I noticed that *Cj/*Cy (depending on if one uses IPA or IEist notation) and *Cw sequences are rare in PIE (with most being the result of schwebeablaut or regular ablaut). Among sequences that aren't the result of schwebeablaut or regular ablaut, *Tw sequences are the most abundant (with *T denoting any dental, including *s), followed by *Kw (*K denoting any dorsal), then by *Hw (*H denoting any laryngeal). The sequences with *j/*y are even rarer, with the same T>K>H pattern as sequences with *w. *C is never a labial or a resonant. Even the *Tw sequence represents a very tiny portion of reconstructed PIE. Any reason why C+glide sequences are rare and why the dentals are the most likely to be *C in both *Cj/*Cy and *Cw while *C is never a labial or resonant in such sequences?

Edit: there are also *Cj/*Cy and *Cw arising from combinations of different roots and elements. The above doesn't refer to clusters of such origin.

  • w and y are semivowels; I think I remember reading recently that a root *Cew- as a zero grade turns the labiovelar to a vowel, *Cu-. Imaginably the same holds for y>i. My lack of knowledge makes me wonder especially about Slavic with swoti ~ sloti on the one hand, and ljublu, sowjet etc, the [i] written as a reverse b of all things (< *lewbʰ- "love", Ger Lieb-). – vectory Nov 7 '19 at 3:33

The dearth of labial+w sequences is attested in many non-Indo European languages. The best explanation seems to be that it is a perceptually challenging sequence, where it is hardest to identify [pw] as being distinct from [p]. More specifically that is because labials and rounding both have a lowering effect on formants. The case of [kw] is in part because "kw" is analyzed as a single labio-velar phoneme, i.e. [kʷetwor], not [kwetwor]. A contrast between [kw] and [kʷ] would really hard to hear.

The [j] vs. [w] asymmetry in onsets is attested in a number of other families, for instance in Bantu onset cluster simplification tend to target [j] rather than [w]. The causes of this hasn't been studied systematically, but I have observed that it is often been that Cj clusters are targeted for some palatalization-like phenomenon, e.g. kj→tʃ, but merger of C+w into a single compromise place of articulation is not so common (nothing in Bantu like the IE kʷ→p/t pattern).

  • Needs an explanation for the absence of *Pj/*Py and resonant+glide sequences, the extraordinary rarity of laryngeal+glide clusters and why *Tj/*Ty is still more common than *Kj/*Ky (though it could have been due to the above mentioned palatalization, which if true could also apply to *Pj/*Py). Excellent job at explaining obstruent+/w/ sequences and providing examples from other language families though. – JMRD Nov 6 '19 at 1:08
  • the phonetics of *kʷ, like all other PIE phonemic symbols, are not solved. You say "would be hard to hear", what's the dependent clause? Personally, I memorize [kʷ] like French "quoi" or English "qu" with rounding, voice (and perhaps a backed tongue), while I can easily hold [kw] the German voiceless, way as in Quark, which is almost [kf] (there was a Ger.SE threat where different realizations of "qu" were observed). Even qof and how all those letters are written have a round big fat O. cp Ger Quartier vs AmE ca. "quorters" (quarters) – vectory Nov 7 '19 at 3:59
  • G'wehr as an apophone of Gewehr tends to voicelessness though. Uneconomic to be made heard, I'd say, . By the way, since I'm always keen to commit alternative facts, please compare *ghwer- "to burn" with the above. – vectory Nov 7 '19 at 4:08
  • PIE did, though rarely, contrast *ḱw and *kʷ (*eḱwos "horse"). – TKR Dec 6 '19 at 20:10

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.