From: Glen Gordon
Message: 4907
Date: 2000-12-03
I'm a confused little camper. Concerning the supposed plural *no:mn-x, why does the following link show the Sanskrit plural of /na:ma(n)/ to be /na:ma:ni/?
http://pauillac.inria.fr/bin/dicdecl?query=naaman&gender=Neu
Before you jump the gun and tell me "But it ends with /-i/, silly. Therefore *-x is attested!", I'm wondering about the extra length in the second syllable that you might help to explain for me. For all the world, it would appear to me to be a suggestion of IE *nom�:n (from an earlier **nom�nx just like *wodr & *wedo:r) where Sanskrit /-i/ has been secondarily added. Yo wassup, homey?
Miguel stated:
>You told me, but you failed thus far to give any examples or any
>precise rules. So what's your view on the poim�:n/d�imo:n thing?
I've just given you a long list of precise rules, Miguel. If you refuse to pay attention to what I'm saying, the discussion becomes very stale and frustrating.
The poim�:n/d�imo:n thing? Greek "pastor" and "demon"? Or are we talking about IndoEuropean itself? Please, let's only deal with IndoEuropean and not random items in any IE language. What's the IE relevance to poim�:n/d�imo:n?
Second, I feel no shame in presenting what you call "imprecise" rules, or as I like to put it, "non-megalomaniac" rules. My strategy is to try to uncover _realistic_ developments in preIE stages rather than attempting to find all-encompassing rules that are doomed to fail. Realistic linguistic changes often involve a main development followed by some modifications and reshufflings that obscure that main development.
I have provided some of these main preIE rules like "MidIE penultimate accent" and now "conjugational vowel harmony". They serve to explain much of why IE is the way it is, just like Grimm's Law serves its own purpose. They cannot be expected to solve everything. That's what Verner's Law is for. :)
Miguel confusingly states in two seperate areas of your last post:
>The **a: in a palatal environment (*g^) failed to back, and developed
>to *e:, distinct from short *e (< **a).
>[...]
>The root had a long vowel (**wa:dn-). Under the stress, this >develops to *o (*w�dr). When unstressed, **a: is first shortened to >**a, and then develops normally to *e (*wedn�s). What's the problem?
There are three very large problems that defy probability:
1) You are heavily dependent on the unlikelihood
that *g^ is a very ancient, seperate phoneme
from IE *g and not caused by Late IE *e.
2) You reconstruct *a: unnecessarily when no
evidence suggests that this vowel should
be vrddhi.
3) There is no order to the development of **a:
which is free to become either a short OR
long vowel without solid explanation.
First of all, if it is true that **a: > *o, except in the environs of an already existant *g^ where it becomes *e:, as you have now claimed, you are stating that *e(:) does not cause allophony in velars (*g > *g^). Either you're saying that the velar palatalisation is a much earlier development than usually viewed by IEists, involving a very tentative **a: instead of *e(:), or you're saying that it never occured at all:
( **ra:g > ) **ra:g^ > IE *re:g^
In all likelihood, *g^ is merely a secondary item, a palatal allophone of *g in the presence of front *e. I think most IEists would accept this as a late development, otherwise we would expect much more ancient "satem" languages than what we actually have. If you are saying that palatalisation is before the existence of *e, you must be saying that first *a: caused palatalisation throughout, then this palatalisation was retracted in places where *a: > *o (VERY UNNECESSARILY COMPLEX!). Then again, maybe you're saying that *g^ and *g are actually two seperate phonemes which is totally senseless altogether.
Second, please provide the exact sound changes for your preIE **a if **a: is supposed to have created all of *o, *e:, *e and *a (??!). Is there short preIE **a at all? What twisted preIE phonological system do you have in mind, exactly? I really can't help but believe that you're just making it up as you go along. If *re:g^ is supposed to be long throughout the paradigm and yet with mobile accent, we should expect that your **a: sound change contradicts itself completely, becoming SHORT *e in the *re:g^- paradigm just as **wa:d- becomes *wed- when unaccented! Make up your mind.
Very ad hoc.
Third and more immediately irrational, there is no rhyme or reason to why it is that accented **a: should become short *o, except before *g^ when it remains long and fronted...?? Excuse me? I'm sorry, there must be a bad connection over here cuz I'm hearin' alot of static but no substance. You're going to have to explain yourself out of this incredible fantasy of yours.
Excuse me if your use of **a: as an answer to every IE problem seems like a loosely-knit, somewhat mad solution evocative of Patrick Ryan's excursions into the wild chaos of superficial thought. Mr Ryan is of course author to this wonderful site with personal reconstructions of Proto-World:
http://www.geocities.com/Athens/Forum/2803/IE-VerbalInflection.htm
... yixes! Get out the white coats, eh?
What I provide presents no fundamental problems. IE *e simply derives from Mid IE schwa and thus could not have palatalized the vowel until Late IE (5000 BCE or later) when the vowel fronting/backing occured. IE *g^ is still to be viewed as an allophone of *g, as is most commonly believed (Why rock the boat?). Further, my Early and Mid IE phonology is pretty much the same as Late IE. It's just straightforward and as simple as I can make it.
IE *re:g^ simply comes from an earlier form *re:k: with schwa *e, and a product of ancient monosyllabic lengthening. Everything is fully explainable and direct. Similarly, *wodr is explained in a simple fashion. No need for *a: where there is no vrddhi, no need to ignore conventional views about IE phonology or phonology in general, no need for next-to-impossible *sw- > *y sound changes. (:P)
You might be confused about one thing though. I view preIE stages to have the following three series of stops and I write them accordingly in Early and Mid IE:
voiceless lenis *t (IE *t)
voiceless fortis *t: (IE *d)
voiced *d (IE *dh)
I'm sure you've seen my yet-to-be-revised site at:
http://glen-gordon.tripod.com/language.html
It says some more on the subject.
>>It's far better explained with alternating short vowels *a and *e >>(schwa) on the root *wat:-/*wet:- (later *wod-/*wed-) which is >>evocative of other alternating *o/*e patterns, by the way.
>
>Well no, that's merely begging the question. My aim was rather to
>_explain_ *e/*o ablaut.
Rot. It is so explained. The earlier *a/*e ablaut is merely related to accent. The *a becomes schwa when unaccented, just like the "a" in English "sofa". Nothing magically complex here. How much elaboration must I provide for you? If your aim was to explain it, then what's holding you up? :)
>Yes, in the mediopassive, which has *o-vocalism throughout the >endings [pl. -wasta(ri), -duma(ri), -anta(ri)]. We don't find that >in the perfect, which has *-e (except before *h2, of course) [sg. >*-h2-e, *th2-e, *-e; pl. *-m-�, *-t-� (*-h1-�/*-s-�), *-r-�].
Yes! The mediopassive was secondarily developed out of perfect and non-perfect conjugation. The vocalism of the endings is primarily *-a in the perfect, initially causing the a-grade (and then o-grade) of stative verb stems. The perfect plural consists in reality of non-perfect endings and can't be relied upon for the original vocalism of the other endings. Dividing *-m� into **-m-� is senseless since it clearly derives from the non-perfect *-m�s where the *-me- part is very much atomic and related to the 1ps pronoun.
The mediopassive conjugation appears to have taken the perfect *-a and attached it to the non-perfect endings like 3ps *-t with an added *-r. This initially produced 3ps mediopassive *-t-a-r. It obviously must have occured before the shift of *a to *o in most positions, producing a later *-to(r). This sound change is already supplied in my list in the previous post, showing that unlike Miguel, I don't make things up as I go along. (Double :P)
Further, I think that *-r was originally a meaningful ending necessary to the mediopassive endings but through time, *-r became an unnecessary meaningless phoneme since *-to distinguished itself from *-t just fine without the need for *-r. The original semantics of the suffixal collage in the mediopassive were now forgotten.
>>The 3ps may not have originally been *-e. Afterall, there is clearly >>an automatic semantic link between the 3ps ending (whether *-e or >>*-a) and the pronominal *e which can offer confusion.
>
>The ending *is* pronominal *-e (a.k.a. the thematic vowel). In the
>third person perfect, we sporadically find (pronominal?) *-s instead
>(Hitt. 3p.sg. *-s; Skt./Avest. 3pl. *-rs).
Yes, I'm fully aware. But I mentioned a possible later confusion with pronominal *e (or even *se) and you have in turn mentioned more possible contaminations via an association with the thematic vowel *-e- that clearly doesn't exist in the perfect (cf. *woid-xa) but does in the indicative. Thus we have two good motivations for an earlier 3ps *-a to become *-e (via pronominal *e AND thematic *-e-). Sweet!
Thus 3ps *-se is truely secondary (that explains Tocharian) and I can sleep at nights now.
>> - *-�x -> *-�:
>> *dwex "two" > *dwa:
>> *(e)kWtwex "eight" > *(e)kWtwa:
>
>So what about *-ah2 in the feminine?
A later attachment of collective *-x. With *dwex and *(e)kWtwex, the *-x was attached _permenantly_ to the stem and didn't come off. The declined stems were *dwex- and *(e)kWtwex- WITH *-x (IE *dwo:- and *okto:-).
>> - e/o ablaut (unaccented *e > *o)
>> - resistance in vowel-harmonious conjugation
>> where e-grade is non-stative
>> and a-grade is stative.
>> - plural *-es remains by analogy
>> with *-mes & *-tes.
>
>Which in turn remain because of? And what about *-men and *-ten?
Endings *-mes/*-tes remain because of the conjugational vowel harmony (e-grade non-stative vowel harmony) that I was so careful to mention on the previous lines to no avail. The suffixes *-men and *-ten are very ancient variants (Mid IE *-m�ne and *-t�ne) and were formed by imitation of the 3pp *-�ne (> *-�n-t). There was never a plural **-n.
>> - accent regularisation in athematic/thematic
>> - thematic gets initial accent
>
>No it doesn't. There are whole categories of nominal and verbal
>thematic formations with the accent on the theme vowel (Grk. phor�s,
>Skt. tud�ti, etc.)
>
>[...] OK, for <phor�s> "carrying" above, read *<snus�s> >"daughter-in-law".
In general, we do so have accent-initial thematics (*bhere- "to carry", *wlkWo- "wolf", et cetera ad nauseum infinitum). The lesser exceptions are a later development. I believe the List talked about things like /tud�ti/ earlier in relation to aorist formations and its accentuation. I still have yet to settle exactly how this developped, to be honest. I do know however that the accent of *snus�s must be a rather late item since accent differences between "adjective" and noun blurred together later on. Clearly, *wlkWos proves that the accent was secondarily placed on the reduced initial syllable and shows the hopelessly entwined relationship between noun and adjective in IE.
>> - athematic gets accent on
>> non-nominoaccusative case endings
>> - adjectives keep old accent
>
>Unh? So now you agree with me that adjectives _did_ exist?
Well okay, it depends on what you mean and when. In general they didn't. The adjectives are a late category in IE at best. The declensional endings are clearly taken from other semantic categories which shows that this must have been a new and underdevelopped concept for IE speakers. A language can live without a formal category like "adjective". There are so-called ancient IE "adjective" stems like *sen- "old", for instance, but no one can say for sure whether they were truely adjective stems or whether they derive from qualitative verbs or even from ancient noun stems (Note the related Etruscan /san-/).
>> - *-om -> *-o:
>> - genitive in *-o: created (later instrumental)
>
>So what about the ins. in *-e:?
An ablaut variation by analogy with the other already-existing ablauted endings like genitive *-es/*-os, of course.
>> - *ego: replaces *meu
>> - meaning "I am here"
>> from *e-ge- (*e "this") plus them.1ps *-o:
>>
>> - indicative *-i established from affixed *ei "here"
>>
>> - 1ps them. n-i. *-o: reestablished as *-om
>
>I assume you mean "present" instead of "indicative". Why don't we
>have present *-o:i; past *o:, then?
I don't think any textbook will tell you that IE was a tensual language. The non-indicative (lack of *-i ending) was used not only when conveying a past action but also a hypothetical one (without reference to time). In this way, the non-indicative *�sm might be much like English "you were" which can be used to mean the past and yet also hypothetical situations like "if you were". No, *-i is not a "present" marker.
We would expect 1ps thematic indicative **-o:i, wouldn't we? Here's a good theory. Remember *-mes and *-tes? They obviously must have lost the indicative *-i too because of their strong association with the plural *-es. What's up? As the indicative shortened to *-mes/*-tes, the non-indicative became *-me/*-te to compensate. Likewise, by similar analogical processes, *-o:i was reduced to *-o: while the non-indicative must have temporarily compensated to *-o (later affixed with non-thematic *-m).
Here's the quick-reference low-down:
indicative analogical
pre-indicative established alterations
---------------------------------------------------------
1pp *-mes *-mesi/*-mes *-mes/*-me
2pp *-tes *-tesi/*-tes *-tes/*-te
1ps *-om > *-o: *-o:i/*-o: *-o:/*-o(m)
All solved, the gLeN way! Is the populus pleased?
- gLeN