Module talk:izh-pronunciation

Final consonant voicing in Ala-Laukka

[Continuing a discussion started in talk]. See 4.8.2 in Kuznetsova, starting from "Отметим однако одну любопытную тенденцию": apparently (and confirmed by Kuznetsova and Muslimov), voicing after a dropped vowel happens where i is reduced before p and s, i.e. pI > b, sI > z, but looks like more or less systematically this happens for ND, who was known for taking the dialectal features to an extreme (see examples in tables 45-46). On the other hand, voicing of regular terminal consonants is quite consistent: lammaz, -d, -vad are clear examples shared amongst most of the informants (apart those who had more contact with Finnish). I'm currently checking what are the limits for voicing of the initially terminal consonants. KirillW (talk) 21:24, 9 November 2023 (UTC)

@KirillW: By the way (regarding your comment about maata), note how in T. 45: kraasata ~ kraasat ~ kraasət (and maybe also ~ kraasəd? that one doesn't seem to be recorded, so maybe only /p/ and /s/ get voiced?); So I assume that similarly maata ~ maat, no? Thadh (talk) 22:33, 9 November 2023 (UTC)

@Thadh, that's exactly the point: -ta > -tə > t, but not > d. I think I suggested at the beginning to render this kind of pronunciation like t(ə), but again can't find what was the outcome of that discussion.

For voicing we have 4 distinct cases:

• Initially terminal /s/ and /t/ > /z/ and /d/, very widespread in Ala-Laukka
• Kaik /kaig/, looks like a special case, but very widespread too
• After reduction /-pi/ > /-b/, /-si/ > /-z/, consistent only with certain speakers in the very extreme South. I suggest we don't include this into the pronunciation module (yet?)
• Partitive case for adjectives which historically had *-eda. Luckily we don't have to do anything with that :)

KirillW (talk) 10:42, 10 November 2023 (UTC)

@KirillW: Another question, does any consonant if followed by a final -i get palatalised when that -i- is dropped? And also, how do you suggest handling final i, just like now (dropping) or marking it as reduced and keeping it between parentheses? Thadh (talk) 12:54, 10 November 2023 (UTC)

@Thadh there's a great summary table 85 in Kuznetsova. The issue is that the answer is as usual "it depends". I see there 2 possible approaches here to consider: 1) palatize always, or 2) palatize only l and t. Both options looks solid, since palatalization is widespread on one hand, but is less pronounced for anything but l and t on the other hand. KirillW (talk) 17:01, 10 November 2023 (UTC)

IIRC, you were suggesting that marking vowels reduced would add unnecessary complication. But it's still kind of the closest we can get to the "average". Just to quote Kuznetsova on the reduction: 1) a, ä, e > ə > Ø; 2) o, ö, u, ü, i > [ŏ̥, ŭ̥, ü̥̆, ĭ̥] > Ø. In this case to be consistent we need to use (ə) for reduced a, ä, e and something in parenthesis for reduced o, ö, u, ü, i. What kind of something should this be to represent the actual sound, but without over-complicating things—I don't know unfortunately. KirillW (talk) 17:10, 10 November 2023 (UTC)