Use this thread to ask questions or share trivia, if you don’t want to create a new thread for that.

[Note: the purpose of this thread is to promote activity, not to concentrate it. So if you’d still rather post a new thread, by all means - go for it!]

  • Lvxferre [he/him]OPM
    3·
    9 days ago

    It’s a bit of a crack theory, but I think that Early PIE had a vertical system. What’s being reconstructed as *e *o are actually *ə *ä or similar, and there’s a missing third vowel. This system is rather common in languages from the Caucasus, and it’s likely that Early PIE interacted a fair bit with them, making it an areal feature.

    I’m saying this based on:

    • Almost every single time you see *ē or *ō in a PIE reconstruction, there’s a missing consonant nearby; e.g. *ph₂tḗr “father” missing the nominative *-s. I don’t think those vowels were already present in early PIE, so this reduces the vowel system even further into *e *o.
    • Vowel raising and centralisation are way more common than vowel lowering. So why are current reconstructions proposing that Proto-Indo-Iranian, Lithuanian, Armenian, Albanian, Tocharian, Hittite, Proto-Germanic, are all lowering PIE *o into *a, instead of claiming the others do the opposite?
    • Under the current reconstructions, almost all IE branches get rid of the syllabic consonants through epenthesis. They almost never do it through simple deletion, even if that’s what you’d expect from typical evolution patterns. There’s a missing vowel there.
    • Syllabic consonants (plus *i and *u) oddly gravitate towards weird phonotactics: initial CC, medial CCC+, or final CC. That reinforces my belief that there’s a missing vowel there, and syllable structure is simpler; perhaps even (C)V(C).
    • A third vowel being conditionally merged into *e = *ə depending on accent would explain ablaut rather nicely.
      • Lvxferre [he/him]OPM
        1·
        3 days ago

        Thank you for linking this paper. His take is the opposite of mine - he proposes current *e was actually *a, instead of *o. It’s actually worth investigating this because at least some of his arguments are fair points, specially #2 (o-grade behaving like zero-grade) and #3 (*o limited distribution).

        It does create a problem, though; in plenty languages you’d have *a *ə→*e *a, as if they swapped places. While phenomena like this are attested (Dixie English comes to my mind*), it’s messy and cross-linguistically rare.

        *e.g. Southern US English renders /äɪ̯/ as [ä:] and merges /ɛ ɪ/ as [ɪ], so if you look from Middle English to now it’s like the vowels were swapped - /i: ɛ/→/ä: ɪ/.

        • antonim@lemmy.dbzer0.com
          2·
          12 hours ago

          Yep, I mean his take is not quite opposite of yours, as it uses the same units and has a similar idea on his mind, just one step is opposite.

          Back when I read the article first time, I tried to figure out the details - I suppose they wouldn’t immediately swap places but would first have a “fission” of *a (= trad. *e) + H into *e, *a, *o, and then *ə (trad. *o) > *o, which might(?) be a more natural late PIE triangular vowel system that would then, I suppose(?), lead to the later outcomes more naturally. So I guess I don’t see the swapping?

          But I might be talking absolute nonsense since I don’t have a satisfyingly firm grasp on traditional PIE reconstruction either… D:

          • Lvxferre [he/him]OPM
            1·
            10 hours ago

            Yup - the core idea is the same, only the “implementation” is the opposite.

            I’m calling it “swapping” because, in some branches, what’s currently reconstructed as *o ended as *a: for example PIE *h₁óynos “one” → Proto-Balto-Slavic *aiˀnas, Proto-Germanic *ainaz. So if his hypothesis is true, in those branches the mid vowel becomes the low vowel and vice versa; this does happen but it requires some specific conditions (like gliding, length, or some other secondary articulation), otherwise the vowels end merging midway.

            *e, *a, *o, and then *ə (trad. *o) > *o, which might(?) be a more natural late PIE triangular vowel system that would then

            On its own *a *e *o wouldn’t be too natural, but if we include *i *u it does. And by then odds are that the later were already “promoted” to vowels.