« What is the deal with word-final ʔ nouns? | Main | New Article: Noun Prefixes in Eastern Berber »




Rāyta and rāytum would violate basic Arabic phonotactics - and the orthography rules out rayta, raytum. That suggests that these forms, at least, still had a glottal stop.

yarā lacks the glottal stop not just in the reading traditions but in standard Classical Arabic too, so I'm not sure what your point is in the last paragraph.


Thanks for your comments! I updated the final paragraph a little bit. I hope it's more clear now.

As for the phonotactics: I'm not sure if I'm too worried about phonotactic arguments. Phonotactics can change, and they do, all the time, of course.

Classical Arabic doesn't allow for superheavy syllables at all, yet, in most dialects Arabic allows for superheavy syllables in word-final position (due to the loss of case vowels).

In the Minabbih dialect which has rāy, this verb did get changed to the phonotactics, e.g. rayt 'I saw'. You might imagine that the moment ʔ was lost, it was instantly shortened, but that doens't have to be the case of course.

raʔaya > rāy(a)
raʔaytu > (rāytu) > rayt(u)

I don't think that we can make the a priori assumption that the stage in between brackets never existed, and I would be inclined to say the QCT confirms it.

I have trouble formulating a sound law where:
raʔaya underwent a shift *aʔa > ā but
raʔaytu did not.

You would have to assume a syncope of a of some sort: rʔaytu, and that syncope would have to post-date the *Cʔ > C development. This would give us a phonotactically problematic word-initial cluster, and we have no evidence for such a syncope (it would be hard to find evidence for it, of course).

Or one could write a conditioned sound law for the loss of ʔ like: aʔaCV > āCV. while aʔaCC remained. But the phonetic motivation for such a sound law seems rather limited. I would rather assume a stage of the language where internal superheavy syllables were allowed for some period of time.

Without the sound law aʔa > ā preceding āyv >āʔv though, the form and becomes inexplicable. We would be stuck with a Classical Arabic-like pronunciation of two verbs which just happened to have the same root structure C-ʔ-y, while all other final weak verbs with -y are written with , and rhyme seems to suggest that they were pronounced ē or ay.

So whatever form we assume for the aʔa > ā shift, I think it needs to have happened for the 3sg.m.


There are also actual cases of super heavy syllables in Classical Arabic occasionally as Ahmad Al-Jallad pointed out to me. The Energic ending on dual verbs creates super heavy syllables as in Q10:89 yattabaʕānni. So clearly whatever phonotactics were in place (and anti-superheavy phonotactics were in place of course with wādiyun > wādīn > wādin), they don't seem to be active in Classical Arabic.


You also regularly get super heavy syllables with derivations from double roots, e.g. ḍāll- "erring" or (al-)ḥāqqa(h).
Of course, one might argue that doubled consonants consonants following a long vowel create a somewhat different environment than two different consonants following a long vowel.


addendum: Though personally, I think that these should be treated necessarily as different environments.


Thanks for your comment Daniel!

I was going to mention those too. But I agree that a case could be made that VVC1=C2 is a different 'kind' of superheavy syllable.


In Classical Arabic poetry CVVC (superheavy) is treated as CVC (heavy) for the purposes of scansion and there is no difference between them. This could imply that superheavy syllables were actually abundant in early Classical Arabic and that phrases like "mā smuka?" were actually pronounced with a long ā.

The standard Classical Arabic syllable types of CV and CVV/CVC may just be a realignment from an earlier system that was mora-timed (like Japanese), and there were two coda-less syllable types C/CV (short) and CVV (long) that may have both been considered one mora. Classical CVC would have developed from bimoraic C.C (bin.t < b.n.t), CV.C (man < ma.n) and CVV.C sequences (mas.mu.ka < maa.s.mu.ka)


As I said, I see no reason to not break phonotactics if it serves the description.

Phonotactics can occasionally be a very strong tool to understand a whole range of historical developments, and super-heavy syllable avoidance, explains all kinds of behaviors of II-weak and III-weak verbs in Semitic. But when it doesn't, it doesn't, and of course Arabic clearly develops means to avoid such phonotactics at some points, and the moment when this happens is rather difficult to pinpoint. I wouldn't mind placing that before the period that Arabic was first set to writing in the Arabic script.

Your mora-based analysis seems sensible, but I'm not too knowledgeable of moraic theory, so I'm hardly one to judge.

The comments to this entry are closed.