Thursday, October 23, 2014

Berber: classification, Tasahlit, roots vs. stems

Today seems to be a good week for comparative Berber linguistics - the day's haul is worth sharing:

Maarten Kossmann has uploaded his preliminary classification of Berber varieties based on shared innovations: Berber subclassification (preliminary version). He divides Berber into seven blocks:

  1. Zenaga block (Zenaga of Mauritania, Tetserrét in Niger)
  2. Tuareg block
  3. Western Moroccan block (SW Morocco, Central Morocco, i.e. Tashelhiyt and most of Tamazight)
    possibly including NW Moroccan Berber (Ghomara, Senhadja de Sraïr)
  4. Zenatic block (Eastern Morocco, Western Algeria, Saharan oases, Tunisia, Zuara) extending towards the east with Sokna, Elfoqaha, Siwa
  5. Kabyle (N Algeria), possibly linked to the western Moroccan block
  6. Ghadames (Libya), probably to be linked to Djebel Nefusa (Libya)
  7. Awdjilah (Libya)
By and large, this appears very plausible, although it should be noted that Tunisian Berber and Zuwara are already somewhat peripheral to Zenati, not sharing western Zenati's innovative distribution of initial vowel dropping, and El-Fogaha is even more so than Siwa or Sokna. (As he notes, the much greater homogeneity and clearer boundaries of Zenati in the west imply that this group arrived in Algeria and Morocco from the east.) But, in principle, it is still necessary to identify specific innovations characteristic of each of these groups. It is also clear that the Zenaga block is by far the first split on the tree, and the list ought ideally to reflect that. But the moderately high degree of mutual intelligibility poses serious obstacles to applying the family tree model to Berber, as he discusses.

The most interesting Kabyle varieties for historical reconstruction are the little-known ones of the extreme east, "Tasahlit". As it happens, Abdelaziz Berkai has just uploaded his recent thesis, a dictionary and sketch grammar of the Tasahlit of Aokas: Essai d’élaboration d’un dictionnaire Tasaḥlit (parler d’Aokas)-français. The quality of his work appears excellent, and this will no doubt be a very useful resource. The choice of dialect, however, is not entirely ideal. It is clear from Basset's dialect atlas, and from the all too rare comments in Rabdi's grammar on neighbouring varieties, that the vocabulary of Aokas is still quite close to that of Bejaia; the really divergent varieties seem to be those of the Babor Mountains and Oued el Bared, approaching Jijel, and those are the ones most likely to give an insight into the dialect of the now largely Arabised Kutama.

I haven't yet had time to properly look at Samir Ben Si Said's thesis, De la nature de la variation diatopique en kabyle: étude de la formation des singulier et pluriel nominaux, but it tackles the synchronically as well as diachronically thorny problem of Berber non-concatenative morphology, and argues for an approach based more on roots than on stems, contrasting with another important study I've been working through lately, Heath's Grammar of Tamashek (Tuareg of Mali).


PhoeniX said...

Thanks for the link on Tasaḥlit! I had not seen that yet.

I just read through the Berber subclassification paper.

Just some thoughts from my perspective:

I think it's incorrect to state that Zwara did not undergo the shortening of the prefix so typical of Zenati. It's just that the prefix has not dropped completely, because Zwara retains accented ə in open syllables.

ə́fus 'hand' (EL) úfus (EA) fús-is 'his hand'.

The places where Zwara has ə- prefixes is exactly in the places where you find the absence of a prefix in Zenatic. Which, due to Zwara's retention can more or less be shown to be a shortening, before it was dropped completely. (which, incidentally makes the development look a lot like shortening, but more irregular).

And the Yefren Nefusi dialect seems to simply have the complete shortening (like Siwa)

I'm not sure about the Tunisian varieties (then again, who is?), but you're probably right about them avoiding that development.

I myself am inclined to state the relationship between Tashl./MA and Kabyle a bit more strongly now. I think the 'loanword-morphology class' that those three have developed almost identically can probably be considered a shared innovation. Besides that innovation, most stuff they share just seems to be shared retentions though.

I think Ghomara is most closely related to Tashlhiyt. The odd innovation of lengthening the initial consonant in the impf. for triradical roots like ɣrs > qqrs is a little too odd to've developed independently in those two languages separately.

Lameen Souag الأمين سواق said...

It's not the shortening of the prefix that's typical of Zenati - that happened in Tuareg too, after all. But yes, the distribution in Zwara is the same as in Zenati or Tuareg (well, when penultimate). In Tamezret and Zraoua it's pretty similar, but afus "hand", afud "knee" do not show shortening, and the schwa remains even when unstressed. I don't know much yet about the rest of Tunisia. (Btw, in Zuwara Mitchell's article has wə́fus, not úfus).

Which loanword-morphology class? Do you mean impfv. ttəCCaC? But that has Tuareg cognates.

Yes, that struck me in Ghomara too - also the final -a in CC' imperfectives.

PhoeniX said...

You're write about wə́fus of course, I misremembered, and forgot to check.

I'm talking about ttəCCaC in terms of the loanword morphology class.

Of course, technically that class is around in Tuareg (and all of Berber for that matter), for non-Conjugation I (and maybe II?) type verbs.

But I'd say it being appropriated for Conjugation I.A verbs whenever they are of Arabic origin is a significant innovation within Tashelhiyt/Middle Atlas Berber/Kabyle.

But of course Conjugation III verbs in Tuareg CăCăC would look exactly like Conjugation I verbs of the type ăCCəC in these languages.

One could argue that those merged, and that that gave rise to two different possibilities for CCC verbs, after which one was assigned to loanwords and the other to native words.

But if that were the case, I would expect to find at least some stem III type verbs to've survived in Tashl./MA/Kb. with cognates in Tuareg. I've looked pretty hard for those (mostly in Tashelhiyt), but they don't seem to exist.

I think it rather developed from Arabic stem II verbs being incorporated in the CCCC verb class, which naturally had the tt-CCaC.

Perhaps the Stem III verbs with CaCC shape was close enough to CuCC impf. tt-CuCuC to form an analogous: tt-CaCaC.

With those two incorporations, the vast majority of tt-...aC verbs were Arabic loanwords, which was then interpreted as a rule: tt-...aC is the imperfective formation for loanwords, which gave rise to the Stem I Arabic loanword formations.

I think that's probably a logical evolution (although of course still a little speculative, there's a reason I haven't written this up as an article yet...), but despite it being logical it doesn't seem likely that it developed in two otherwise unrelated language groups, which is why I think it's a shared innovation between Kabyle and Tashl/MA.

If you don't accept that development, you'd probably have to argue for all of Zenatic to lose the tt-CCaC formation as a shared innovation, which is also possible.

PhoeniX said...

Could you elaborate on the Zenati shortening?

From what I can tell from your classification in 'Berber and Arabic in Siwa (Egypt)' you see the shortening of the prefix vowel as a two-stage process:

One taking place in disyllabic words that is common to Tuareg and Zenatic:

aCVC > ăCVC (like afus > ăfus)

But as you mention, it seems to be irrespective of syllable count in Tuareg.

And then a second Zenati specific shortening taking place in trisyllabic nouns:

afunas > funas, aɣanim > ɣanim.

Does ifiɣər > fiɣer /igidər > gidər also count for that one? (Because Zwara partakes in that one).

Is it really necessary to split those two up? Because neither shift is regular in any of the languages (e.g. Riffian has afunas but ɣanim/žiḏā/fiɣā; ṯamuaṯ 'country' but ṯmāṯ 'beard')

In my impression the shortening in Tuareg applies to both the bisyllabic and trisyllabic nouns. Although it seems a little less regular in the trisyllabic nouns.

Anonymous said...

Thèse sur Tamazight le "Berbère" de
Tunisie ( Douiret)

An outline of the Shilha (Berber) vernacular of Douiret (Southern Tunisia) . A télécharger ici :

Lameen Souag الأمين سواق said...

Within Siwi, the status of afus > fus is clearly synchronically distinct from afunas > funas. In penultimate position, prefix vowel dropping is purely lexical, and you get contrasts like ta-ẓa "dish" vs. t-sa "liver". In antepenultimate position, however, prefix vowel dropping is productive enough to apply systematically to Arabic loan adjectives whenever the first syllable of the stem has a full vowel: contrast šarəf, t-šarəf-t "old (person)" with a-qdim, ta-qdəm-t "old (thing)". Diachronically, most instances of penultimate prefix vowel dropping are associated with *ʔ, while antepenultimate prefix vowel dropping shows no such correlation. More work is needed, but yes, I think the two phenomena are probably historically separate despite their similarities.

igidər > gidər probably is the same change as afunas > funas, but it seems impossible to reduce izimər > izmər (and igidi > ijdi, inisi > insi) to either of these two changes. And unlike the other two, it really does seem to be specific to Zenati.

I don't think there's necessarily anything irregular going on in tamurt vs. tmart - rather, like aduf "marrow" vs. fud "knee", it shows that penultimate prefix vowel dropping was conditioned by a factor other than syllable structure. The distinction is not just in Tarifit; every Zenati language I've looked at has a full prefix vowel in tamurt (Siwi taməṛt) and a reduced one in tmart, and that consistency would hardly be likely if there weren't some underlying factor governing it. "Cow" is more problematic - I need to look at its distribution...

Lameen Souag الأمين سواق said...

La thèse de Gabsi a l'avantage d'être presque le seul travail sur le parler de Douiret. Mais, pour le parler de Tamezret, le site de Ben Mamou offre une description beaucoup plus précise et mieux transcrite, malgré le fait que l'auteur n'a pas reçu un doctorat pour ce travail :

PhoeniX said...

Those are some good points!

There's certainly an article waiting to be written about prefix-shortening in Zenatic...

Just to challenge some of the ideas:

You mention aduf. In my forthcoming article in WZKM where I re-examine the o-/u- initial nouns in Ghdadames and Nefusa I discuss that word. But an early version of the idea is already present in my MA thesis.

aduf is part of a cluster of three words which have a funny vocalic distribution. The other words are lum 'hay' and suf 'stream'.

These words have in common that they have a vowel i in the stem of Kabyle/Tashelhiyt/Tamazight, and a vowel u in Zenatic and Tuareg.

Moreover lum and suf have Ghd. oləm Nef. uləm and Nef. usəf.

I argue that this odd i~u correspondence must mean they had similar phonetics at some point. And I suggested that the Zenaga glottal stop that we find for 'marrow': äḏiʔf shows that the glottal stop plays a role in it.

I argue that maybe the original form was *adiʔf/aliʔm/asiʔf, and that maybe iʔ yields u in Zenatic/Tuareg (drawing upon the rounding effect of ʔ which also causes ăʔ > o in Ghd./Nef.

If you accept that this similar vowel distribution suggests that they probably had the same Proto-Berber stem-shape, then the fact that aduf has a prefix vowel and lum/suf do not is an issue if that shortening is supposed to be regular.

It's not like there's no way around this of course (you could reconstruct *adiʔf/*aʔliʔm/*aʔsiʔf for example and save the i/u correspondence AND explain why adiʔf doesn't have the prefix shortening the other two words have, for example), but like you said it needs further examination.

David Marjanović said...

[ʔ] has a rounding effect? ~:-|

...Does it perhaps have a lowering effect that turned ă into [ɑ] or thereabouts? [ɑ] easily turns into [ɒ] "by itself".

PhoeniX said...

Yes maybe, it's not exactly satisfying, I agree.

But whatever the case, the reflex of *aʔ and *aw both turn into o in Ghadames. And there seems to be some kind of connection between *ʔ and *w (Zenaga often has variation between the two, and some verbal nouns have a w showing up in the position of *ʔ).

So yes, if what we reconstruct as *ʔ is an actual proper regular glottal stop, then it probably isn't a rounding effect of a glottal stop, because it doesn't have one.

But the actual phonetic value of what we reconstruct as *ʔ might be a bit more complex than that.

David Marjanović said...

I see.