Zompist Bboard Again

WeepingElf wrote: ↑Tue Feb 21, 2023 9:35 am [I have seen it; I am not convinced. It requires the change q > χ to happen at least three times: 1. in Non-Anatolien IE; 2. in Non-Luvic Anatolian; 3. in Non-Lycian Luvic. Also it is questionable to me to build such a theory on a language which we can read but don't really know for sure how those letters were pronounced. And even if the Lycian reflex of the laryngeals was indeed [q] and not [χ] or whatever, what speaks against a change χ > q in Lycian? Maybe there was a substratum involved which had [q] but not [χ].

All possible. But there are also sounds that are unstable and tend to develop in one direction; e.g. /s/ > /h/ has happened at least three times independently in the history of IE (PIE to Greek, Indo-Iranian to Iranian, Brythonic to Welsh; plus IIRC /h/ is one of the outcomes in Albanian; plus it has happened again to final /s/ in some varieties of Spanish). Spirantization of velar or uvular stops is frequent as dirt, while the reverse fortification is much rarer.

WeepingElf wrote: ↑Tue Feb 21, 2023 9:35 am Kloekhorst is a creative and adventurous scholar, but some of his idea strike me as unjustified.

He's someone who knows his stuff extremely well, so I'd hesitate long before betting against him.

hwhatting wrote: ↑Tue Feb 21, 2023 9:42 am

WeepingElf wrote: ↑Tue Feb 21, 2023 9:35 am [I have seen it; I am not convinced. It requires the change q > χ to happen at least three times: 1. in Non-Anatolien IE; 2. in Non-Luvic Anatolian; 3. in Non-Lycian Luvic. Also it is questionable to me to build such a theory on a language which we can read but don't really know for sure how those letters were pronounced. And even if the Lycian reflex of the laryngeals was indeed [q] and not [χ] or whatever, what speaks against a change χ > q in Lycian? Maybe there was a substratum involved which had [q] but not [χ].

All possible. But there are also sounds that are unstable and tend to develop in one direction; e.g. /s/ > /h/ has happened at least three times independently in the history of IE (PIE to Greek, Indo-Iranian to Iranian, Brythonic to Welsh; plus IIRC /h/ is one of the outcomes in Albanian; plus it has happened again to final /s/ in some varieties of Spanish). Spirantization of velar or uvular stops is frequent as dirt, while the reverse fortification is much rarer.

WeepingElf wrote: ↑Tue Feb 21, 2023 9:35 am Kloekhorst is a creative and adventurous scholar, but some of his idea strike me as unjustified.

He's someone who knows his stuff extremely well, so I'd hesitate long before betting against him.

Yes, he knows his stuff very well; he is the current top authority on Anatolian languages, even a notch above Craig Melchert. And you are right that many sound changes are pretty much unidirectional.

All of this talk about the "Graeco-Aryan" aspect system has me wondering. Correct me if I'm wrong, but it seems that although the morphology used for creating said aspects is common among many of the other IE branches, the actual semantics of said formations are inconsistent from branch to branch, even between Greek and Indo-Iranian. It might be more consistent with the evidence to state that Common IE had a variety of possible verb formations with diverse meanings floating around its verbal system, and the Graeco-Aryan verb is a common innovation from this earlier stage into a more organized verbal system revolving around three major aspects. The other branches that are traditionally seen as not retaining the common verb system, have instead just reorganized the diverse verbal formations differently without ever having passed through a distinct three-aspect stage to begin with.

WeepingElf wrote: ↑Tue Feb 21, 2023 9:55 am [Kloekhorst] is the current top authority on Anatolian languages, even a notch above Craig Melchert

I was under the impression (admittedly from not very much) that his ideas about wider Indo-European studies were extremely controversial and not widely accepted. Is he only an authority for Anatolian languages or is this impression simply mistaken?

abahot wrote: ↑Tue Feb 21, 2023 3:38 pm All of this talk about the "Graeco-Aryan" aspect system has me wondering. Correct me if I'm wrong, but it seems that although the morphology used for creating said aspects is common among many of the other IE branches, the actual semantics of said formations are inconsistent from branch to branch, even between Greek and Indo-Iranian. It might be more consistent with the evidence to state that Common IE had a variety of possible verb formations with diverse meanings floating around its verbal system, and the Graeco-Aryan verb is a common innovation from this earlier stage into a more organized verbal system revolving around three major aspects. The other branches that are traditionally seen as not retaining the common verb system, have instead just reorganized the diverse verbal formations differently without ever having passed through a distinct three-aspect stage to begin with.

We can't exclude that. This actually echoes an idea of Roland Pooth, but he assumes a much more radically different nominal and verbal system of categories and morpholgy for PIE. His assumption is that the systems attested in the known IE languages are the result of parallel developments in the individual families, based on a common tendency towards nominative - accusative systems, from the originally very different system he proposes. Now, you don't need to accept his orginal system, but he maybe is on to something with the mechanism; that would be comparable to how articles, have-perfects and other features spread through modern European languages.

abahot wrote: ↑Tue Feb 21, 2023 3:43 pm I was under the impression (admittedly from not very much) that his ideas about wider Indo-European studies were extremely controversial and not widely accepted. Is he only an authority for Anatolian languages or is this impression simply mistaken?

His ideas are under discussion. He's working at the cutting edge of IE research, and the deepening of knowledge about Hittite, Luwian, and the other Anatolian languages is still overthrowing old assumptions about PIE. One problem is that the circle of scholars who actually know Anatolian well enough to argue with him and who are interested in working on the reconstruction model of PIE (as opposed to, say, just using the standard model to etymologize this or that word in an Anatolian language) is relatively small. Whether his ideas will become accepted or not, we'll see in two or three generations. A century is a short time in IE linguistics

hwhatting wrote: ↑Wed Feb 22, 2023 3:47 am He assumes a much more radically different nominal and verbal system of categories and morphology for PIE.

I think there's something to be said for that, honestly. Not for his ideas specifically, which I only know a bit of, but for the general idea that IE reconstructions as a whole are too Graeco-Aryo-centric, ascribing to PIE what might make more sense to be described as an areal innovation along the southern edge of the IE dialect continuum since those languages were recorded earliest.

Some of those features in question may include things such as:
- The tripartite aspect system.
- The realization of *o as a back rounded vowel (in this case, secondarily lost in Indo-Iranian)
- The realization of the *D^h series as aspirated in some sense.

@abahot: I don't disagree with any of the things you say, although the only thing I have a firm opinion on is that the tripartite system is a post-PIE regional development.
As for Pooth, he goes way beyond saying that the Graeco-Aryan model is an innovation; for him, even Anatolian has already travelled far down the road from the original system towards the nom-acc system.

I am fairly conservative with regard to Late PIE, because the standard model accounts of the facts in the non-Anatolian IE languages well enough and most amendments that have been proposed only make things worse; but I am quite adventurous with regard to Early PIE, because the standard model does not account of the Anatolian branch well, and there are numerous hints at it having had a structure quit different from Late PIE. For instance, Early PIE may have been an active-stative language, or at least a spltit-ergative one where animate nouns followed an accusative paradigm and inanimate nouns followed a "privative ergative" paradigm (i.e. one in which they have an absolutive case but not an ergative case because they were disallowed as transitive subjects).

hwhatting wrote: ↑Wed Feb 22, 2023 3:47 am His ideas are under discussion. He's working at the cutting edge of IE research, and the deepening of knowledge about Hittite, Luwian, and the other Anatolian languages is still overthrowing old assumptions about PIE. One problem is that the circle of scholars who actually know Anatolian well enough to argue with him and who are interested in working on the reconstruction model of PIE (as opposed to, say, just using the standard model to etymologize this or that word in an Anatolian language) is relatively small. Whether his ideas will become accepted or not, we'll see in two or three generations. A century is a short time in IE linguistics

So basically, he is to Anatolian languages, what Everett is to Piraha?

Dear gods, thats the most terrifying question/sentence I've written in at least ten years.

keenir wrote: ↑Thu Feb 23, 2023 8:59 pm So basically, he is to Anatolian languages, what Everett is to Piraha?

No, not really. There are a lot more scholars knowledgeable about Anatolian than there are about Piraha, so when Kloekhorst states things about Anatolian, that can be easily checked, and while I'm not following closely, I haven't heard that he's been accused of serious errors or misrepresentations of the facts; it's just that only a small subset of these scholars is ready to look into how to best interpret the orthography of Anatolian languages in phonetic terms, or what this interpretation or facts about morphology in Anatolian mean for PIE. And other IE scholars working on modelling PIE often have only a cursory and outdated knowledge of Anatolian, so they either have to accept Kloekhorst's conclusions at face value or ignore them / handwave them away, because they conflict with their views on PIE and they lack the detailed understanding to see why Anatolian evidence points a certain way.

Indeed, the cases cannot be compared. Everett apparently is the only scholar working in Pirahã, and given his extraordinary claims about the structure of that language, some scholars suspect that his work is spurious. (A few years ago, I had the words "The Piltdown Man spoke Pirahã, invented cold fusion and died of spiroptera carcinoma." in my signature for a while.) Anatolian, in contrast, is a language group studied by many scholars, there are text editions, grammars, dictionaries and a large body of research literature for anyone interested to consult, and Kloekhorst is just one of many scholars working on it - the currently leading scholar perhaps, but not the only one. That said, I don't buy all of his ideas about Anatolian, much less on PIE, but I respect him.

abahot wrote: ↑Wed Feb 22, 2023 11:44 pm Some of those features in question may include things such as:
...
- The realization of *o as a back rounded vowel (in this case, secondarily lost in Indo-Iranian)

Don't Italic and Celtic reflect back rounded *o? And Albanian has *ō > *ö > e, although I vaguely remember this having problems.

Nortaneous wrote: ↑Fri Feb 24, 2023 12:50 pm

abahot wrote: ↑Wed Feb 22, 2023 11:44 pm Some of those features in question may include things such as:
...
- The realization of *o as a back rounded vowel (in this case, secondarily lost in Indo-Iranian)

Don't Italic and Celtic reflect back rounded *o?

They do! And that means that the Old European Hydronymy (if it really reflects an old IE language rather than being Vasconic or just bogus) can't be Italic, Celtic or (if that's a thing) Italo-Celtic, because there *o has merged with *a.

Nortaneous wrote: ↑Fri Feb 24, 2023 12:50 pm Don't Italic and Celtic reflect back rounded *o?

The idea is that the realization of *o as a back rounded vowel is only shared by IE languages at the south(west) of the dialect continuum -- Italo-Celtic, Hellenic, Phrygian, and Armenian. The theory that this is an innovation may be wrong of course, but in general it's telling that a number of common features seem to be present in this "area" of the IE dialect cloud. The exact details seem to differ from branch to branch, but that's something we see elsewhere in IE, like the satem and ruki sound changes, for example.

abahot wrote: ↑Wed Feb 22, 2023 11:44 pm - The realization of *o as a back rounded vowel (in this case, secondarily lost in Indo-Iranian)

Also Tocharian *o apparently causes some rounding in preceding vowels and doesn't cause palatalisation, although it later unrounds to /e/ in Toch B (/a/ in Toch A).

eg. *okʷs > *ëk > TA ak, TB ek
*ǵómbʰos > *këm(β)ë > TA kam, TB keme
contra eg. *gʷeneh₂ > *śänā > TA śäṃ, TB śana

Znex wrote: ↑Fri Feb 24, 2023 5:13 pm Also Tocharian *o apparently causes some rounding in preceding vowels and doesn't cause palatalisation, although it later unrounds to /e/ in Toch B (/a/ in Toch A).

eg. *okʷs > *ëk > TA ak, TB ek
*ǵómbʰos > *këm(β)ë > TA kam, TB keme
contra eg. *gʷeneh₂ > *śänā > TA śäṃ, TB śana

Interesting. I always sort of treated Tocharian vowel changes like a black box, being too complex and poorly understood to offer any useful information about their predecessors, but perhaps that view was mistaken.

The Tocharian vowel developments are complex and confusing to laypeople, but AFAIK the experts understand them.

The discussion in this area is about two things; (1) the number of central vowel phonemes in PIE and (2) the quality of these phonemes.
A bit of history on (1):
In the beginning, IE scholars, assuming Vedic to be nearest to PIE, postulated a three-vowel system with long and short pairs: a(:), i(.), u(:). the bigger number of central vowels (with e(:) and o(:) beides the three others) in the Europeam languages (Tocharian and Anatolian weren't known back then) was explained by unsystematic splits.
Then the Neogrammarians came along, established the principle of sound laws, and also found traces of e(:) and o(:) in Indo-Iranian (Brugmann's law, palatalisation before old e(:). They established the Classical five-vowel model, with a e i o u in short and long pairs.
Next step: the laryngeal theory explained many instances of a(:) as e(:) coloured by laryngeals. As a reaction, a significant group of iE scholars came to the conclusion that a(:) didn't have phonemic status in PIE and was always due to colouring (by laryngeals, velars) or the result of a secondary schwa (schwa secundum) in certain zero-grade positions in individual families or as a regional development. But there's also a significant number of scholars that maintains that there was a (rare) phoneme a(:) in PIE not caused by colouring. Seeing that i(:) u(:) alternate with the semivowels y, w, some IE scholars count them not as vowel phonemes, but as allophones of consonants, comparable to the syllabic nasals and liquids. So, according to whom you ask, you get anything from two to five vowel phonemes (in short and long pairs) for pre-split PIE (some traditionalists also count the syllabic nasals and liquids as vowels on top of that). If you go further back to pre-ablaut PIE, it's generally assumed that all long vowels are due either to laryngeals or ablaut, so pre-ablaut PIE did not have a length contrast, and further that e, o are only ablaut variants of an ur-vowel, which is variously designated e or a. So, in not counting the high vowels, some scholars arrive at a one-(phonemic)-vowel system for pre-ablaut-PIE.
On (2): Seeing that there is no continuation of PIE *o(:) as o(:) in Anatolian, and based on arguments on the quality and phonemicity of the palatal / velar/ labiovelar series in PIE, and also seeing that PIE *o shows up as a long vowel in the Brugmann positions in Indo-Aryan, some scholars have argued that pre-split PIE had not e, o, but something like a, ɑ with maybe a non-phonemic length distinction that developed into e, o in (some of) the European branches. Now, if Tocharian shows clear traces of a rounded quality for PIE *o, then we either need to assume that the rounding to *o must have taken place before Tocharian split off, in which case the traditional assumption of a development *o > a for Balto-Slavic, Germanic, and Indo-Iranian still applies, or that the rounding to *o happened independently in the branches or as an areal phenomenon (when Tocharian was still spoken more westerly, maybe).
On (1), I haven't yet seen a convincing explanation that can derive all instance of a as caused by colouring or schwa secundum, so I assume a five-vowel system for pre-split PIE, which at the same time means that a system in (2) like a, ɑ would include a split of from two phonemes into three e, a, o, for which the conditions need to be explained. More generally, these issues are heavily debated and there's no consensus; but if you want to use anything but the default traditional five-vowel system in your reconstructions, it's better to state that up-front, if you want to publish them.

hwhatting wrote: ↑Sun Feb 26, 2023 7:04 am Now, if Tocharian shows clear traces of a rounded quality for PIE *o, then we either need to assume that the rounding to *o must have taken place before Tocharian split off, in which case the traditional assumption of a development *o > a for Balto-Slavic, Germanic, and Indo-Iranian still applies, or that the rounding to *o happened independently in the branches or as an areal phenomenon (when Tocharian was still spoken more westerly, maybe).

Or (following Adams, Hamp, and Fellner) that Tocharian didn't split off early.