The Lost Vowels of Pre-Etruscan Syncope

It's been a while since I've wrote about issues concerning Aegean or Etruscan linguistics. However, lately the issue of lost vowels in Pre-Etruscan Syncope sprang to my busy mind.

Many suffixes that are consonant-final in Etruscan appear to come from vowel-terminating suffixes at some Pre-Etruscan. In fact the vowels were probably lost at a stage when Rhaetic, Lemnian and Etruscan were still the same undifferentiated tongue. So far, I think I can ascertain what some of these lost vowels were in some suffixes and words. For example, the Etruscan derivational suffix -aχ appears to have been in origin *-aku. “How can I be so certain?” you ask? Luckily, alternations exposing the past are still present in Etruscan such as seen in *araχ “falcon, hawk” (glossed as arakos by Hesychius) versus aracuna “(one) of the hawks or falcons” where former *u is preserved when the pertinentive suffix -na is attached. I've also ascertained so far that the intransitive participle -θ was once *-ta whereas the homophonous agentive suffix -θ (as in the names Aranθ and Vanθ) was once *-ti. I could be wrong but that's my theory so far.

Since all vowels seem to equally disappear word-finally during Pre-Etruscan Syncope, I would naturally assume that the transitive participle suffix -u must consequently come from an earlier diphthong *-au. In Minoan, this ending may indeed be preserved as a diphthong in the inscribed word DI-NA-U /'tʃinaw/ “moulded” in inscription HT 16 for example (c.f. Etruscan zinu).

That's where I'm at in regards to Pre-Etruscan Syncope so far but I still have a lot more questions to answer (and it certainly would be nice if an extensive Minoan document floated our way soon in order to make its translation a helluvalot easier).

Determining typical forms behind Semitic verbal loans in Pre-IE

If one looks to Norman French and Middle English as a typical example of intense linguistic contacts between two historical languages in order to understand better the Proto-Semitic and Proto-IE contacts during the Neolithic, one may notice that only a small number of verbal forms in French loans typically surface in English. For example, many verbs were simply borrowed from the presentive form (c.f. French (il) part vs. English to part). However there are also many verbs which were borrowed into English based on French infinitives (c.f. French rendre vs. English to render which fossilizes the infinitive ending in -re).

Given that, I start to wonder if maybe it would be more organized on my part to compare Semitic and PIE verbs according to only a few specific verbal forms. So I've been thinking about how to answer the question “If I were to pick only two Proto-Semitic verb forms as sources of PIE loans, which would I pick that would fit most or all of the data the best?”

Based on the handy Semitic Binyanim pdf, my answer at this point would now have to be: 1) the nominative-declined active participle of the shape *CāCiCu and 2) the nominative-declined infinitive of the form *CaCāCu. This could account for almost all Semitic verbal loans that pop up in Mid IE, if we assume that Mid IE speakers simply ignored vocalic length (i.e. interpreting both PSem *a and *ā as MIE *e), that MIE employed a fixed penultimate accent, and that the rule of Proto-Semitic accent by contrast was that it was to be either placed on the first available non-wordfinal “heavy syllable” (i.e. CV: or CVC) from left-to-right or on the initial syllable by default. Predictably, the irregular essive verb *yiθ (becoming PIE *h₁es-) would be an outlier from this general pattern and “to be” is a rather oddball verb cross-linguistically speaking.

So, I guess I need to update my SemiticPreIEloans.pdf document on esnips to reflect this. I hope that sounds a bit more organized than what I've been saying so far. Little by little, I'm gettin' there hopefully. Cross fingers.

What do I "know"?

Both Mehri wēda and Akkadian wadū (variant of idū) make it uncertain whether it's appropriate to reconstruct *w- or *y- as the first radical of the Proto-Semitic (PSem) triliteral meaning “to know”^[1].

Considering the Proto-Indo-European (PIE) root *weid-, it's just too tempting to wonder if there's a connection. If the original PSem root had *w-, then it would predictably become *y- in Western Semitic languages as is the case with other Iw-verbs like *wθb “to sit”.

The question is: What, if any, PSem form can plausibly account for the semantics and phonetics of the PIE root if this was indeed a Neolithic loanword? So I've been consulting a handy pdf called Semitic Binyanim for a hint at a sturdy answer. One form that may fit could be the active G-stem participle which is reconstructed to be of the vocalic structure of CāCiC-. So in theory, I'd then expect *wādiʕ- “knowing” and this would be one option to explain the PIE root, if correctly formed that is, since PSem *ā appears to correlate with PIE *ei also in the equation of PSem *θalāθi = PIE *treis “three”. I really need to find an in-depth book on Proto-Semitic grammar though...


NOTES
^[1] Bibliographic Bulletin (1982), University of Virginia, p.193 (see link).

A list of possible Proto-Semitic loanwords in PIE

I decided to get organized and produce a pdf that I plan on updating as new information comes forth concerning Proto-Semitic (PSem) loanwords in Proto-Indo-European (PIE). Hopefully, if anything, it will provoke interesting discussion or cause one to ponder on the exact cultural and linguistic relationship between Proto-Semitic and Proto-Indo-European during the Neolithic... or perhaps disgust and ire from the peanut gallery. Oh well, it's worth a try:

SemiticPreIEloans_20080817.pdf

Hosted by eSnips

Genitival Misanalysis

I've been flipflopping in terminology, not knowing whether to call this rule Nominative Misanalysis or Genitival Misanalysis. I think I'll stick with the latter since this also affects inanimate thematic stems which don't have a nominative case seperate from the accusative. This rule involves the faulty segmentation of stem-final *-s and *-m that I believe had happened in a stage of Late IE. This rule helps explain the origin of the declensional system of thematic adjectives which curiously use *-s in the animate nominative singular but *-m in the inanimate nomino-accusative (n.b. Latin masculine bonus vs. neuter bonum). I've traced this odd quirk to the genitive endings *-ós and *-óm, the latter of which probably had originally been used for a seperate case altogether (perhaps once meaning “amongst” or “amidst”) which later acquired a collective nuance (hence its use in inanimates) and then finally a plural sense (hence the ending being used in animate genitive plurals).

At any rate, today I've updated my pdf entitled Diachrony of Pre-IE with more elaboration on the sections in the Late IE period labelled Genitival Misanalysis (formerly Nominative Misanalysis), Development of Adjectival Case System, Acrostatic Regularization and Agglutination of New Thematic Genitive Marker.

DiachronyOfPreIE_2008Aug10.pdf

Hosted by eSnips

All of this for your intellectual pleasure. Enjoy!

Inanimate thematics that failed to be converted in Pre-IE?

In my previous post, I explained that I recently pondered on how Syncope would have affected inanimate thematic noun stems. If I'm right that many animate thematic stems in Mid Indo-European (MIE) were preserved by resisting Syncope due to phonotactic reasons such as avoiding awkward syllable codae, then one should expect that inanimate thematic stems should have behaved similarly.

After pondering on a potential Semitic loan *qárnu "horn" and its likely path of borrowing into PIE, I came to the conclusion that inanimate thematic nouns indeed did survive into early Late IE (eLIE) but that they were endingless in the nominative and accusative. This means that this would have resulted in wordforms such as *kérnə "horn" in the nomino-accusative. I fleetingly mentioned at the end of my last post that if these stems had survived unmolested, Indo-Europeanists would be speaking of *e-stems (i.e. **kérne). Instead however, this thematic stem appears to be reconstructed as *krnóm. This and other inanimate thematic stems exhibit interesting effects such as the curious disappearance of *m in the locative case (e.g. *yugóm "a yoke" but *yugó-i "among a yoke") which leads many to assume that this was a functional ending rather than the product of misanalysis and missegmentation by early Indo-European speakers. So I take it that these endingless stems must have been converted by adding the genitive ending *-óm by analogy with inanimate thematic adjectives but were reanalysed as thematics ending in *-ó- with a pseudo-ending *-m in the nominative and accusative cases. I don't think that the final nasal stop had anything to do with the accusative case ending which was restricted to the animate stems anyway.

Yet, if these stems were "converted", surely the process wasn't entirely complete, was it? Could there be still some evidence of bare thematic stems in the language that failed to be converted? So now I'm starting to wonder whether this might explain the curious shape of the numeral *pénkʷe "5". If this numeral had been originally treated as a collective inanimate noun in Mid IE, its failure to be converted would be natural if it came to be solely understood as a numeral rather than a generic noun. Thus MIE *kérna "horn" > eLIE *kérnə > *kr̥n-ám (conversion) > *ḱr̥nó-m (missegmentation) yet MIE *pénkʷa > eLIE *pénkʷə > *pénkʷe (no conversion in numerals). Word-final schwa in early Late IE is expected to become PIE *-e# since it only becomes *-o- word-medially in the presence of a following voiced segment and this then explains the thematic vocative case which ends in *-e (i.e. the vocative case was endingless causing this Pre-IE schwa to be word-final).

What happened to Pre-IE's inanimate thematics?

This past week has been a little busy (i.e. a new job, looking after my parents' bratty dog, summer tanning, etc.) and so I've been slacking a little on my blog. Mea culpa. Of course, hectic life doesn't stop me from thinking and theorizing about ancient languages. NEVER! It only stops me from typing for a while. It's probably just as well because it gives me time to develop some more elaborate ideas to share. So I just noticed something in the past few days that I had never realized before concerning inanimate thematic nouns.

Over the years, I've slowly come to the conclusion that Syncope didn't obliterate all the schwas in early Late IE (eLIE) and that a few resisted the rule due to reasons regarding proper syllabicity. So a disyllabic wordform in Mid IE (MIE) of the shape *CVC.CV- should have preserved its final schwa because if it had disappeared, it might create problems with the distribution of the consonants in the remaining syllables^[1]. For example, I believe that the PIE word *h₁éḱwo- "horse" had survived Syncope (i.e. MIE *ʔékwa- > eLIE *ʔékwə- and not **ʔéku-). I also concluded thus far that some animate thematic nouns were derived from genitive-declined stems that are an artifact of a stage of Pre-IE when adjectives were once formed by way of the genitive case (either *-ás or *-ám at the time) and did not originally agree with the case of the nouns they modified. Inanimate adjectives were declined with *-ám while animate adjectives were declined with *-ás until they were "misanalysed" by prehistoric speakers as vowel-ending stems with terminating consonantal endings, nominative *-s (e.g. *wĺ̥kʷo-s "wolf") and a (nomino-)accusative ending *-m (e.g. *yugó-m "yoke"). Or at least, so say I.

That being said, I was recently investigating a potential loanword from Semitic, namely the "horn" word. In Proto-Semitic, the word is *qárnu while in Proto-Indo-European it's *ḱr̥nóm. At first, I was becoming disheartened by the apparent incongruence in accent, phonetics, etc. However, when I allowed my mind to explore what might have theoretically happened to thematic inanimate stems during Syncope, I came to realize that my theory had a gap that needed filling out. You see, if one ponders what would happen to a inanimate, rather than animate, vowel-ending stem in MIE during this period of time, we have to conclude that the final schwa should be preserved in the same way as animate stems when of the shape *CVC.CV-. If so, we end up with an interesting grammatical possibility because since inanimate nouns were endingless in the nominative or accusative case, my theory suggests that there should have originally been inanimate nouns ending only in a schwa.

So let's say that Proto-Semitic *qárnu was borrowed into Mid IE. We might rationally expect a form like *kérna, more than anything. Syncope would then fail to affect the final vowel in order to avoid syllabicity issues, thus a bare inanimate wordform *kérnə in the nominative and accusative cases. Obviously, at this point, this is not PIE as we know it but once I traced this interesting hypothesis to this point, I realized that I might be able to fill in the rest of the diachronic development. I think what happened next was that, for whatever reason, bare stems were undesirable to speakers and that once former genitive-declined adjectives were misanalysed as thematic stems, this provided a means to "complete" the awkward wordform with an imaginary "inanimate nomino-accusative" pseudo-ending *-m. Thus, this stem as well as other inanimate thematics must have been converted to this new "genitivized" form, thus *ḱr̥nóm with accent on the final syllable in imitation of genitive plurals. As extra food for thought, if this conversion had not taken place, inanimate thematics in Proto-Indo-European might have ended in *-e instead.

NOTES
^[1] These stem-final schwas are not always preserved and it appears to depend on individual phonotactic considerations. For example, I reason that MIE *ʔékwa- "horse" and *kérna "horn" would likely have resisted Syncope because *kw and *rn wouldn't have been valid syllable codae in early Late IE, the former violating sonority hierarchy and the latter showing adjacent tautosyllabic resonants which would have offended language-specific phonotactic restrictions at the time. On the other hand, MIE *nakʷta- "night" appears to have been reduced to early Late IE *nakʷt- (> PIE *nokʷt-) without interference, but then again, this is natural since the resulting cluster *kʷt was a perfectly valid syllable coda.

UPDATES
(Aug 09 2008) Added an important footnote about the phonotactic conditions behind stem-final vowel loss or preservation.