Gen.
>**nu:gt-ás, should have given *n.kWtés, and *n.kWt- is
>attested in Sanskrit and Germanic, but for the most part we
>have analogical *nokWtes or *nekWtes.
>The most interesting form is Greek núx, nu:któs.
>Cowgill's law states that in Greek PIE *o > u between a
>labial and a resonant (Examples from Sihler: *nokWts > núx
>"night", *bholjom > phúllon "leaf", *molah2 > múle: "mill",
>*h3noghWs > ónux "nail", *nogWnos > gúmnos "naked", *morm- >
>múrme:x "ant", *h3noh3mn > ónuma "name", *kWetwores >
>písures "4"). The formulation as it stands cannot be
>correct, or Sihler would not have included it in his chapter
>on Phonology (but rather "Phynology"). I believe that in
>most of the examples given, Greek /u/ reflects an original
>**u:, kept distinct from **a: > *o.
>========
>I think I'm ready to agree with that.
>Arnaud
>==========
>>why should *i and *u not be kept
>>as *i and *u ?
>>Why should they be lost > zero ?
>>On account of what ?
>>I would just keep *i and *u unchanged.
>
>Tell that to the Slavs.
>
>We are discussing PIE2 and PIE3...
>Arnaud
>============
>In fact, stressed *i and *u are not lost (see above), theyYou can find a lot more in the list archives.
>merge with *a (except at the edge of morphemes, and except
>for labializing and palatalizing effects on neighbouring
>consonants).
>
>Unstressed *i, *u are lost, like unstressed *a.
>The only peculiarity is that long *i: and *u: were
>apparently also lost in unstressed position, unlike *a:,
>which, though shortened, remained and attracted the stress
>(Gen. *pa:d-ás > *ped-és > *péd(e)s).
>=============
>I don't understand that for the time being.
>Arnaud
>============