Re: On the ordering of some PIE rules

--- In cybalist@yahoogroups.com, Piotr Gasiorowski <gpiotr@...> wrote:

>
> On 2007-01-28 18:50, tgpedersen wrote:
>
> > Now it's my contention that immigrated Iranian-speakers and their
> > local followers in Galicia?/Przeworsk?/Thuringia? set themselves
> > apart from the not-yet-converted locals by the Iranian habit of
> > spirantizing stops before consonants and that in mutual
> > competition and phonological ignorance they tried to outdo each
> > other, thereby extending the application field of that Iranian
> > (but ultimately old-school PIE) rule.
>
> Why didn't they spirantise the *d series as well?

Oh, I haven't got to that one yet. The main thing is to find an
environment where it would have had a *t series allophone. I'm working
on it.

> In Iranian, spirantisation affected _all_ stops.

I know; I may have presented this too early, before I figured out what
environments produced the allophones which became the Germanic
outcomes of the voiced and voiced aspirates, not to speak of Vetner.
That will be tough!

> And if they were so intent on out-Iranianising the Iranians, why are
> actual Germanic loans from Iranian, such as *paTa-, so boringly
> straightforward instead of ending
> up as *faTa- or the like?

Since it is also Slavic, it is tempting to ascribe it to a late
influence, when spirantization (I mustn't say Grimm) had run its
demographic course. Count in *gatu- "gate, street" (Iranian, Baltic,
Baltic Fennic, Germanic) there too.

> > This first Germanic sound shift, as it turned out to be, set them
> > socially apart from the locals in the NWBlock area as a mark of
> > their superiority (as the second Germanic sound shift still does
> > in those parts), whereas in the former Celtic area of Southern
> > Germany those skalkaz they would feel superior to were those Celts
> > who had learned the new Lingua Franca from scratch, with minimal
> > influence from their too distantly related mother-tongue, so that,
> > in order to maintain the social distance, members of the New Order
> > had to perform the same trick on their own sociolect over again,
> > which became known later as the second Germanic or High German
> > sound shift.
> >
> > Now, to get back to your question, those Celtic loans will have
> > entered Germanic as low-status words at a time where both shifted
> > and unshifted sociolects of Proto-Germanic existed. Their
> > high-status equivalents, with Germanic sound shift, have survived.
>
> Funny how you continue referring to "the shift" after allegedly
> having done away with it.

I'm trying to please the diehards ;-) No, seriously, after the
generalization process, for which I haven't yet a name, had run its
course the *result* was the Grimm shift.

> Well, there's no getting round the fact that Germanic obstruents
> correspond regularly to extra-Germanic ones. If loans from
> Celtic split into "high" and "low" variants, why don't we find a
> similar stratification in the inherited vocabulary,

But we do! That's what all the fuss about NWBlock was about, and as I
gave a few examples of, some of them may be "low Iranian". Most likely
though, the laeti of Northern Germany provided the bulk of the
unshifted vocabulary, when they, sufficiently terrorized by wild hunt
attacks, became Germanic lower class.

> and why did no single Latin loan undergo Grimm's Law? The
> traditional explanation is clear: loans from Celtic are on the whole
> older than those from Latin and _some_ of them were borrowed before
> GL applied.

And how does my idea fail that test? You want GL to apply on a fixed
date, I want it to be gradual over an interval. In both cases there is
a before and an after.

> >> And how can your "allophone rearrangement" hypothesis be tested?
> >> Does it make any predictions different from the standard theory
> >> of GL as a Germanic sound change?
> >
> > I made it with the intention of simplifying the existing set of
> > rules so if it does predict a different outcome, which it might,
> > it is not intentional. Therefore I think it should judged on the
> > lesser criterion of Occam instead.
>
> Isn't the traditional scenario more parsimonious? It makes no
> untestable assumptions about PIE allophones.

No, it makes the untestable assumption that there weren't any. Real
languages are not like that, as any speaker of Polish will tell you ;-)

from Burrow: The Sanskrit Language
(on the second palatalization)
"
To begin with the distribution of palatalised and non-palatalised
forms must have depended entirely on the nature of the succeeding
vowel, and consequently an alternation between the two must have been
active in the paradigms of noun and verb. In the parallel
palatalisation of Slavonic such alternation exists, e.g. between
vlUkU'wolf' nom. s. and vluc^e voc. s., and between peko, 'I cook' and
pec^etU 'he cooks'. Since the natural tendency of linguistic evolution
is to smooth out such irregularities (as is done later in Russian,
etc.) it is likely that the Slavonic palatalisation did not long
precede the beginning of the literary tradition. On the other hand in
the case of Indo-Iranian the change had taken place early enough for
the working of analogy to become widely effective. Variation in the
paradigms of noun and verb after the Slavonic style has been
eliminated, except as between vocalic and consonantal suffixes (loc.
s. va:cí, loc. pl. va:ks.ú). On the other hand the alternation remains
active between different nominal derivatives (bhoga- : bhoja-, etc.).
At the beginning of a root alternation between palatal and guttural
remains active only in the case of a few roots as those quoted above.
Mostly it is eliminated, and in this respect Sanskrit shows a greater
tendency to innovation than Old Iranian ; cf. Skt. akar 3rd sg. root
aor., Av. c^o:r&t_ (early Aryan ac^art < ekert), and agamat, a-aor.
Av. jimat_. In the latter case the proper name Jamadagni- ('who goes
to the fire') preserves the earlier, pre-Vedic form. In the
reduplication of the perfect, etc., the alternation always remains,
based on the fact that the vowel of the reduplicating syllable was
originally e (jagá:ma < gWegWo:me, etc.).
"

> And I'd like to repeat my question, slightly rephrased: does your
> hypothesis explain anything that the traditional account doesn't?

Well, it does make the strange PIE assibilation *-tt- -> *-tsts- ->
*-tt-/*st-/*-ss- part of something bigger and more natural. It
rectifies the strange sequence of first RUKI, then assibilation where
the latter applied over more branches that the former, and it supplies
a motivating force (regularization) for several IE soundshifts
unmotivated since Grimm (and this time it's not Germanic 'muth und
stolz', but possibly Iranian ditto).


Torsten