Reconstruction talk:Proto-Slavic/vitь

*vitъ vs *vitь

Russian, Polish, and Slovac descendants originate from Proto-Slavic *vitь (with soft yer at the end). Should we leave them here? Furthermore, the proposed etymology from Balto-Slavic *witis presumes Slavic **vьtь. I guess the laryngeal in the root is to be blamed for the long -i-. Probably, reconstructing Balto-Slavic *wˀitis > pre-Slavic *weitis is better?

Anyway, the form with hard yer should be from pre-Slavic *weitаs. Bezimenen (talk) 16:11, 8 May 2019 (UTC)

@Bezimenen ˀ is not a separate phoneme, so it can't be placed just anywhere in a word. Derksen of the Leiden school holds the position that it was an actual consonant, but it's not mainstream among linguists concerned with Balto-Slavic so we don't follow his position. What is understood as mainstream is that Balto-Slavic has short or long vowels from laryngeals according to the same rules that are found in other Indo-European languages: HV becomes short V, VH becomes long V̄. The catch with Balto-Slavic in particular is that the latter results in an acute syllable, which the symbol ˀ denotes, so it's more properly V̄ˀ. But that's just what ˀ indicates in this case, a syllable-wide effect and not a consonant. It is always written after a long vowel or diphthong and not anywhere else.

Regarding the Slavic form, *vitь is what Derksen gives, and it doesn't look like any of the descendants disagree on that, so it should probably be moved.

I think you are correct to point out that the Balto-Slavic form is wrong, and the Lithuanian form vytìs agrees on that. But the Lithuanian form also shows that it's not *weitis, since that would become *vietis. Rather, Lithuanian y indicates Balto-Slavic long *ī. The next question is whether it was acuted or not (*ī or *īˀ?). We list the Slavic form with accent paradigm b, and the Lithuanian form belongs to paradigm 4. Both of these accent paradigms arose from a non-acuted root syllable, so that probably answers that question, and *wītis is the correct form. I'd like to know what the Slavic paradigm b is based on, though, since Derksen doesn't give it, and it disagrees with the Lithuanian noun, which has mobile accent (=Slavic accent paradigm c). @Victar, as the one who added the accent paradigm, do you know what it's based on? —Rua (mew) 16:40, 8 May 2019 (UTC)

Don't recall. Don't care. Delete table if you want. --{{victar|talk}} 16:44, 8 May 2019 (UTC)

Hmm ok, then we really only have the Lithuanian evidence to go on, which isn't as much. —Rua (mew) 16:47, 8 May 2019 (UTC)

Derksen gives the verbs výti amd vît as being from the same root, and both of those unambiguously indicate an acute. Why the Lithuanian noun has an accent paradigm indicating a non-acute root I can't say. —Rua (mew) 16:55, 8 May 2019 (UTC)

@Rua: I meant the -e- only for the pre-Slavic form *weitis (as an emphatic innovation specific for Slavic). As you point out, the Baltic descendants do not follow this reconstruction.

PS What is the official convention which Wiktionary follows on Balto-Slavic? I've encountered various notations - e.g. some using -o- and -ā-; others -a- and -ā-; third writing ˀ as ʔ. It's really messy (mainly due to misuse by users like me). Bezimenen (talk) 16:56, 8 May 2019 (UTC)

I don't see the need to reconstruct ei for Balto-Slavic when the Lithuanian form explicitly indicates ī, and the Slavic form could reflect either ei or ī. That makes ī the most obvious option.

The official convention on Balto-Slavic transcription should be on Wiktionary:About Proto-Balto-Slavic, but nobody has written it yet it appears. In any case, to transcribe from Derksen into our notation, short vowel + ʔ becomes long vowel + ˀ, and then any remaining short o becomes a. There is no reconstructable difference between short o and a in any Balto-Slavic language, both always emerge as the same vowel. I suppose people like Derksen claim a difference so that their theories work out nicely, but those theories aren't widely accepted anyway. —Rua (mew) 17:02, 8 May 2019 (UTC)

@Rua: I don't suggest to reconstruct -ei- for Balto-Slavic. My initial assumption was that pre-Slavic (the form of Balto-Slavic that would eventually give proto-Slavic) developed -e- post factum. However, after some checking, it seems this is unnecessary. Variations like Lithuanian mùšti (“to hit”) vs Lithuanian mū̃šis (“battle”) and Lithuanian lìkti (“to keep”) vs Lithuanian lỹkis (“remainder”) suggests that the lengthening was productive in general in Balto-Slavic. In other words, we should just reconstruct *wīˀtis, not **witis.

On the other hand, though, PIE *-tis requires e-grade. Oh, I give up... I need to re-read some theory on Balto-Slavic. Bezimenen (talk) 17:15, 8 May 2019 (UTC)

I did some digging and found that Derksen distinguishes *wʔi- from *wiʔ-, but we give both as *wīˀtei (look in the references), and they are completely indistinguishable in the languages that preserve both. Derksen also gives *bū́ˀtei, a verb that very clearly has a long acute vowel, as *bʔu-. I don't understand the logic behind his reconstructions (he and especially Kortlandt seem like they're on their own planet sometimes), but at least this suggests that we should treat Derksen's glottal stops as acutes regardless of the ordering of neighbouring vowels. This, in turn, means that his *wʔitis becomes our *wīˀtis. If that's the case, I don't really get where Lithuanian gets accent paradigm 4 from, though. —Rua (mew) 17:30, 8 May 2019 (UTC)

Regarding the e-grade in *-tis, I'd be careful. Lots of cases all over Indo-European attest this with a zero grade root throughout the whole paradigm, and fixed accent on the ending. Analogical levelling was very commonplace in the *-tis and *-tus nouns, as well as all other athematic paradigms. So it's best not to assume that descendants have any particular grade or accent, but instead that the attested grade and accent must come from at least one of the forms. There's no rule that says the grade and accent must come from the same form either, descendants can have genitive root grade but nominative accent or vice versa. —Rua (mew) 17:30, 8 May 2019 (UTC)

Yeah, I've noticed this. Balto-Slavic, in particular, tends to use 0-grade like in *mirtis (“death”). *wīˀtis is probably an exception of this rule of thumb. Bezimenen (talk) 17:36, 8 May 2019 (UTC)