Reconstruction talk:Proto-Georgian-Zan/ber-

*რ- → ჯ-

The current version claims this sound change is regular; it is not. There are MANY, many lexical items that do not undergo such a sound-change. The reason is that Proto-Kartvelian must have had two separate phonemes: a palatalized *rʲ and a regular *r, and only the former underwent the regular sound-change above. The proof is that there are even reconstructible minimal pairs such as *pur 'heat' and *purʲ 'buffalo', and only the latter undergoes this change. Trwier (talk) 09:07, 26 December 2024 (UTC)

You're not proving anything with a "minimal pair" like that. One is a verb, the other is a noun. In *pur- 'buffalo' we have purʲ- → puǯ- because the r- was followed by an -i. In *pur- 'heat', we don't have that because the r- wasn't followed by an -i. კვარია (talk) 15:36, 26 December 2024 (UTC)

No it's more complicated than that. There are plenty of cases with the same phonological context. For example in causative verbs in Megrelian, you can have an -ინ suffix, and here we see /r/ in Zan languages, not /ǯ/. There are also cases in other lexical items where /i/ follows *inside* the morpheme in Megrelian, e.g. ფურინი purini 'fly (away)' and ფურიალე puriale 'steep rock'. (No one is going to say that Megrelian ფურინი is not a very old lexeme in Kartvelian!) In such cases you are forced to argue that the rule is irregular or morphophonological if you insist that there is just one unitary *r phoneme. A more plausible explanation is that the rule applies to a different phoneme, namely *rʲ. With the great evidence that Proto-Kartvelian had a palatalized series of obstruents (especially sibilants), this is not a very major claim. Trwier (talk) 17:55, 26 December 2024 (UTC)

You're comparing apples and oranges. The split is clearly seen in Zan q̇vaǯi but q̇var-; čxinǯi but čxir-. {{R:ccs:Fähnrich:2016}} says the following: Im Auslaut von Nominalstämmen affrizieren Mingrelisch und Lasisch das r, so daß die Entsprechung georg. r : mingr. ǯ : las. ǯ : swan. r entsteht. You said "minimal pairs", in the plural. You should provide the supposed data set so the distribution of said sound change can be evaluated. Perhaps Fähnrich will ammend his point of view following your discoveries. კვარია (talk) 18:43, 26 December 2024 (UTC)

You missed my point about the phonological conditioning factor: minimal pairs are important but they are less important than understanding the conditioning factor. It is simply not the case that all instances of Zan /ǯ/: Geo /r/ correspondences can be explained by a following /i/ in Zan. I already gave you examples of words that failed to undergo sound-shift you are arguing for. Some other examples from Megrelian: v-o-skur-in-u-a 'ich mache Angst'; v-o-bur-in-u-a 'ich werfe'; zor-i 'lauter Ton mit dem Horn'; ko-b-dzir-i 'ich sah, ich fand'; zir-il-i 'eingesammelt'; mu-zar-i 'weibliches Kalb'; p-tir-i 'ich zerrte, schleppte'; txor-il-i 'Graben'. All of these have a correspondence of Geo /r/ : Zan /r/, and that is because despite the fact they have /i/ following, the palatalization rule is not triggered by that environment, but rather by an earlier phonemic contrast between *rʲ and *r. Only the former, *rʲ, underwent the palatalization rule to /ǯ/; the latter, *r, did not.

There are also cases where /r/ in Megrelian is secondary and therefore also fails to undergo this sound-shift: zisxir-i 'Blut', txir-i 'hazelnut', from the lateral fricative *ɬ. You could explain these facts by chronological rule-ordering, but nonetheless they are surface counterexamples to the idea that /i/ triggered the palatalization to /ǯ/ in Zan. Trwier (talk) 19:38, 26 December 2024 (UTC)

muzari does not belong here; that reconstruction is simply spurious because the adduced words are late Iranisms. All the other roots that you just showed here, with the exception of zori the failure of which to undergo the rule may be explained as a simple onomatopiea, were verbal stems. Zan words denoting 'blood' are irrelevant since as you say the /r/ is secondary, what is even the point of this remark? Don't muddy the waters. კვარია (talk) 19:51, 26 December 2024 (UTC)

Almost everything you say here is grossly inaccurate or misleading or irrelevant:

- 'Onomatopoeia' is a complete red-herring: it's a nonexplanation.

- No person who takes sound-symbolism seriously would say that lexemes like 'throw', 'find', 'collected', 'bring', 'ditch' are onomatopoeia.

- the fact that some of these are verbal stems is irrelevant, because the conditioning factor was phonological, not morphological.

- there are in fact plenty of other nominal stems, if that is what you seek: e.g. Megrelian კვარი, corresponding to Georgian კვერი; Megrelian მასარი, Georgian მესერი; Megrelian ჟირი, Georgian ორი; Megrelian რინა (the noun form of 'to be'), Georgian არ-; another minimal pair would be *sur- 'complete, whole' vs. *surʲ- 'ivy' (both are noun stems).

- the point of mentioning the secondary /r/ is to show precisely how the process is phonological, not morphological.

Are you seriously going to suggest that a lexeme like 'two' (ჟირი/ორი) somehow doesn't count, or is not Kartvelian enough? If, as you say, the conditioning factor were purely morphological, that would predict Megrelian *ჟიჯი, when in fact we know that is not the form found in that language.

The real answer to all these issues is that the original generalization is wrong: there were two rhotic phonemes in Kartvelian, *rʲ and *r, and only the former of these underwent the affrication rule. Trwier (talk) 22:30, 26 December 2024 (UTC)

> No person who takes sound-symbolism seriously would say that lexemes like 'throw', 'find', 'collected', 'bring', 'ditch' are onomatopoeia.

To whom do the words that these particular roots are sound-symbolic belong? Who said that exactly? You constructed some false narrative specifically to disprove it. That's a strawman. The positional character of this sound change was noted by Rogava a long time. It was from then on accepted in Kartvelology, including by notable Kartvelologists like Klimov and Fähnrich. There is scientific consensus. The cases where this sound change did not operate were also noted a long time ago, you haven't opened any new ground here. You're the one pushing your personal point of view about the necessity to reconstruct Proto-Kartvelian *rʲ. This is just your opinion. An opinion just as marginal as reconstructing a lateral fricative in *zisxli whereas the general consensus is that a syllabic *l must be reconstructed here. Clearly, this conversation has run its course as you've now resorted to strawmans. კვარია (talk) 16:51, 27 December 2024 (UTC)