Re: Odp: Odp: A few PIE questions

From: Piotr Gasiorowski
Message: 289
Date: 1999-11-18

 
----- Original Message -----
Sent: Thursday, November 18, 1999 12:21 AM
Subject: [cybalist] Re: Odp: A few PIE questions

About PIE infinitives:  Does this mean that if you were to say the sentence 
"I went to the restaurant to eat" in PIE, the infinitive "to eat" would be 
expressed in the form of an abstract noun in the dative case rather than the 
verbal root with the abstract noun dative ending?
 
The verb and the abstract noun were derived from the same root, but their stems could look rather different because of differences in their vocalism, stress pattern and suffixation. The important thing is that the noun was not derived directly from the verbal stem but was an independent, albeit related formation. Its incorporation into the verbal paradigm took place independently in the daughter languages. It was also at that late stage that deverbal nouns corresponding to morphologically complex verbs (denominals, causatives, etc.) began to be formed in a productive fashion.
 
And if it is the first 
choice, are there any suffixes added to a verbal root or particular gradation 
of vowel length that are common to PIE when forming abstract nouns.  Also, 
did they tend to be a particular gender (like in Greek, where feminine is 
predominant)?
 
In Vedic alone there were 16 different formations corresponding functionally to the later (Classical Sanskrit) regular infinitive in -tum. Most of them can be traced back to PIE; they were datives (most often) but also accusatives, locatives or ablatives of abstract nouns. In non-Anatolian IE root (that is, uxsuffixed) abstracts and formations in *-ti-, *-tu-, *-ta:- were indeed feminine, but there were also neuters in *-es-, *-onom (as in Germanic). Hittite had neuter deverbals in -war/n- and -tar/n-, formed very productively for all sorts of verbs.
 
Different suffixes imposed different restrictions on the form of the root to be used with them. For example, *-ti- was inherently stressed and was added to roots in the zero grade (like *mn-tis from *men- 'think'), and root abstracts had the o-grade (or so it seems to me at the moment). However, most of the productive formations in post-PIE times abandoned apophony as an unnecessary burden and simply copied the vocalism of the present tense of the verb (the case of Germanic and of the first conjugation in Latin).
 
Also, how would you express an infinitive with the verb "to want, wish"?  It 
doesn't seem to me you could use the abstract noun formation, so would you 
have to use the optative mood to express the idea of wishing or wanting 
rather than a verb meaning "to want" plus an infinitive?
 
 
Well, that's what the optative is for. But I wouldn't exclude the possibility of employing a deverbal noun with verbs like 'want'. The very fact that in so many languages infinitives historically derived from deverbals are used in this manner suggests that the construction is not unnatural. 'Want' may take a noun complement in the accusative -- why not an action noun? With datives, the semantics was perhaps something roughly like this: 'I exercise volition for the sake of an action'. Compare this with the development of going to in English: the construction I am going to die originally meant 'I'm going to (the place of) my death'.

<<One almost foolproof criterion for telling thematic from athematic verbs is 
the form of the first person singular. The ending of thematic verbs is -o: 
versus the athematic -mi.>>

This may be a stupid question, but are you referring to the daughter 
languages here as the foolproof criterion?  And, since Greek and Sanskrit 
display both types of endings (as opposed to say Latin where there is only -o 
verbs), are Greek and Sanskrit the main daughter languages?  And if so, does 
Sanskrit take precedence over Greek if there is conflicting information (for 
example the case of PIE *bher- where the Sanskrit form is bherami and the 
Greek form is phero)?  Of course, *bher- is an athematic verb, but I just 
wanted to make sure about Sanskrit taking precedence over Greek.
A very good question. Sorry for the lack of precision in my hasty answers. I was referring to PIE reconstructions, not to any individual daughter languages. Neither Greek nor Sanskrit should be regarded as 'the favourite daughter'. As far as well-attested IE verbs are concerned, they mostly agree and sometimes disagree. In the former case it is likely that the agreement goes back to PIE, and in the latter you have to look for further evidence elsewhere. To make matters worse, the PIE verb system was already characterised by 'polymorphism', that is, there were whole constellations of related verb stems and the same root could have thematic and athematic variants as well as more complex derivatives.
 
By the way, Sanskrit has no direst reflex of *-o: in the first person sg.; it has -a:mi, which is an Indic innovation, reslulting from the analogical addition of -mi to the original ending. I have to correct my 'foolproof criterion' if it is to work for Sanskrit: the thematic class has -a:mi, and the athematic class has -mi there. Note that there is no disagreement in the case of *bhere-, which is THEMATIC in both Sanskrit and Greek. Most other correspondences bear out its thematic character in PIE, only Latin has athematic forms in 2nd & 3rd sg. and the imperatives (fers, fert, fer, ferte).
 
Piotr