La diversité des noms de prières islamiques et ses origines

(English translation)

Presque chaque musulman, quelque soit sa confession ou sa langue, et chaque arabe, donne aux cinq prières quotidiennes de l'Islam des variantes des même noms : Fajr / Ṣubḥ avant l'aube, Ḍhuhr à midi, `Aṣr à l'après-midi, Maghrib au coucher du soleil, et `Ichā’ au soir. Mais à partir de l'ouest de l'Egypte, la situation est tout autre, et on trouve des noms totalement différents pour les cinq prières. Par exemple, comparez les noms bien connus donnés ci-dessus avec ceux qui sont utilisés dans la plus grande langue berbère, le tachelhit (sud Marocain): ṣṣbaḥ, tizwarn, takʷẓin, tiwutch, et tin-yiḍs. Sauf le premier, ils semblent être sans relation.

Or, les noms des prières au Moyen-Orient n'ont pas toujours été aussi standardisés que maintenant. Al-Boukhārī nous donne le hadith suivant (#516):

“... Sayyâr ben Salama a dit : "J'entrai avec mon père chez Aboû-Barza. Mon père lui demanda comment l'Envoyé de Dieu faisait la prière canonique. “Il faisait, répondit-il, al-hajīr, que vous appelez al-’ūlā, aussitôt que le soleil déclinait ; puis, quand le soleil avait baissé, il faisait al-‘aṣr et, celle-ci finie, on avait le temps de retourner à sa demeure située à l'extrémité de Médine (pendant que le soleil était encore bien vivant)." J'ai oublié ce qu'il a dit d'al-maghrib. "Et il préférait de retarder al-‘ishā’, que vous appelez al-‘atama ; il n'aimait pas dormir avant cette prière, ni causer après elle. Il retournait d'al-ghadāt au moment où on y voyait de façon à reconnaître son voisin de prière, et il y récitait de soixante à cent versets du Coran.””

En siwi, la langue berbère de l'Egypte, on utilise encore une série de noms qui pourraient avoir été pris presque directement de ce hadith : ils appellent les cinq prières sra (le matin), luli, la`ṣaṛ, mməghrəb, et l`ətmət. On trouve des traces de ces noms même plus loin ; au songhay (une langue de Mali et Niger), Dhuhr s'appelle aluula.

Ce même hadith explique le nom tachelhit du Dhuhr : tizwarn veut dire littéralement "les premières", une traduction littérale de l'arabe al-’ūlā. Ce terme n'est pas seulement répandu en berbère ; il est également, grâce au zénaga, la source du mot wolof (Sénégal), tisbaar. En soninké, la langue de l'empire médiéval de Ghana entre la Mauritanie et le Mali, une autre traduction littérale nous donne sállì-fànà (“prière-premier”), qu'a emprunté le bambara et beaucoup d'autres langues ouest-africaines.

Un autre hadith, moins bien attesté (Maṣḥaf `Abd al-Razzāq 2067) explique également le nom tachelhit de l'Icha:

“De Yaḥyā ben al-‘Alā’, d'al-A‘mash, d'Abū Wā’il qui a dit : J'ai demandé Ḥuḏayfa, et il m'a dit : Pourquoi m'as-tu demandé ? J'ai dit : Pour la conversation. Il a dit : “‘Umar ben al-Khaṭṭāb, que Dieu l'agrée, nous mettait en garde contre la conversation après ṣalāt al-nawm (la prière du sommeil).”

En comparant avec d'autres versions du même hadith, on voit que la prière en question est l'Icha. Et en fait, tin-yiḍs en tachelhit veut dire littéralement "celle du sommeil". Cette forme est répandu en berbère, et elle a été traduite littéralement en soninké comme sákhú-fó (sákhú "sommeil", fó "chose"). Cette forme à son tour est introduite en songhay (saafoo) et quelques autres langues de la région.

Cela implique que :

• La terminologie islamique du berbère a été créé très tôt dans l'histoire de l'Islam, avant que ces formes variantes n'ont disparus de l'usage arabe ;
• Les soninké et les wolof ont appris l'Islam en grande mesure des berbères, et pas directement des arabes ;
• Les soninké ont joué un rôle important dans la propagation de l'Islam à des autres groupes ethniques en Mali et Niger.

Il implique aussi, plus généralement, qu'il ne faut pas trop vite traiter les traditions islamiques spécifiques à une région particulière comme des innovations.

(Ce billet résume une article à moi qui sera publié au Bulletin of SOAS, titrée "Archaic and innovative Islamic prayer names around the Sahara" [Noms archaïques et innovatifs des prières islamiques autour du Sahara].)

Diversity in Islamic prayer names and its roots

(Traduction en français)

Practically all Muslims whatever their sect or language, and all Arabs, refer to the five daily prayers of Islam by versions of the same names: Fajr / Ṣubḥ before sunrise, Ḍhuhr at noon, `Aṣr in the afternoon, Maghrib at sunset, and `Ishā’ in the evening. West of Egypt, however, things look very different: there we often find quite different names for the five prayers. For example, contrast the familiar names above with those used in the largest Berber language, Shilha (southern Morocco): ṣṣbaḥ, tizwarn, takʷẓin, tiwutsh, and tin-yiḍs. Except for the first, there seems to be no relation.

However, Middle Eastern Islamic prayer names haven't always been so uniform. Al-Bukhārī reports the following hadith (#516):

“... from Sayyār b. Salamah: “My father and I entered into the presence of Abū Barzah al-’Aslamī. My father asked him: “How did the Messenger of God, blessings and peace be upon him, pray the prescribed (prayers)?” He replied: “He used to pray al-hajīr, which you (pl.) call al-’ūlā, when the sun declines (from the meridian), and pray al-‘aṣr such that one of us could return to his home at the far end of Medina while the sun was still lively.” I forget what he said about al-maghrib. “And he used to prefer to delay al-‘ishā’, which you (pl.) call al-‘atamah, and he used to hate sleeping before it or speaking after it. And he used to return from the ghadāt prayer when a person could recognise the one sitting next to him, and read sixty to a hundred (verses.)””

Siwi, the Berber language of Egypt, still uses a series of mostly Arabic-derived prayer names that might as well be taken straight from this hadith: they call the five prayers sra (morning), luli, la`ṣaṛ, mməghrəb, and l`ətmət. Traces of these names are found further afield too: in Songhay (Mali/Niger), Dhuhr is referred to as aluula.

It turns out that this hadith also explains the Shilha name for Dhuhr: tizwarn literally means "the first ones (f.)", a literal translation of Arabic al-’ūlā. This form is not just widespread in Berber, but is also (via Zenaga) the source for Wolof tisbaar. In Soninké, the language of the medieval Ghana Empire between Mauritania and Mali, another literal translation yields sállì-fànà (“prayer-first”), which has been borrowed into Bambara and many other West African languages.

A similar, much less well-sourced hadith (Maṣḥaf `Abd al-Razzāq 2067) likewise explains the Shilha name for Isha:

“From Yaḥyā b. al-‘Alā’, from al-A‘mash, from Abū Wā’il who said: I asked for Ḥuḏayfa, and he said: Why have you asked for me? I said: For conversation. He said: “‘Umar b. al-Khaṭṭāb, may God be pleased with him, used to warn against conversation after ṣalāt al-nawm (the sleep prayer).”

Comparison to other versions makes it clear that the prayer being referred to is Isha. As it happens, Shilha tin-yiḍs means, literally "that (f.) of sleep". This form is widespread in Berber, and was literally translated into Soninke as sákhú-fó (sákhú "sleep", fó "thing"). The resulting form was borrowed into Songhay (saafoo) and several other regional languages.

All of this tells us three things:

• Berber Islamic terminology was created very early in Islamic history, before these variants disappeared from Arabic usage;
• Soninké and Wolof speakers adopted Islam in large part from Berbers, not directly from Arabs;
• Soninké speakers played an important role in the spread of Islam to other ethnic groups in Mali and Niger.

It also has a wider moral, though: that we shouldn't be too hasty to dismiss region-specific Islamic traditions as innovations.

(This post summarises about half of an article of mine which is forthcoming in the Bulletin of SOAS, under the title of "Archaic and innovative Islamic prayer names around the Sahara".)

So how different are Algerian and Egyptian Berber?

In the previous post, we looked at how hard Egyptian Arabic was for an Algerian to understand (answer: not that hard) and how it diverges from the rules of Algerian Arabic (answer: a lot). What if we try the same exercise for Berber – specifically, Kabyle vs. Siwi? Obviously, I don’t speak Kabyle fluently or even well; if you do, feel free to correct me or give me your own impressions. However, with a bit of help from a dictionary, I think it’s worth a try. There aren’t any Siwi stories recorded online at the moment, about Juha or otherwise, but here’s a short fable with a sad ending retranscribed and retranslated from Laoust’s grammar:

Azidi dilla g adrar, itessu aman. Tizmert ttella adday. Azidi yeṃṃ-as: “Itta xeḅḅecṭ-i aman nnew?” Tizmert teṃṃ-as: “Aman dillan g ɛali, iteggezen i gda!” Yeṃṃ-as: “Ɛam-nuwwel nic uṭnaxa, cemm edduqqaṭ ṭaren nnem!” Teṃṃ-as: “Nic n aseggasa!” Yeṃṃ-as: “Namma eṃṃa nnem namma axxa nnem!” Baɛdin yečč-ét.

There was a jackal on a mountain, drinking water. A ewe was below. The jackal said to her: “Why have you muddied my water?” The ewe said: “The water is above, and goes down to here!” He said: “The year before last when I was ill, you stamped your feet (disturbing him with the noise)!” She said: “I’m from (I was born in) this year!” He said “Or (it was) your mother, or your aunt!” Then he ate her.

Only seven words (out of 44) have no cognates in Kabyle as far as I know – in three cases, this is because one language or the other has borrowed an Arabic term:

• azidi “jackal”: in Kabyle this would be uccen.
• yeṃṃ-as “he told her”: in Kabyle this would be yenn-as.
• itta “why”: in Kabyle this would be ayɣer. The Siwi form is from i “to, for” and -tta < tanta “what”, a local variant of widespread Berber matta, which Kabyle has replaced with the Arabic loan acu.
• ɛali “above”, from Arabic: in Kabyle this would be asawen, but the Siwi form is easy to guess from Arabic.
• iteggezen “they go down”: in Kabyle this would be trusun.
• ɛam-nuwwel “year before last”, from Arabic: in Kabyle this would be sell-ilindi, but the Siwi form is easy enough to guess if you know Algerian Arabic.
• namma “or”: the first syllable is cognate to Kabyle neɣ, but the word has changed enough to make guessing difficult.
• axxa “aunt (mother’s sister)”: in Kabyle this would be xalti, from Arabic.

So in terms of vocabulary, the situation is pretty similar to what we saw between Algerian and Egyptian Arabic. However, as with the previous example, there are many more subtle differences – and those differences are of a more significant kind. If we look at grammatical differences alone and ignore phonetic or semantic ones, we notice that:

• g adrar “in the mountain”: in Kabyle this would be g wedrar; Siwi has no “état d’annexion”.
• dilla, ttella: “he is at, she is at”: Kabyle yella, tella, with no d- prefix. adday “below”: Kabyle does have a noun adda “below”, but it can only be used in combination with certain prepositions, not on its own as here.
• xebbecṭ-i: Siwi marks the 2nd person singular (“you”) with just -(a)ṭ; Kabyle uses t-...-ḍ.
• nnew “my”, nnem “your (f.)”: in Kabyle this would be inu, inem.
• iteggezen: Siwi marks the 3rd person plural (“they”) with y-...-en; Kabyle, like all other Berber languages, uses -en alone.
• i gda: in Kabyle, i is usually used just for the dative, but in Siwi it’s used for destinations in general; the g- in gda was originally the preposition “in”, but in Siwi it became part of the word for “here”.
• uṭnaxa “I am/was ill”: the -a suffix is a Siwi verbal form marking the perfect, frequently used in subordinate clauses to mean “while”. Kabyle doesn’t have such an ending, and would just use uḍnaɣ.

This contrasts with what we saw between Algerian and Egyptian Arabic, where very few of the textual differences were strictly grammatical. Of course, a longer text would have revealed more grammatical differences between Algerian and Egyptian, for example in the formation of comparatives – and would reveal many more between Kabyle and Siwi. This makes sense; for many centuries, Siwi has been much more isolated from Kabyle than Algerian Arabic has been from Egyptian Arabic, and the expansion of Berber happened earlier than that of Arabic, so they’ve had longer to develop separately.