Richardson’s Sokna phrases 1: Of jobs and geography

A while ago I came across a hitherto unpublished early source for Sokni: Richardson’s notes, which I plan to write up for publication at some point. James Richardson was an English traveller who made two trips to the Libyan desert in the mid-nineteenth century, gathering notes on half a dozen languages in the process which have ended up at the UK’s National Archives.  The phrase I’m going to look at today comes from a few pages which Richardson persuaded a man from Sokna named ʕAlī bən əl-Ḥājj ʕAbd al-Ṭawīl (sic!) to write for him in 1850, rhymingly titled كتاب المحاورة الإنسانية في اللغة العربية والسوكنية Kitāb al-Muħāwarah al-‘Insāniyyah fī al-Luɣah al-ʕArabiyyah wa-s-Sawkniyyah “Book of Human Dialogue in the Arabic and Sokni Language”, although this particular phrase is in a different hand, probably that of Richardson’s assistant Yusuf Moknee.  The phrase clearly describes Richardson himself:

English: An Englishman is going from Mourzuk to Ghat to spend some time with the Touaricks.
Sokni: مار سال انقليز ماشى سال تفزنى يغات اد ماتر اودان التواق
Arabic: رجل انقليز يحب يمش من مرزق الى غات باش يتفرج عند ناس التوارق

The Sokni is to be read as: mar sal əngliz maši sal tafəzni i ɣat ad [i]matər udan əttwa[rə]g. “A man from England is going from Murzuq/Fezzan to Ghat to watch Tuareg people.”

  • mar: man (recorded by Sarnelli), pl. imarriwən.  The word is rare in this meaning – the only Berber language to use a cognate that I know of is El-Fogaha – but the widespread forms for “beard” (tamart) and “hard work / power” (tamara) look potentially related.  If it is proto-Berber, one might be tempted to derive Latin Mauri and English “Moor” from this term, rather than from reconstructed Phoenician “westerner”, or Greek amauros “dark, faint”, as is usually suggested.
  • sal: from (attested in Sarnelli’s texts 1 and 5 – see previous post).  Forms with s for “from” are pan-Berber, but the -al is rather more surprising.  I’m inclined to compare it with ən, a fossilised morpheme roughly meaning “chez” that shows up in the Siwi complex prepositions sən “from chez”, gən “chez”, in “to chez” (cp. Tamajeq ehǎn “tent, house, family”, Zenaga ān “a dozen tents; family of”.)
  • əngliz: England.  Presumably from Italian inglese via Turkish and Arabic.
  • maši: going; this is a straightforward borrowing of colloquial Arabic māši, rather surprising in a Berber context
  • tafəzni: translated here as Murzuq, but Paradisi translates the same form in El-Fogaha as “Fezzan”. At that time Murzuq was the capital of Fezzan, so the usage makes sense.  Fezzan was known to the Romans as Phasania; the –i in this word suggests that the original was something like *fǎzǎni, but it might also be a fossilised survival of the locative ending -i found in Awjila and Ghadames.
  • i: to (attested in Sarnelli’s texts), but in most Berber languages it only marks the indirect object, whereas in Sokni and Siwi, as illustrated here, it marks the destination too.  The double function is probably calqued from Arabic, in which l– does the same two things.
  • ɣat: Ghat, a Tuareg town in southwestern Libya.
  • ad: pan-Berber irrealis (“aorist”) marker, prefixed to the verb (lost in Siwi)
  • i-matər: i-, which I’ve restored, is the 3rd person singular masculine subject agreement marker, “he”.  matər is “watch” rather than “visit”, as the Arabic translation as well as Sarnelli (“guardare”) confirm.
  • udan: people (attested in Sarnelli under “gente”, with no corresponding singular.)  Fairly widely attested in Berber. Siwi retains the corresponding singular, aggʷid, but uses a different plural.
  • əttwarəg: Tuareg.  It could also be əttwarəq, since the Arabic transcription doesn’t distinguish g from q. I’ve restored the missing r; the handwriting of this phrase is strikingly messy, unlike the rest of the manuscript.

In terms of vocabulary, the interest of this sentence for those already familiar with Sarnelli is limited to the geographical terms – while not especially obscure, tafəzni, ɣat, and əttwarəg are not otherwise recorded in Sokni.  Grammatically, the use of maši – an unanalysed Arabic participle – is striking; El-Fogaha uses at least one conjugated Arabic verb, as we will see, but I haven’t spotted any parallels in Sokna yet.

Lameen Souag


Sokna Text I: A Man and a Pomegranate Tree

This post is the first in a series of (re)translations of the published texts of the Berber language of the Sokna oasis in west-central Libya, in the same manner as the series of translations of Aujila Berber on this blog. I give here text, translation, and linguistic notes; the comments section, as always, is open for contributions.

The Berber language (once?) spoken in the oasis of Sokna (سوكنة) in west-central Libya is particularly interesting for our purposes. We don’t really know if it is still spoken, or by whom, because the most recent documentation of the Sokna language was in the 1920s. In fact, the single publication with any information about Sokni is a small collection of texts translated into Italian, accompanied by a rather unfulfilling Italian-Sokni glossary (and hardly any grammatical or linguistic commentary) by an Italian, Cap. Dott. Tommaso Sarnelli (1924). The texts that I present here will be re-translations based on Sarnelli’s transcriptions. Despite its merit in being the only substantial record of Sokni, his publication has numerous shortcomings. The glossary, for example, does not include any Arabic loans, and sometimes does not even include Sokni words from his own texts! There is no linguistic analysis or commentary, and no attempt to give even a basic sketch of the grammar. One also wonders how he arrived at his transcriptions, since he states that the stories were first written down for him in Arabic script (Sarnelli, 4). Presumably they were read aloud later, though this is not explicitly mentioned.

According to Sarnelli’s main informant, Shaykh Hassouna Ben Mohammed ad-Dakshi (“considered the best living speaker of the native language…”), in 1915 only about 40 or 50 people could understand the language, while only 4 or 5 could speak it fluently (ibid, 3). This does not bode well for our hopes of finding any native speakers.

Sarnelli seems to have been aware of the work of Rohlfs, a German explorer of the Sahara, who claimed to have drafted a dictionary and grammar of Sokni in the 1870s (unfortunately ‘lost’ during Rohlfs’ other travels in the desert). Rohlfs had characterized Sokni as “the most imperfect and the poorest of all Berber languages”; but Sarnelli strongly disagreed with this and found fault with a few of Rohlfs’ comments.

The Sokni language is relatively closely related to that of el-Fogaha (الفقهاء), a nearby oasis in central Libya, but they probably should not be grouped as one language as the scholarship up until now does. There is clearly some historical connection between the two; for example, inhabitants of both claim historic descent from the same geographic origin—Sāgya el-Ḥamra in southern Morocco. We hope to explore the nature of this relationship by working simultaneously on the published texts of Sokna and Fogaha Berber.

N.B. For some reason, Sarnelli sometimes puts asterisks (*) after long vowels–I have no idea what this is supposed to indicate, but I’ve left them in the transcription anyway. WLA means Western Libyan Arabic.


íǧǧěn ně mmār ilâ eššéžret n arěmmûn, arěmmūn-énnes mo̱qqâr

One man had a pomegranate tree, his pomegranate(-tree) (was) big.

  • íǧǧěn ně mmār ‘one man’. Genitive particle with epenthetic vowel; corresponding lengthening of initial m of mār?
  • mār ‘man’ pl. immariwen. This word for man is only otherwise attested  for Berber in El-Fogaha (see Richardson’s Sokna I, forthcoming).
  • i-lâ ‘have’
  • eššéžret ‘tree’ (< WLA eššǝžǝra ‘tree’), with -t of fem. construct state.
  • n genitive particle ‘of’
  • arěmmūn-énnes n. + 3sg.m.poss.pron.suff. ‘his pomegranate (tree)’. Rather than from Ar. rumān ‘id.’, the final long vowel hints at a possible Phoenician loan(?), in which Sem. ā  > Phoenician ū (via Canaanite ō), cf. Colin, Etymologies Maghrébines, Hespéris 1927: 88.
  • mo̱qqâr ‘big’ adj. deriv. of common berber verb *mɣr ‘to be big’. Compare Auj. amoqqǝrán ‘big, old’. Or, could be a stative verb.

iwî laḥmél n ameḳtâr iyi-ssẹ́lṭān, yessûṣēl-t.

He brought a donkey-load to the Sultan, and caused it to arrive(?).

  • iwî of âwid ‘he brought, carried’ (?)
  • laḥmél ‘load’ < Ar. l-(ǝ)ḥmǝ́l ‘load’, WLA CCvC nominal pattern with l-; takes gen. n ‘load of…’
  • ameḳtâr ‘ass, donkey’
  • iyi-ssẹ́lṭān dir.part. iyi + ssẹ́lṭān ‘sultan’ borrowed from Ar. with es-sulṭān ‘to the sultan’
  • yessûṣēl-t 3sg.m.impf. must be a causative derivation of Ar. yūṣel ‘to arrive’. The only way to explain the ē seems to be the tendency of imperfects of causatives to insert a long vowel before the final root consonant. (cf. Auj. impv. sg. š-îšef; impf. 1sg. š-išâfḫ ‘to sieve’), with a shift /ā/ to /ē/ in Sokni. Compare Auj. uṣǻlen /uṣǝ́lǝn/. The shift in the Sokni form seems to be some unusual imāla-like phenomenon, but I haven’t yet checked for comparison. The -t is the 3sg.m.dir.obj. clitic.

arěmmûn yèssěkkém-t laḥdîm, dě mār ět-tamektārt-énnes bẹ́dden d imî n tásqā.

The servant brought in the pomegranate (tree), and the man and his she-donkey halted under the entrance of the house.

  • yèssěkkém-t 3sg.m.impf. of causative ‘he brought it, made enter’ + -t 3sg.m.dir.obj. suffix. For the verb, cf. Fog. sékem ‘introdurre’ < ákem ‘entrare’.
  • laẖdîm n. < dial. Ar. l-exdīm ‘the worker, servant’. WLA noun shape.
  • ět-tamektārt-énnes ‘she-ass’ + 3sg.poss.pron. ‘and his she-ass’, note assim. of d/dě to fem. circumfix t-…-t *d-tamektārt > *t-tamektārt > ěttamektārt
  • bẹ́dden 3pl.impf. of bẹd ‘they stopped, halted’
  • imî ‘entrance (lit. mouth)’
  • n tásqā ‘of the house’

yuséd azzabṭi ěssẹ́n imarrîwen ẹ́nġả’’n iǧǧẹn ně mmār, suggarẹ́n-tën i-láḥbes.

An officer came with two men (who had) killed one man; they wanted them to (go to) prison.

  • yuséd of âsed ‘he came’. Compare Auj. yušâd; the shift to š is unconditional.
  • azzabṭi ‘officer’? from Ar. eð-ð̣ābiṭ (dial. ẓābiṭ?)
  • ěs-sẹ́n ‘two’, epenthetic vowel ě (to avoid initial CC)
  • imarrîwen ‘men’
  • ẹ́nġả’’n 3pl.m.aor of ‘to kill’. Not sure why Sarnelli inserts ’’.
  • iǧǧẹn ně mmār ‘one man’ (poss. constr. with gen. n)
  • suggarẹ́n-tën of issúgger ‘to want’+ -ten 3pl.m. dir.obj.suff. ‘they wanted them’
  • i-láḥbes from Ar. ḥabs ‘to prison’. note syllable structure of Ar. borrowing: CCvC as in WLA

šâwrěn fěll-âsen essẹ̌lṭân, yěnn-âsên: sagít-ten iyi-láḥběs.

They consulted about them, the sultan said to them: bring them to prison.

  • šâwrěn < Ar. šāwara III ‘to consult’ (cf. Ar. šūra ‘council’)
  • fěll-âsen ‘about them’ -ā
  • yěnn-âsên of mẹ́n ‘to say’ + 3pl.m.indir.obj.suff. ‘to them’
  • sagít-ten 2pl.impv. + 3pl.m. dir.obj.suff. ‘bring them’
  • iyi-láḥběs dir.part. iyi + Ar. noun ‘to the prison’. It is odd that the immediately preceding sentence has i- and here is iyi-.

sagín-těn, ítni dè bāb n tamektârt.

They led them – they and the owner of the she-ass.

  • sagín-těn + 3pl.m.dir.obj.suff. ‘they led them’
  • ítni 3pl.m.pron. ‘they’
  • ‘and’
  • bāb ‘father, owner’

ummuḥábsen sānā*.

They were imprisoned for a year.

  • ummuḥábsen (-mm- usually a middle/reciprocal prefix) of Ar. root ḥabbas ‘to imprison’
  • sānä ‘year’ < Ar. sana ‘year’

s igö*f ně lě‘âm yěnšẹ́d essẹ́lṭān sě bāb n tamektârt ẹ́mmik iṣâr dg-īs we bā tā usân-ās.

At the beginning of the year, the Sultan asked about the owner of the donkey, how (he) happened (to be) in it, and what they gave to him.

  • s ‘with, by, at’
  • igöf ‘head, beginning’ cf. iġọ̣́f
  • lě‘âm ‘one year’ < Ar. ‘ām ‘year’
  • yěnšẹ́d ‘he asks’ < dial. Ar. yenšed ‘to ask’
  • must mean ‘about’ here
  • ẹ́mmik ‘how’
  • iṣâr ‘happened’ < Ar. ṣār ‘to happen’ (?)
  • dg-īs ‘in’ 3sg.m.
  • bā tā ‘what’
  • usân-ās (misprint in Sarnelli for ušân?) of ‘they gave’ + ‘him’ 

ufân-t dē lěḥábs.

They found him in the prison.

  • ufân-t of af ‘to find’ + 3sg.m.dir.obj.suff. ‘him’
  • lěḥábs ‘prison

yěsseḥědō*r-t essẹ́lṭān we yěstâ‘dr-ās wu yěnnâ-s: nettâ-k.

The sultan caused him to appear (before him) and asked him for pardon and said to him: I had forgotten you.

  • yěsseḥědōr-t 3sg.m.impf.caus. (< Ar. ḥaḍara ‘to make appear’). Note lengthening of final vowel to long /ō/ (lengthening seems to be a phenomenon with the Sok. caus.), but why /ō/ and not /ē/?
  • yěstâ‘dr-ās vb. < Ar. yista‘ðar ‘to excuse oneself, to ask for pardon’ + 3sg.m.indir.obj.suff. ‘him’.
  • nettâ-k (royal plural?) + 2sg.m.indir.obj.suff. ‘you’ (cf. Foq. úttu ‘to forget’; Ghd. ǝttu ‘to forget’; Kb. əţţ ‘to forget’; Nef. étta ‘to forget’). This word is misprinted in Sarnelli as rettâk, but the handwritten Arabic looks like n; definitely not r.

we yěkkér dîd-ēs yěssekém-t iyi-lḵā*znet n elmā**l yěnn-âs: âwi ělḥāml-énnek sāl bā tā tsuggârě-t.

And he got up with him and brought him in to the treasure room (and) said to him: take your load of what you want.

  • yěkkér of ékker ‘he rose up’
  • dîd-ēs conj. + 3sg.m.obj.suff. ‘with him’
  • yěssekém-t ‘he brought it in’ (parsed above)
  • iyi-lḵāznet n elmāl dir.part. iyi + noun phrase (i)lḵāznet ‘storeroom’ < Ar. xazna + elmāl ‘wealth, treasure’ < Ar. māl ‘id.’
  • âwi 2sg.m.impv. ‘carry, take’ (?)
  • ělḥāml-énnek n. ‘load’ (note difference from earlier laḥmél ‘id.’) + 2sg.m.poss.suff. ‘your’
  • sāl ‘from’, note Pan-Berber s (see Richardson’s Sokni I, forthcoming).
  • tsuggâr-ět 2sg.impf. ‘you want’

iwí íg̃ǧẹn n assrẹ́m yennâ-s essẹ́lṭān: ẹ́mmik sā?

(He) took one axe. The sultan said to them: how (is it) so? (i.e. why?)

  • íǧǧẹn n assrẹ́m ‘one axe’ (poss. construction with gen. particle n)
  • ẹ́inmik sā ‘how so’

yènn-âs: ummuḥébsaẖ sě sebéb n aremmûn, suggâreḵ atténkōṭ, satrîži!

(He) said to him: (I was) imprisoned because of the pomegranate, I want to cut (it), please!

  • yènnâ-s ‘he said’ + 3sg.m.indir.obj.suff. ‘him’
  • ummuḥébsaẖ ‘(I was) imprisoned’ (-mm- passive) < Ar. ḥabbas ‘to imprison’
  • sě sebéb n ‘by reason of’ (constr. parallel to Ar. bi-sabab + iḍāfa), sebéb ‘reason, cause’ < Ar. sabab
  • suggâreḵ ‘I want’
  • atténkōṭ impf. ‘to cut’ cf. énkōṭ (< Ar. naqada with q > k?). The text in Arabic script doesn’t display a long ō, however.
  • satrîži 2sg. (cf. Nef. itréžža “pregare (persone)”) < dial. Ar. 2sg. itrāžī “wait” (?). Not really sure what this form is supposed to be.

ušān-ās laḥmēl-énnes uráǧ.

(But) he gave him his load (of) gold.

  • ušān-ās of aš ‘they gave’ + 3sg.m.indir.obj.suff. ‘him’
  • laḥmēl-énnes n. ‘load’ + 3sg.m.poss.suff. ‘his’. The second long vowel /ē/ is rather odd, and the Arabic text reads simply الحمل. Perhaps a resyllabification of earlier laḥmél (“mobile schwa”) [p.c. Lameen].
  • uráǧ n. ‘gold’

-A. Benkato

Awjila Text I

This series of articles are (re)translations, and revised transcriptions of the Aujila Texts presented in Paradisi (1960b), lexical data is drawn from Paradisi (1960a).

These texts were collected by Umberto Paradisi from October 1959 until May 1960. Four informants helped him with the collection of these texts. These same four informants were his main informants for the collection of his vocabulary. The texts vary in length and content, from historical stories, wise lessons, jokes and fairy tales.

This first text presented here is an example of a story whose main goal os to form a wise lesson.

The translation aims to give a word-per-word gloss of the text, and the transcription aims to be a phonemic transcription. Sometimes it will not be completely possible to retrieve the phonemic quality of a word, and several options will be given.

These texts are revised versions of those published on my personal blog.


Márra gan amə́dən aməqqərán u qəríb a-yəmmút

‘There once was an old (litt. large) man and he was about to die’

  • márra ‘once’, cf. Ar. marra(t-an) ‘id.’
  • gan‘there is, there are’, perhaps this is the participle of the root G* ‘to do’.
  • amə́dən[1] ‘man’, an analogous singular formation based on the common Berber plural noun míddən ‘people’.
  • aməqqərán ‘large’, definite adj.
  • u ‘and’ cf. Ar. wa- ‘id.’
  • qəríb ‘near’ adj. cf. Ar. qarīb ‘id.’
  • a-yəmmút[2] ‘to die’ fut.3sg.m.

w-iwín n išf illúm aməẓẓín-nəs w-ifk-ísin əlḥə́zmət n təġariwín w-in-isín:

‘and one day he gathered his sons and gave them a bundle of sticks and he said to them:’

  • w- ‘and’, variant in front of vowels.
  • iwín ‘one’, usually combines with genitive particle if uses attributively with a noun. As is the case here. There is also a lengthened variant of this word, iwínan, as of yet it is unclear what the conditioning is of this variant, they seem to appear in similar contexts.
  • n ‘of’ genitive particle.
  • išf  ‘day’, this word surprisingly lacks an accent. Perhaps because the common phrase iwín n išf has become an accentual unit.
  • illúm ‘to gather’ pf.3sg.m. cf. Ar. lamma ‘id.’ Arabic C2=C3 verbs are always found as C1C1uC2(C2) in Aujila.
  • w- ‘and’
  • ifk- ‘to give’ pf.3sg.m., usually found with the final vowel -a, this is lost when followed by an indirect object suffix.
  • -ísin 3pl.m. Indirect Object suffix. Note the place of the accent, which is the same place where it would have been, had the word ifká not been suffixed.
  • əlḥə́zmət ‘bundle’, cf. Ar. ḥuzma(t) ‘id.’
  • n ‘of’ genitive particle.
  • təġariwín ‘sticks’ plural of təġarit[3], the plural suffix attracts the accent to the final syllable.
  • w- ‘and’.
  • in- ‘to say’ pf.3sg.m. This word always takes a indirect object suffix, regardless if there is an actual object or not. The default object is -is.
  • -isín 3pl.m. Indirect Object suffix. Note the difference in accentuation to the earlier attestation of this suffix. The reason for this discrepancy is unclear.

“yaxát arẓát-tət.”

“take (and) break it”

  • yaxát ‘to take’, the imperative of the root , the singular imperative is yax, with devoicing of ġ in final position, this final devoicing sprad to the plural forms.
  • arẓát- ‘to break’.
  • -tət 3sg.f. Direct Object suffix.

U yuġí-tət kull iwinán s-ġar-sín iġə́lli a-irẓí-tət, yəzmər-ká,

‘And each one of them took it, and wanted to break it, (but) could not’

  • u ‘and’.
  • yuġí- ‘to take’ pf.3sg.m.
  • -tət 3sg.f. Direct Object suffix.
  • kull ‘all, each’ cf. Ar. kull ‘id.’
  • iwinán ‘one’, the accent in this text is different from the form listed in Paradisi (1960a) which has iwínan. The reason for this is unclear. Perhaps because it is used nominally, rather than adjectivally.
  • s-ġar- ‘from’ This preposition also occurs as simply s-, the exact conditioning when one or the other is used, is unclear. The form s-ġar-  (litt. ‘from at’) probably came to be to distinguish this preposition from the instrumental preposition s-.
  • -sín 3pl.m. post-prepositional pronoun.
  • iġə́lli ‘to want’ pf.3sg.m.
  • a-irẓí- ‘to break’ fut.3sg.m.
  • -tət 3sg.f. Direct Object suffix.
  • yəzmər ‘to be able’ pf.3sg.m.
  • -ká Negative particle. Always accented.

bəʕədén yuġí-tət ššárəf u yufukkí-tət u yəfká kull iwín təġarít.

‘afterwards, the old man took it, and untied it and gave each one a stick.’

  • bəʕədén ‘afterwards’  cf. Ar. baʕda an ‘after’.
  • yuġí- ‘to take’ pf.3sg.m.
  • -tət 3sg.f. Direct Object suffix.
  • ššárəf ‘old man’ cf. Ar. šārif ‘old (camel mare)’
  • u ‘and’.
  • yufukkí- ‘to untie’ cf. Ar. fakka ‘to separate, disjoin, disconnect, sever, sunder’
  • -tət 3sg.f. Direct Object suffix.
  • u ‘and’.
  • yəfká ‘to give’ pf.3sg.m.
  • kull ‘all, each’.
  • iwín ‘one’, here we find iwín, while in the previous sentence we found iwinán in almost the exact same context. This is an indication that these words can be freely used instead of one another.
  • təġarít ‘stick’.

Bəʕədén kull iwín yuġá təġarít u yərẓí-tə́t físa físa.

‘Afterwards, each one took a stick and broke it easily.’

  • bəʕədén ‘afterwards’.
  • kull ‘all, each’.
  • iwín ‘one’.
  • yuġá ‘to take’ pf.3sg.m.
  • təġarít ‘stick’.
  • u ‘and’.
  • yərẓí- ‘to break’ pf.3sg.m.
  • -tət 3sg.f. Direct Object suffix.
  • físa físa ‘easily’ cf. Dial. Ar. fissaʕ ‘quickly’ from fī (a)s-sāʕa(t) ‘in the hour’, curious loss of ʕ, this does not usually happen.

Bəʕədén in-isín: “akká, kəmmím am təġariwin-íyək,”

‘Afterwards he said to them: “Here, you are like these sticks,”

  • bəʕədén ‘afterwards’.
  • in- ‘to say’ pf.3sg.m.
  • -isín 3pl.m. Indirect Object suffix.
  • akká ‘here’, similar in usage to French voilà.
  • kəmmím 2pl.m. personal pronoun.
  • am ‘like’.
  • təġariwin ‘stick’, lost the accent because of the following suffix that attracts the accent.
  • -íyək demonstrative suffix ‘this, these’.

“ə́ndu təllummám iman-nəkím məʕá bəʕə́ḍkum a-ttəqqímam am təġariwin-íyək,”

“If you gather yourselves, you will stay together like these sticks”

  • ə́ndu[4] ‘if’.
  • təllummám ‘to gather’ pf.2pl.m.
  • iman- ‘oneself’, reflexive pronoun, followed by a pronominal suffix that agrees with the subject.
  • -nəkím 2pl.m. pronoun suffix.
  • məʕá bəʕə́ḍkum ‘with each other’, complete phrase is from Arabic: maʕa baʕadkum ‘id.’
  • a-ttəqqímam ‘to stay, reside, remain’ fut.2pl.m.
  • am ‘like’.
  • təġariwin ‘sticks’.
  • -íyək demonstrative suffix ‘this, these’.

“təllumám-ká iman-nəkím a-ttəqímam am iwatán n təġarít izmíra kull iwín a-yərẓi-tət bišwáš.”

“(If) you don’t gather yourselves, you will remain like one stick, everyone will be able to break it easily”

  • təllummám ‘to gather’ pf.2pl.m.
  • -ká Negative particle. Note that the accentuation did not remove the accent in the previous word.
  • iman- ‘oneself’, reflexive pronoun.
  • -nəkím 2pl.m. pronoun suffix.
  • a-ttəqqímam ‘to stay, reside, remain’ fut.2pl.m.
  • am ‘like’.
  • iwatán ‘one’ in the feminine form of iwínan.
  • n ‘of’ genitive particle.
  • təġarít ‘stick’.
  • izmíra ‘to be able’ res.3sg.m. [5]
  • kull ‘all, each’.
  • iwín ‘one’.
  • a-yərẓi- ‘to break’ fut.3sg.m. Interesting difference in vocalization compared to the earlier a-irẓí-. The realization of the 3sg.m. prefix as i- or yə- seems to in free variation.
  • -tət 3sg.f. Direct Object suffix.
  • bišwáš ‘easily’ cf. Egyptian Ar. bi-šweš ‘a little’.

U in-isín: “llummát iman-nəkím, a-issəʕəd-kím ṛəbbi.”

‘And he said to them: “gather yourselves, and god will help you.”

  • u ‘and’.
  • in- ‘to say’ pf.3sg.m.
  • -isín 3pl.m. Indirect Object suffix.
  • llummát ‘to gather’
  • iman- ‘oneself’, reflexive pronoun.
  • -nəkím 2pl.m. pronoun suffix.
  • a-issaʕəd- [6] ‘to help’ fut.3sg.m.m cf. Ar. sāʕada ‘to help’.
  • -kím 2pl.m. Direct Object suffix.
  • ṛəbbi ‘god’.

-M. van Putten


[1] Perhaps amédən.
[2] Unlike other Berber languages, the stem of the imperative and the aorist are distinct. Functionally, it is better designated as a future tense, similar to Lanfry’s futur in Ghadamès (Lanfry 1968: 333-334). Therefore, I have decided to stick to the terminology future rather than aorist.
[3] Perhaps taġarit.
[4] Perhaps úndu.
[5] The tense indicated as res. refers to resultative, this tense indicated an action that was finished in the past whose result affects the present. It is formed by placing an -a behind the verbal in the perfect, and after any object suffixes that may follow it. If there is a ə in the syllable before the -a clitic, it is raised to i. This form is Shared with Siwa Berber, albeit, with a slightly different function (Souag 2010: 389-392).
[6] Perhaps a-issəʕəd, indistinguishable in Paradisi’s transcription, and perhaps indistinguishable phonetically. Considering the Arabic origin with a long vowel in the first syllable of the root, Aujila probably has this too.


Lanfry, Jacques. 1968. Ghadamès. Etude Linguistique Et Ethnographique. I. Textes; Notes Philologiques Et Ethnographiques. Algérie: Fort-National.

Paradisi, Umberto. 1960a. “Il Berbero Di Augila. Materiale Lessicale.” Rivista Degli Studi Orientali 35: 157–177.

Paradisi, Umberto. 1960b. “Testi Berberi Di Augila (Cirenaica)” Annali. Nuova Serie 10: 43–91.

Souag, Lameen. 2010.  Grammatical Contact in the Sahara. Arabic, Berber, and Songhay in Tabelbala and Siwa. PhD Thesis.

Welcome to the Oriental Berber blog

This blog is aimed at studying and exploring the many details of  “oriental” Berber languages.

We use the word “oriental” in a particular sense here. Firstly, we indicate that our scope is the Berber languages of the eastern part of North Africa (mostly Libya, but also parts of Tunisia and Egypt). This is a strictly geographic grouping; we do not claim, nor is there necessarily, a straightforward genetic relationship between the different Berber languages of these countries. Secondly, it is also the case that these languages have on the whole received only marginal attention from researchers, and in some cases are scarcely documented. We thus want to highlight their differences from other, more well-studied varieties of Berber (such as those of western North Africa). Thirdly and poignantly, we also want to point out that, for much of the last century, the only scholarship on these languages was that of colonial Italian, British, or French researchers or travelers. To a certain extent those academic traditions shaped the final form of their studies, and hence the documentation that is today available. Fortunately, some of the languages here have become the subject of recent – and very good – scholarship; part of our goal is to continue this trend.

The position of all of these varieties is somewhat precarious, spoken as they are by geographically scattered minorities, and the smaller ones are seriously threatened: in particular, Awjili is endangered, and El-Fogaha and Sokni may already be extinct.  While recent political changes in Libya have opened up new opportunities for the use of Berber in the public sphere, particularly in the northwest, the long-term effects remain to be seen.

This blog will focus on several aspects on the study of the “oriental” Berber languages. A large part of the content of the blog, initially, will consist of the (re)translation of texts, provided with more extensive grammatical analysis, and sometimes, retranscription based on a more phonemic approach than the common transcription methods used by the Italian scholars who have dominated the field.

Other posts will include notes on linguistic history, discussions regarding interesting grammatical features, phonemic analysis, and so forth, and notes about these languages’ interactions with the surrounding (dominant) Arabic dialects.

Through this, we attempt to get, and give, as complete an overview of the knowledge there is to be gained from studying these fascinating languages with the documentation that currently exists. Through this blog, we also hope to bring them to the attention of a wider scholarship, and to encourage further research in the field.

Research on these Berber varieties is particularly timely, for the important reason that, especially in Libya, they are now coming to be used increasingly in the media. Prior to this year, there was essentially no media – written or audio – in Berber in Libya; it was suppressed by the regime. We also take an interest in this, and from time to time will post and comment on Berber media. There is no doubt that some interesting sociolinguistic and inter-dialectal phenomenon are going on.

We will finish off with a short overview of some of the languages to be discussed on this blog:

  • Awjila Berber (Awjili) is spoken in the oasis of Awjila, Libya. From a historical point of view, it is a fascinating language, as it is one of the few that retains Proto-Berber (as v). Other languages that have retained this consonant are the Tuareg languages (as h) and Ghadames (as β). Syntactically and morphologically Awjila is an interesting language, as it has lost much of the typically Berber features such as ‘state’, clitic fronting and has quite a different verbal system from other Berber languages.
  • The only Berber language that is spoken in Egypt is Siwa Berber (Siwi), in the oasis of Siwa in the western desert. Like Awjila, it has undergone intensive restructuring of the grammatical system, and fascinatingly, seems to share several of these grammatical features with Awjila.
  • Ghadames, an oasis in western  Libya on the border with Tunisia, is the home of Ghadames Berber (Ghadamsi), the other Libyan language that retains the Proto-Berber Ghadames is a fascinating language for historical linguists as it also shows some traces of the long lost Proto-Berber consonant . Patterns in the oriental Berber languages are the lack of ‘state’ marking, and a radically different verbal system than the more familiar Berber languages of western North-Africa. The verbal system of Ghadames may just be the most exotic reconfiguration of all the languages of this region.
  • North of Ghadames, still in western Libya, we find the Nefusa Berber (Nefusi) languages spoken around the Nefusa mountains, in the cities of Nalut, Jadu, Kabaw, and Yefren (to name a few). These languages have received quite substantial academic attention, from the perspective of oriental Berber. Nevertheless, further research, especially into its linguistic history, will be well worth it.
  • High up north on the coast of Libya, we find Zwara, where the Zwara Berber language is spoken. This language has received quite considerable attention due to the recent posthumous publication of Mitchell’s work, edited by Harry Stroomer and Stanley Oomen (Mitchell et al. 2009). The Zwara language is not generally considered to be part of the Eastern Berber group, and is rather associated with the Northern Berber languages, similar to Tunisian Berber. Nevertheless, this language could use more attention, and maybe in the future of this blog we will focus on it.
  • Sokna Berber (Sokni) was (or is still) spoken in the oasis of Sokna in west-central Libya. Our only record of Sokni comes from 1924, when only a few dozen people were reported to still speak the language. Though distinct from Fogaha Berber, there is some historical relationship between the two.
  • El-Fogaha Berber, traditionally considered to be the same language as Sokna Berber, seems to be lexically quite divergent. A more in-depth study of this language, will definitely give a clearer indication of the underlying relations between these two languages.