4

The "maximal onset principle" says that, in many (most) languages, consonants will attach to a syllable onset rather than a coda when given the choice. For example, "walking" /wakɪŋ/ in English is syllabified as wa.kɪŋ, even though wak is a valid syllable, because the k prefers to be in an onset.

Wikipedia's article on Turkish, however, says that:

Turkish syllabification in multisyllable words follows maximal coda rule

I've never heard of a "maximal coda rule" before, and the examples in the same Wikipedia article don't seem to show one clearly. However, it's easy to imagine how it would work, and I'd readily believe it happens in some language—I've just never come across it.

So: is the "maximal coda principle" real? If so, what languages does it appear in? If not, is there a theoretical reason why it can't, or is it an accidental gap ("that's just the way it is")?

  • 2
    Ask the editor who added the sentence. To me the sentence doesn't even make sense, because "V.CV" clearly does not follow "maximal coda rule" if taken literally. – Nardog Dec 17 '18 at 19:33
5

The quoted sentence from the Wikipedia article isn't very clear, and I wouldn't be confident that the author knew what they were talking about.

Syllables and syllabification rules are very theoretical topics. I think most theories that recognize the syllable as a unit of analysis do include reasons for the supposed "maximal onset principle" (or some similar rule). But because it's such a theoretical concept, it doesn't necessarily correspond in any simple way to all of the data that we have access to.

Arrernte is an example (fairly well-known, I think) of a language that has been analyzed as containing only "VC(C)" syllables. You can see an alternative analysis invoking the concept of "moraic onsets" here: "Arrernte moraic onsets: stress and beyond", by Nina Topintzi and Andrew Nevins.

As Nardog said in a comment, "V.CV" in Turkish is obviously a counterexample to the supposed tendency to maximize codas. As far as I can tell, the actual situation in Turkish is simply that complex onsets are not usual (something that the article itself points out in the same section), so we have "VC.CV" not because of some principle of "maximizing the coda" of the preceding syllable, but just because the alternative syllabification "V.CCV" would go against the tendency/historical constraint/whatever is behind Turkish not having complex onsets outside of loanwords. Turkish apparently tolerates complex codas in at least some circumstances (e.g. kent "city"), so you could look at how sequences like kenti (the definite accusative form of kent) are syllabified for more evidence about whether Turkish actually follows any sort of "maximal coda rule".

My impression is that syllabification in modern Hebrew follows a similar pattern. Biblical Hebrew is described as having no word-initial consonant clusters (and also, I believe, no complex onsets in any position). Modern Hebrew does have word-initial consonant clusters, not only in borrowed vocabulary but also in native vocabulary (because of schwa loss in certain conditions), but my understanding is that word-internal VCCV in Hebrew is still considered to be syllabified as VC.CV (at least usually; I don't know if there might be some exceptions). From an Optimality Theory perspective, "Issues in the Phonology and Morphology of the Modern Hebrew Verbal System: Alignment Constraints in Verb Formation", by Yael Sharvit (1994) lists "*COMPLEX: No complex onsets or codas" as a constraint on Hebrew syllable structure (taken from Prince & Smolensky (1993)); I would guess that this could also be postulated as a constraint that is active in Turkish syllabification.

| improve this answer | |
3

I'll say yes there is, at least if you assume OT. In fact, there is some credibility to the notion from Saami languages, as argued by Bye. Baertsch argues for an approach that has something akin to Coda Mazimization (as opposed to just Ons and *Coda). Oda in this paper argues for coda maximization in a non-OT functional phonetic analysis.

The argument from N. Saami is not trouble-free, but the basic fact is that /beajsku/ → [beajsˑ.ku] 'vermin (acc.)', where s has to be in the coda, even though [sk] is a legal onset. "Coda Maximization" would then tilt the scales in favor of parsing s in the coda, where otherwise one would expect the longest possible string of Cs to go into the onset.

Incidentally, there is a competing parse of "walking" as [wɑk.ɪŋ]. It is often said that stressed syllables have to be heavy in English, and that parse is one way to satisfy that requirement. And the aforementioned Saami fact is about stressed syllables (where much of the phonology of the language is about maximizing stressed rhymes and minimizing unstressed rhymes).

I imagine the (former) Turkish claim is simply wrong.

| improve this answer | |
  • I don’t know Saami, but that example seems dubious to me based on my experience with Indo-European languages that have sC “onsets”. Classical Latin for example has s.C word medially, and sC- word-initially, but this isn’t considered to be a counterexample to the onset maximization rule. – brass tacks Dec 17 '18 at 22:47
  • Rather, it’s often analyzed as evidence that word-initial sC- is not a true syllable onset. Anyway, I still have to read the linked papers – brass tacks Dec 17 '18 at 22:53
  • I just skimmed the papers, and the Saami one definitely seems relevant, but complicated! – brass tacks Dec 18 '18 at 7:47
  • What's OT in this context? – jk - Reinstate Monica Dec 18 '18 at 10:36
  • @jknappen: OT = Optimality Theory – brass tacks Dec 18 '18 at 10:37

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.