Saturday, April 23, 2016

Arabic substrate etymologies as urban legends

In Arabic as in English, social networks have a constantly flowing undercurrent of poorly sourced, manipulative stories being shared and reshared by people who vaguely think they sound right. Over the past, say, five years, I've noticed the emergence of a linguistically interesting new subgenre within this miasma of lies and half-truths: etymological tables purporting to prove the massive contribution of Berber, or Syriac, or (more rarely) Coptic, or perhaps some other pre-Arab substrate to the local Arabic dialect. These tables, in my experience, never cite an academic source, and rarely cite anything at all; closer examination generally reveals a farrago of correct etymologies and bad guesses. For example (from the preceding links):
  • Tunisian məlɣiɣa ملغيغة "fontanelle" really is from Berber tamelɣiɣt, a word widely attested in Berber and with no obvious Classical Arabic counterpart...
  • but Tunisian gdər قدر "pot" is of course from the Classical Arabic qidr قِدْرٌ, which ought to be familiar even to elementary school students; the Berber cognates cited are borrowings from Arabic.
  • Tunisian bəkkuš بكّوش "dumb, mute" is slightly less obvious, but again from Arabic: it's an irregular expressive formation from 'abkam أَبْكَمُ, substituting the dialectally rather productive suffix -uš. The suffix might be from Berber, but the root is not.
  • Syrian (and Algerian) dālye دالية "grape-vine" may well be from Aramaic; the word is attested in Syriac with the right meaning (dālī-ṯ-ā "vine-branch, vine"), and belongs to a semantic field where Aramaic borrowings are to be expected from a very early period. Within Arabic, this word was already noted as a regional synonym of karmah in the 10th century by the Palestinian geographer al-Maqdisi.
  • However, Syrian mnīħ منيح "good" has nothing to do with Aramaic; it's a local version of widespread dialectal Arabic malīħ مليح, with nasality assimilation. This adjective exists both in Classical Arabic (malīħ) and in Syriac (malīħ-ā) with the meaning of "salty"; in an era where salt was more expensive than now, this naturally tended to imply "tasty". There is no reason to assume either language borrowed this word from the other, since the root is proto-Semitic and the template is productive in both languages. However, only in dialectal Arabic did it go on to develop the sense of "good", which it now has in a wide variety of dialects including North Africa.
  • More problematic is Syrian wāwā, a baby-talk word for "pain" used (as far as I can see) neither in Syriac nor in Classical Arabic. Syriac does have wāy "woe!", but so does Coptic - and, if it comes to it, English "waaah!" is closer than either. Onomatopeia is a better explanation than borrowing or inheritance in this case.

The optimistic take on this is that it shows that there's a real public demand in the Arabic-speaking world for information on etymology and on substrate influence. The pessimistic take is that people just want "information" confirming what they want to believe - in this case, that they're not really that Arab after all. (The converse case also exists, of course - recall Othmane Saadi - but I haven't seen as much of it circulating on social media, though that may just reflect my own bubble.) The reality is probably somewhere in the middle.

14 comments:

M said...

Thanks for your new post !

So, if I understand well, Dalia (that is used in Morocco as well) is likely to be an Aramaïc word.
However, I would like to know when did the word appear for the first time in Arabic dictionnaries? Was it a word that was used only in the Levant, where the substrate is Amaraïc and that has, afterwards, be added to the dictionaries ?

Lameen Souag الأمين سواق said...

If it is from Aramaic, it was certainly borrowed within the Levant or Iraq, and was brought to North Africa by Arab immigrants. Very likely it was borrowed before Islam. In any case, it seems to be attested in Iraq in an Arabic manuscript of 852 AD (http://www.apd.gwi.uni-muenchen.de:8080/apd/asearch.jsp?searchtable1=1001&searchwordstring1=d%C4%81liya%E1%BA%97&showdwords=true&showlemmata=true), and no doubt a closer look would uncover earlier attestations.

Imed Adel said...

Talking about etymologies, is it possible to get the root of words in Darija such as dda/yddi, ja/yji, xda/yax°(u)d, qra/yqra, and nbi? I have tried following the method used by Arabic grammarians, but many of these words are problematic: Do you consider 'i' as 'ey' or just a vowel 'i'? Do you take into consideration the etymological glottal stop or not (you shouldn't because you are trying to generate the root from the existing words and not from their etymological forms)?

I've also tried the Berber way: drop all the vowels and the repeated consonants. Quite useful, but is it right?

P.S. Take the verbs klika/ykliki (to click) and pṛugṛama/ypṛugṛami (to program) for example, applying the Arabic method gives 'KL(Y)KY' and 'PṚ(W)GRMY' while the Berber Method gives 'KLK' and 'PṚGṚM'.

Lameen Souag الأمين سواق said...

To be useful for lexicography, a root should ideally be something from which you can predict all the derived forms, and which you can find given any of those derived forms. Whether it consists only of consonants or includes vowels as well should be a secondary consideration, derived from language-internal facts rather than taken as a priori. The customary Berber way fails in this respect not only for Darja but for Berber as well: not only is it impossible to predict the conjugation of a Berber verb from its consonantal "root", but completely unrelated verbs end up having the same "root". In both Standard Arabic and Darja, it would seem theoretically preferable to include unpredictable short vowels in the root. However, such vowels are relevant to inflection but not to derivation, which makes it lexicographically convenient to drop them. Long vowels, on the other hand, often have to be treated as part of the root in both Standard Arabic and Darja. In Standard Arabic this fact is disguised by rewriting them as semivowels, but I see no reason to extend this practice to Darja. In particular, typical French loan verbs' vowels are stable in all derived forms (except for the final vowel), so there's really no possible justification for excluding their vowels from their roots.

xđa and kla are irregular verbs, so the root is naturally a bit trickier; but derivations like makla, maxđ suggest a root axđ, akl.

Imed Adel said...

Thank you so much. So you basically you suggest the roots "ddi, ji, pṛugṛami, kliki, -axd, -akl"?

What do you suggest for a noun without a verb? Should it be used as it is?

Since my questions are related to a work about the names of species in Darja, do you suggest marking the interdentals (e.g. dib or ðib)? Whenever I think about this issue, I remeber the words tlata/θlaθa/tlaθa, kaġət/kaġəθ, and yfədlək/yfəðlək. Can you give me your opinion about this? (You marked the interdentals in all of your works and writings).

Imed Adel said...

Sorry for confusion between interdentals and dentals.

Lameen Souag الأمين سواق said...

Yeah, those would do fine as roots for these words. As for nouns without verbs, that would depend on their morphology, and arguably more generally on their shape. You could quite reasonably argue that if xəbbaz "baker" has the root xbz, həjjal "widower" should be assigned the root hjl, for instance; but it would be much harder to argue that maṣṣu "mason" should have any particular root, because nothing about its shape would allow you to break it up into a root plus a discontinuous morpheme.

Lameen Souag الأمين سواق said...

As for interdentals and dentals, it depends what you're trying to do. For documenting the language, of course you should mark them. For creating a common standard, it could be argued that not marking them makes it easier for speakers of dialects that don't distinguish them.

Imed Adel said...

Thank you a lot. If I'm not marking interdentals, should I use "ḍ" for every word pronounced with [dˤ] or should I use it only when it can be pronounced with both [ðˤ] and [dˤ]? I mean for example, should I write "ḍuwwaṛ" or "duwwaṛ" (in the second case, the "ḍ" is used with words such as "ḍbeɛ" with [ðˤ])?

Lameen Souag الأمين سواق said...

That's quite a list - thanks for sharing it. Probably most of it is of Berber origin, but I can see several problematic cases - šlāɣəm and sərdūk, for example, are quite unlikely to be Berber. gnina for "rabbit" is definitely not from Berber, but rather from Latin or more recent Romance. Reviewing the whole list would be a huge job...

Lameen Souag الأمين سواق said...

Not a bad argument as far as it goes, but a closer look at the regional variation indicates otherwise. Within Arabic, it's attested with q and g (qunn in Bechar); such q/g variation is well-attested for Romance borrowings, but very rare in borrowings from Berber. Moreover, in light of qunn, gnina is easily recognised as a diminutive of *gunn, just as cuniculus is of cun-; Andalusi Arabic qunayla reflects the Romance diminutive suffix. Finally, within Berber agnin has a very limited distribution, and nothing suggests that it can be reconstructed for proto-Berber; the proto-Berber word for hare is clearly ayerẓiẓ. Almost certainly, the few Berber varieties that have this word got it from Arabic, and Arabic got it from Romance/Latin.

As for sərduk, I'm not sure it's even attested in Berber at all; certainly it has no connection with the dominant Berber roots for "rooster" and "chicken". Behnstedt and Woidich consider its etymology unknown, but probably it's some kind of compound based on Arabic dīk.

Lameen Souag الأمين سواق said...

Another non-Berber one is šnāfa "lip", cf. Arabic šafah. A better attested, clearly non-borrowed Berber form for "lip" is aḍlis.

Lameen Souag الأمين سواق said...

There are, yes, but the latter is geographically much more widepread than the former. As a matter of fact, the distribution of the two (awtul being limited to the north, basically) makes it possible to imagine that at some point maybe awtul was the word for (wild) rabbit, while ayerẓiẓ was the hare proper.

Imed Adel said...

A recent tendency among Tunisian Darijists is trying to prove that everything in Tunisian Arabic is actually from Punic. This includes even the verb conjugation.

It's weird how people ignore all the linguistic papers and try to come up with their own conclusions....