The Great Proto-Indo-European Thread

[KathTheDragon]

Probably postpositions-turned-clitics. I'm not sure if any of them have good IE etymologies, but it seems likely that at least some do.

[2+3 clusivity]

The traditional reconstruction of PIE phonation in stops is unvoiced, voiced, and breathy-voiced. Why couldn't the breathy voice have been slack voice instead? Are there any examples of languages with an unvoiced/voiced/slack-voiced distinction?

Someone -- weeping elf, I think -- posted a link to a various of various languages with three voicing/etc possibilities cross-linguistically and how, in that light, likely the non-glottalic reconstruction is. Looking back, I am not seeing it. Maybe it's in L&L.

[WeepingElf]

The traditional reconstruction of PIE phonation in stops is unvoiced, voiced, and breathy-voiced. Why couldn't the breathy voice have been slack voice instead? Are there any examples of languages with an unvoiced/voiced/slack-voiced distinction?

Someone -- weeping elf, I think -- posted a link to a various of various languages with three voicing/etc possibilities cross-linguistically and how, in that light, likely the non-glottalic reconstruction is. Looking back, I am not seeing it. Maybe it's in L&L.

I don't remember posting such a thing, but I think I have seen it, don't know who posted it and where.

[kanejam]

What is the origin of the tocharian's secondary cases?

They all appear to be from postpositions attached to the end of the early Proto-Tocharian oblique case which continues the PIE accusative.

Someone -- weeping elf, I think -- posted a link to a various of various languages with three voicing/etc possibilities cross-linguistically and how, in that light, likely the non-glottalic reconstruction is. Looking back, I am not seeing it. Maybe it's in L&L.

I don't remember posting such a thing, but I think I have seen it, don't know who posted it and where.

Try this. I don't know if it's quite what you're thinking of, but it does discuss the PIE stop system typologically as well as diachronically.

[2+3 clusivity]

Try this.

Yes. That is it. Thanks.

[Astraios]

Question: Why is the derivation of ἀφρός “foam” from *n̥bʰrós “cloud” here said to be problematic because of semantic mismatch? Surely seafoam looks a lot similar to fluffy white clouds...

[hwhatting]

Question: Why is the derivation of ἀφρός “foam” from *n̥bʰrós “cloud” here said to be problematic because of semantic mismatch? Surely seafoam looks a lot similar to fluffy white clouds...

Well, although the standard Greek etymological dictionaries (Frisk, Chantraine, Beekes) say so, we can still use our own judgement. The semantic mismatch is not that big if you take "cloud" as the original meaning; it's bigger, when you assume the original meaning was "rain".

[jal]

Wiktionary says "rain-cloud, with semantic narrowing to rain, cloud". It also says "[ἀφρός ‎(aphrós, “foam”)] is usually dropped from consideration due to heavy semantical mismatch". I call bullshit on the latter. As has been pointed out above, cloud -> foam is a rather simple semantic drift. And as we all know, semantic drift can produce about any kind of change in meaning.


JAL

[Neek]

ETYM Meillet BSL 31 (1931): 51f. connected Arm. p'rp'ur 'foam' (which does not belong to <σπείρω>), but the ἀ- is problematic (*h2- would give a- in Armenian), and the *bh presupposed by Greek did not give p'. Not to Skt. abhra- [n.] 'cloud' (because of the meaning), not to ► ὄμβρος because the rule of de-aspiration before resonant is not valid. Not here ► ἀφρίους· ἀθέρας (H.).

It's a muddled mess, to be honest. He'd rather link abhra~ὄμβρος~imber/imbris, because they're semantically related, but even Beekes finds trouble with ὄμβρος as pʰr does not yield br, nor does mpʰ > mb, and the ultima accent would prevail (I'd wager that ὄμβρος is more likely a borrowing from a dialect where mpʰr > mbr).

ἀφρός is a perfect match, but the meaning is unusual. Not because n̥bʰrós is cloud, but because n̥bʰrós is a narrow meaning in IE relating to Dyḗws ph₂tḗr's sphere. There's a consistency among IE languages to keep n̥bʰrós within that sphere: Most IE languages that retain n̥bʰrós want to keep it up in the air. Reducing it from a sense of "stuff in sky" to "stuff from churning sea" with no other meanings inbetween is a bit unusual.

[jmcd]

The intervening forms are likely lost in PII. Think of a word in another language like the English word 'nice': if we lacked the intervening sematic change, we'd be lost for seeing the link.

[Neek]

I am not suggesting that I agree with Beekes 100%, but I understand why he isn't decisive on the link between n̥bʰrós~ἀφρός. I find the form too perfect to deny the relationship, just that the semantic shift from IE is unusual to raise question.

[2+3 clusivity]

Transient grey to white volumetric substance/object: pretty tight conceptual sphere imo.

[McMeatLoaf]

"A very close phonetic match, but somewhat problematic semantic development, I'd rather assume a pre-IE substratal term" /Boutkan

[Salmoneus]

As we've said before, a surprising number of IEists would, given any glimmer of a chance, prefer to assume a pre-IE substratal term...


(the semantic difficulties here appear to me to be negligible. And I don't subscribe to the people-in-the-past-were-aliens hypothesis in which everyone carefully weighed their analogies so as not to illicitly transfer words from the domain of one deity to that of another. (they didn't even take that much care over the deities themselves! they shuffled around and were lost and adopted with remarkable facility, as evidenced by how little theological (in the narrow sense) vocabulary can actually be reconstructed for PIE))

[2+3 clusivity]

So, I just learned of the oddity of celtic plural pronouns, does anyone know of a good publically available source on the proto? My google-fu is failing me an Beekes, Sihler, and Fortson do not seem to have much to say in their reconstruction texts.

Also, if anyone has a good source on Proto-Germanic, is there anything on the source of the /*-t/ in the dual nominatives?

Much thanks from my weary eyes.

[Valdeut]

Also, if anyone has a good source on Proto-Germanic, is there anything on the source of the /*-t/ in the dual nominatives?

It appears that an uninflected numeral two was suffixed to the older dual pronouns. So *wé-dwo > *wet and *yú-dwo > *jut. Similar developments are found in Lithuanian mùdu (m), mùdvi (f) and jùdu (m), jùdvi (f).

Note that there was a pre-PG sound shift where final non-high short vowels were lost. In the process, a *w that became final and post-consonantal was also lost. If the word had final accent, it was retracted to the previous syllable. This change preceded Verner's law, the shift to intial stress, the shift of unstressed *e > *i, the shift of final *m > *n, the loss of final *n leaving final nasal vowels, and *the loss of final *t in unstressed syllables.

The effect of this change can be seen in the pronouns *uns < *unswé < *n̥swé (from earlier *n̥smé) and *unk < *ungwé < *n̥h₃wé by Cogwill's Law (from earlier n̥h₃mé).

See Ringe, From Proto-Indo-European to Proto-Germanic, p. 209 (with a reference to Cogwill).

[jmcd]

If it's pre-PG, does that mean it can be found in Baltic and Slavic as well?

[hwhatting]

If it's pre-PG, does that mean it can be found in Baltic and Slavic as well?

I don't think anyone seriously assumes a Balto-Slavo-Germanic node, there are just some shared innovations that look more like areal features. So if a process is said to be pre-PG, it just means its results can be reconstructed for PG.

[Valdeut]

If it's pre-PG, does that mean it can be found in Baltic and Slavic as well?

Calling the change Pre-PG was to provide chronological information more than anything else. Pre-PG usually, I think, refers to the stage of the Germanic parent language after it split from other branches of IE (with which it may nonetheless share some innovations due to contact) but before PG itself. So a Pre-PG change is by definition included in forms reconstructed for PG. The boundaries of the Proto-Germanic period are obviously quite arbitrary, though. On Wikipedia, this is actually the first change listed under the Early Proto-Germanic period, before Grimm's law (the term Proto-Germanic is often used loosely to refer to a period during the evolution of Germanic, rather than strictly to refer to "the latest common ancestor of all Germanic languages", which would probably be right before the split of East and Northwest Germanic).

I don't think this particular change is found in either Proto-Baltic or Proto-Slavic.

[dhok]

Was bored a few days ago in class and started thinking about the etymology of 'tooth'.

The Greek and Armenian cognates of 'tooth' are ὀδούς and atamn, which demand an initial *h₃, so the usual reconstruction is *h₃dónts. This looks like a present active participle, and it's very tempting to say that it derives from *h₁ed- 'to eat'...but the laryngeal is wrong. There's a dialectal Aeolic form ἐδούς, but it's usually explained as having been contaminated with *h₁ed-. The result is that a root *h₃ed- 'to bite' has been reconstructed only on the basis of *h₃dónts (there are other roots *h₃ed- with other meanings that are much more secure).

So here's the leap of logic that occurred to me: what if, whether by contamination with the *o of the next syllable or by some strange analogical or retained process, *h₃dónts really is from *h₁ed-, but the initial laryngeal has undergone o-ablaut? I recognize that this is basically the Indo-Europeanist equivalent of getting stoned and asking how mirrors can be real, but there might be something here?

[Pole, the]

Maybe a voiceless laryngeal *h1 turned into a voiced one due to assimilation with the following *d?

[hwhatting]

The version I have seen (e.g. that's what NIL assumes) is that the PIE was *h1d-(o)nt-, so the Aeolic form woul be regular, and that in the Classical Greek form the expected *edont- was assimilated to odont-. NIL also sees the Armenian form as derivable form *h1d-, but doesn't go into details.

[WeepingElf]

As I am re-reading some old e-mails Allan Bomhard, Tropylium and I exchanged in 2014 and where we discussed, among other matters, pre-PIE phonology, an idea about the PIE glottalic theory emerged in my mind.

If one goes back to what I call "PIE0", the stage before the Great Vowel Collapse, there are just two ejective stops: *t' and *k'. After all, **p' seems to have missed, and the velar series split into three only later. The probably closest living kin of IE, Uralic, has no ejectives. It may have had voiced spirants to match the Standard PIE breathy-voiced stops, if that is what *ð, *ð' and *x were. But it had no ejectives.

However, the ejective or "emphatic" stops *t' and *k' could have emerged from other obstruents: *t' < *ts and *k' < *q.

Does that make sense to you? Or is it nonsense?

[jmcd]

COuld it be that the other series of plosives initially appeared only in loanwords, from Caucasian langauges for example?

[WeepingElf]

COuld it be that the other series of plosives initially appeared only in loanwords, from Caucasian langauges for example?

I can't say it could not be, though I don't feel that is likely. However, I do think that the transformation of whatever system Proto-Indo-Uralic (or whatever) had into the "Glottalist PIE" system happened under areal or substratal influence from a NWC/NEC/Para-NWC/Para-NEC/whatever language group.