Monday, December 28, 2015

Raisins from Carthage to Siwa

Most Berber varieties have borrowed the word for "raisin" from Arabic, eg Kabyle azbib, or use a compound "dried grapes", eg Shilha aḍil aqurar. However, in Tunisia, Libya, and Egypt the situation is rather different, as this Facebook post illustrates:
Location Word for "raisin" In Arabic script
Djerba izummucen
Jadu iz/ẓemmuken ايزموكن
Nalut ijemmusen ايجموسن
Zuwara ijemmucen ايجموشن
Yefren, El-Qalaa ijummucen, ijemmac ايجمّوشن, ايجوموشن, اجماش
Siwa ijeṃṃusen إجموسن

The variation in the consonants is not completely regular, but note that there is a regular correspondences between k in Jadu and š in Yefren and Siwa (from palatal *ḱ), and that sibilant harmony is a fairly productive process in Berber.

As far as I know, this word's etymology has not previously been investigated, so I was happy to discover it this morning quite by chance. It happens to be attested in an ostracon from about 2000 years ago (give or take), found at the site of Al Qusbat, on the Libyan coast east of Tripoli:

This line in Neo-Punic - that is, the later Phoenician dialect spoken in North Africa - starts ldn`ṭ' `sr kkr' ṣmq, rendered by Jongeling and Kerr (Late Punic Epigraphy, 2005:24) as "for Donatus, 10 talents of dried fruits". As usual for Phoenician, the interpretation relies mainly on its much better documented close relative Hebrew: in this case, the relevant comparison is to the ṣimmuq-îm צִמֻּקִים֙ "raisins", attested 4 times in the Hebrew Bible. In Hebrew, the root of this word, ṣmq, means "to dry up"; in Arabic, the same root yields the rare forms ṣāmiq "thirsty", ṣamaqah "milk that has gone off". The direction of borrowing is therefore clear: from Phoenician into eastern Berber.

Now most of the attestations of this form are in a region where intense Punic influence is completely unsurprising: the coast of Tripolitania and southern Tunisia. However, any Classicist will remind us that Phoenician rule stopped at the Arae Philaenorum: eastern Libya was in Greek hands, and Phoenician never had any significant presence there. What, then, is this word doing in Siwa? The answer is simple, as I discuss in the introduction to my book Berber and Arabic in Siwa (Egypt): modern Siwi seems to derive mainly from a Berber variety spoken much further west, which reached Siwa only during the Middle Ages. There very probably was a Berber language spoken in Siwa before that, but if so, it has left very few traces.

15 comments:

John Cowan said...

Alternatively (and I know this is rank speculation), it might have been an Egyptoid language. I've always thought the isolation of Egyptian within AA was fairly unnatural: as one of the first Dachsprachen, it must have had relatives at one time that it overset. Granted that at the oracle of Amun they spoke Imperial Egyptian (and eventually Greek), what did the locals speak?

Piotr Gąsiorowski said...

Late Middle Egyptian also borrowed the same Semitic word for 'raisin' (according to Hoch, Semitic words in Egyptian texts of the New Kingdom and Third Intermediate Period). It was written ḏa=ma₃=qa (= *ṣammūqa, Hoch 2014: 388).

PhoeniX said...

It'd be nice if the Late Middle Egyptian would somehow solve the problems here, but that doesn't seem to be the case.

Me and Lameen discussed this word earlier, if we reconstruct this too an earlier stage of the language it would look something like:

*i-ẓămmuḱ-ăn (singular, although unattested: *a-ẓămmuḱ)

Technically, *ḱ should yield *y intervocalically when the first vowel is a full vowel (*a, *i, *u or *e). In the singular *ḱ would yield š.

The issue is in the velar. Proto-Berber had a uvular stop *q, why is it being reflected with a voiceless palatal stop?

The most obvious solution might be that the *q had already shifted in these dialects to (which it has in all Berber languages now, but in Roman times it was probably still a stop in some places). If that is the solution, then one wonders why *ḱ is being used rather than *k. The contrast between the velar and palatal stops is VERY small in Proto-Berber, this is an indication that the contrast perhaps did not exist at all. It'd be nice if one day we could prove that definitively.

Lameen Souag الأمين سواق said...

John: That is very speculative, but the only piece of data we have on the language at that time could be read as supporting it:

the Ammonians, who are colonists from Egypt and Ethiopia and speak a language compounded of the tongues of both countries.
Ἀμμώνιοι, ἐόντες Αἰγυπτίων τε καὶ Αἰθιόπων ἄποικοι καὶ φωνὴν μεταξὺ ἀμφοτέρων νομίζοντες. (Herodotus 2.42)

Piotr: Great find! But how was it pronounced at the time? Ancient Egyptian didn't have a phoneme /ṣ/, did it?

Lameen Souag الأمين سواق said...

PhoeniX: Yes, the *ḱ here is difficult to explain. For why it isn't *q, another possibility would be that, either in Berber or in substandard Punic, dissimilation took place to avoid having two emphatics in one word - just as a lot of modern Arabic dialects would have turned it into *smq. I haven't recorded a singular for it in Siwi, but some of the Libyan Facebook posters attest it: أجمموس for Al-Qalaa, ازموك for Wifat (Jadu?).

Piotr Gąsiorowski said...

PhoeniX, I didn't want to imply that the word was borrowed by Berber from Egyptian. Lameen's hypothesis of a specifically Punic loan is very compelling. I just wanted to point to a parallel case: the same commodity, a related sorce, a neighbouring region. The Egyptian word is attested in an inscription from the reign of Amenhotep III (18th Dynasty, the early 14th c. BC). It was the high point of the Egyptian Empire, much of Syria-Palestine was under Egyptian control, relations between Egypt and Semitic-speaking countries were lively, and hundreds of Semitic loans entered Ancient Egyptian at that time.

Lameen, Middle Egyptian was probably an ejective palatal stop/affricate. It could have been the closest thing to Semitic emphatic , especially if the latter was an affricate. The reverse substitution (AE --> Semitic ) is also attested.

Anonymous said...

Keep the root in mind on which a noun is based it’s also interesting to compare arabic > مَذاق [maḏāq] with western Tamaziγt/Berber mḍy, and Eastern Tamaziγt/Berber mṭg.
In Tachelhit asrs: (raisin)زبيب

Lameen Souag الأمين سواق said...

asrs for raisin? Which region is that from? Mohamed Chafik only gives aḍil aqurar for زبيب in his dictionary.

I've often wondered whether Siwi meddaq "taste (v.)" is from Arabic maḏāq, Berber mḍy, or both...

Y said...

It's straightforward to imagine the Phoenicians introducing dried fruit to North Africa, which then adopts the Phoenician name. To support this, is there anything known about the beginning of viticulture in North Africa?

What about other words for dried fruit? Is there a Siwi or other Berber equivalent to Hebrew דְּבֵלָה, Phoenician dblt 'dried fig'?

(Aside, the word צִמֻּקִים֙ in your post carries a stowaway cantillation mark at the end.)

Jim said...

"To support this, is there anything known about the beginning of viticulture in North Africa?"

Grapes occur naturally over a wide area in the Middle east, Europe and North Africa but the species seems t have been domesticated in the northern Zagros/southern Caucasus. That would comport with a flow of lexical borrowing from east to west.

Lameen Souag الأمين سواق said...

Y: There are quite a few Phoenician loans in the domain of agriculture; depending on the dialect, they include pomegranate, onion, almond, walnut, cucumber, and in one case even olive (though the latter is usually expressed by an inherited word). "Grape", however, is typically aḍil, a word with no apparent connection to Semitic. I can't find anything like dblt for dried fig either (it tends to be tazart). The impression I form is that the Phoenicians added cash crops and luxuries to an already well-established system of small agriculture. Dried fruits would fit that, as being better suited to export.

Jim: That's what we expect for just about any element of the Neolithic agricultural package, but the question is always at what period? Did it come in with the Phoenicians, or with Afroasiatic, or somewhere in between?

Anonymous said...

Hi Lameen, "asrs" was/is used in the Sous region [Morocco], you can find it in this book: Alqamus alamazighi-alarabi, Ibrahim al Isafani, 18th century

There could be adopted names of fruits/vegetables imported from asia/middle-east, but it isn't always the case. There are examples where you can see a local name given to a fruit. In medieval times there where two words used for pomegranate.

Jim said...

"but the question is always at what period? Did it come in with the Phoenicians, or with Afroasiatic, or somewhere in between?"

I guess the question in answer to that is whether the Afroasiatic expansion - or rather, the Semitic/Berber split - pre-dated or post-dated the Neolithic revolution. My guess is that it happened mid-way, and that some cultigens came along and others didn't, for whatever reason. And obviously any domesticated after the split could not have been part of the initial package of cultigens.

kpozin said...

All of the locations you mentioned were previously part of the Ottoman Empire. The Turkish word for grape is üzüm and has derivations including üzümcü (grower or seller of grapes).

A very possible explanation is that all of these local terms were not Semitic in origin, but rather adaptations of the Turkish word for grape (which allegedly was *jüŕüm in Proto-Turkic); the similarity to צמוק is probably just a coincidence.

(Incidentally, the word for raisin in Russian is also изюм [ɪˈzʲum], which is known to be a Turkic loanword.)

Lameen Souag الأمين سواق said...

Interesting, but almost certainly coincidental. There is no ending -uk in Berber (or, as far as I know, Turkish) that would explain the Jadu form, and I would expect z rather than ẓ from a Turkish word with front vowels anyway. Moreover, Siwa wasn't brought under Turkish rule until Muhammad Ali, and all Turkish words in Siwi seem to have arrived via Arabic, rather than being borrowed directly.