Phoenix's blog

Sîdi Ḥámed ĕz-Zarrûq

So, I spent the better part of my day breaking my head over the before last text that is available for Aujila. I've had serious trouble translating some parts of this text, and the text type is quite radically different from most of the stories we've heard so far. This seems to be a religious legend aboud Sîdi Ḥámed ĕz-Zarrûq whom I assume is the same as Ahmad Zarruq.

I'll present my translation, apologies in advance that it won't be as satisfyingly easy to understand as previous texts, and I hope readers can help me out clear up some of the weirder pards.

Gan iwínan n-amedən ašál n-awílən

'There was a Aujilan man'

• Gan 'There is, was'
• iwínan n-amedən 'one man'
• ašál n-awílən 'village of Aujila', apparently phrase functions as an adjective in some way, I'd expect a preposition that means 'from' or maybe just n- 'off'.

márra yuġáya ksúm u yušád yəxṭíma af ammúd n Sídi Hamid az-Zarrúq

'One day he bought meat and came and visited the mosque of Sídi Ḥáməd əz-Zarrúq'

• márra 'once'
• yuġáya 'he bought'
• ksúm 'meat'
• yušád 'he came'
• yəxṭíma 'he visited'
• af ammúd n Sídi Hamid az-Zarrúq 'at the Mosque of S.Ḥ.Z.' It's interesting that the word ammúd means 'mosque', this is a, not very broadly attested, word for 'prayer' in Berber.

yufí-tən ʕə́mma mudán lʕə́ṣər

'He found them praying the afternoon prayer'

• yufí-tən 'He found them'
• ʕə́mma seems to be a particle to indicate progressive aspect < Ar. ʕamma 'to be or become general, universal, common' [Lameen:]< Arabic ʕammāl (also used in Siwi.)
• mudán 'they prayed'
• lʕə́ṣər 'the afternoon prayer' < ʕáṣr 'afternoon, afternoon prayer'

yúna ammúd-i w-ídd-əs ksúm

'He entered the mosque and (t0ok) with him the meat'

• yúna 'he entered'
• w-ídd-əs 'and with him'

iggí-t dít-a, w-əlimám sídi Ḥáməd əz-Zarrúq u baʕád ikkə́mməl yəʕádd irrə́wəḥ u šummán ksúm wa yərfíʕa

'He put it in front of him, the Imam was S.Ḥ.Z. and after he finished and he went to return home and he cooked the meat that he had brought'

• iggí-t 'he put it'
• dít-a 'in front of him' This takes a different suffix than the preposition idd-, I assume -a is related to Arabic somehow? [Lameen:] could reflect 3msh -hu, which goes to -a in some dialects. Also used with dəffər here though. [Phoenix:] That's what I thought. There seems to be completely free variation between using the Berber and Arabic suffixes in these cases. Very odd.
• w-alimám sídi Ḥáməd əz-Zarrúq'and the imam was S.Ḥ.Z.', I'd expect the copula d- here. Nowhere to be found.
• u baʕád 'and after'
• ikkə́mməl 'he finished'
• yəʕádd 'he went'
• irrə́wəḥ 'he went home'
• šummán 'to cook' 3pl.m. Is it 3pl.m. because the family is that is cooking along is automatically incorporated into the sentence without further introduction? [Lameen:] yeah, "they" is presumably the people of the house. I'm surprised it's not feminine plural, actually.
• ksúm wa yərfíʕa 'meat which he brought'

ʕala-má fəkkan-ís afíw yəlġăm a-yə́mm.

'As much as he was giving flames to it, it refused to become cooked'

• ʕala-má 'as much as', clearly an underlying Arabic form, but I couldn't find this exact phrase in Wehr.
• fəkkan-ís 3pl.m.impf. 'to give', this must be a 3pl.m. So šummán probably is as well.
• afíw 'fire'
• yəlġəm 'to refuse' 3sg.m.pf.
• a-yə́mm 'to be cooked' 3sg.m.aor.

Baʕadén yəʕádd yə́nšəd ərrəfəqá-nnəs wi ižinána nəttín-id-sín ksúm.

'Then he went to ask his companions whom he had divided the meat with'

• Baʕadén 'then'
• yə́nšəd 'he asked'
• ərrəfəqá-nnəs 'his companions' < Ar. rafīq pl. rufaqāʔ 'companion'
• wi ižinána pl. pronoun + 3sg.m.pf. 'to divide'
• nəttín-id-sín 'he-with-them'

Nan-ís: nəkkəní nəšummi-tíya u nči-tíya

'They said to him: we cooked it and we ate it'

• Nan-ís 'they said'
• nəkkəní 'we'
• nəšummi-tíya 'we cooked it'
• nči-tíya 'we ate it'

In-isín: nək ʕala-má fəkká-s afíw yəlġə́m a-yə́mm

'he said to them: I was giving it as much fire (as possible), it refused to be cooked'

• In-isín 'he said to them'
• fəkká-s 'to give' 1sg.impf. + indirect object 3sg.

Nan-ís: ddíwa dgíta?

'They said to him: What did you do?'

• ddíwa 'what'
• dgíta 'to do, put' 2sg.pf.

 In-isín: baʕád ušíġd uníx ammúd-i, ufíx-tən ʕə́mma mudán.

'He said to them: After I came and entered the mosque, I found them praying'

• Repetition of an earlier phrase, now in the 1sg.pf.

gíx dít-i ksúm u qqăimíx mudíx, wén-ma kəmmə́lx ʕaddíx; wa d-əlá ṣərána.

'I put the meat in front of me and I started praying, as soon as I had finished I went; That is what happened'

• gíx dít-i ksúm 'I put the meat in front of me'
• u qqăimíx mudíx 'and I started praying', notice the odd diphthong-like notation in qqăimíx.
• wén-ma 'as soon as' < Dialectal Arabic wēn 'where' < Ar. wa-ʔayna 'and where' + mā? Could not find this construction in Wehr.
• wa d-əlá ṣərána. 'that is what happened', we've seen this phrase in the previous text.

Nan-ís ərrəfəqá-nnəs ʕádd an-ís-t i-sídi Ḥamə́d ăz-Zarrúq.

'His companions said to him: go and tell it to S.Ḥ.Z.'

• No further notes needed I think.

Yəʕádd in-ís-t am-alá nan-ísa

'He went and he says it like they said to him'

• am-alá 'like'

wén-ma in-ís-t i-sídi Ḥáməd ăz-Zarrúq, in-ís:

'As soon as he said it to S.Ḥ.Z., he said to him:'

nək lukán wa ammudán də́ffər-i a-iččí-t afíw, maʕádč a-mmudə́x s-ḥíddan

'If he who would pray behind me would be eaten by fire (that is: burn in hell), I would no longer lead them in prayer.'

• This is a pretty essential part of the text and I can't make sense of it at all.
• Paradisi translates: Se colui che prega dietro di me andasse all'inferno (lett: lo mangiasse il fuoco), non pregherei più per alcuno. Which Google Translate lovingly translated as: 'If the one who prays behind me to go to hell (litt: the fire eats him), do not beg for any more.' Which is not very helpful.
• lukán 'if'
• wa ammudán də́ffər-i 'who prays behind me' with a aor.ptc. ?
• a-iččí-t afíw 'the fire will eat him'
• maʕádč 'no longer' [Lameen:] maʕádč is Arabic mā-ʕād-š. In my dialect this would be "not yet", but I think in Libya this would be "no longer"; [kato:] Interesting that your dialect has maʕádč for 'not yet', Lameen. Here, it'd be "no more, no longer", and is placed before the vb as in ELA.
• a-mmudə́x 1sg.aor. [Lameen (paraphrased):] mmud + s-  = 'lead in prayer' calque of Ar. ṣallī bi- 'to lead in prayer'.
• s-ḥíddan 'to anyone' [Lameen:] ḥiddan is of course Arabic ʔaḥad-an "anyone" - with tanwīn! My own dialect has šayən for "nothing", so the survival of tanwīn in this context is not unprecedented, but it is interesting.

In-ís: ksum-áya ġár-ək ʕádd kəffə́n-t, u ʕádd mtí-t žəbbánət

'He said to him: That meat that you have, go wrap it up, and go bury it in the cemetary.'

• ksum-áya 'that meat'
• ġár-ək 'you have' ġár on it's own seems to be almost exclusively used for possesive constructions.
• ʕádd really seems to be used a lot as a sort of Auxiliary verb here, maybe to strengthen the imperative?
• kəffə́n-t 'wrap it up' < Ar. kaffana 'to wrap s.th. up'
• mtí-t 'bury it'
• žəbbánət I'd except a preposition here af žəbbánət or žəbbánət-i but instead the sentence says: "go bury it the cemetary", this is very odd to me.

Yəʕádd yəmtí-t u yəqqím ixəbbár míddən

'He went to bury it and he started telling the people'

• yəmtí-t 'he buried it'
• yəqqím 'he started to...'
• ixabbár 'he said' impf.3sg.m. < Ar. xabbara 'id.'
• míddən 'people'

wén-ma slán-t míddən n ašál kúll ušánd ġallíyən ammudán kull də́ffər-a w-ammúd məššə́k iġəlli-ká a-yúġ míddən n ašál.

'As soon as they heard this, all the people of the village came and wanted all to pray behind him, and the small mosque could not fit (litt: did not want to take) the people of the village.'

• slán-t 'they heard it'
• míddən n ašál kúll 'all the people of the village'
• ušánd 'they came'
• ġallíyən 'they wanted'
• ammudán 'to pray' 3pl.m.aor.
• kull this second kull surprises me a bit because of it's position
• də́ffər-a 'behind him'
• məššə́k 'small' m.sg.
• iġəlli-ká a-yúġ 'it did not want to take' interesting way of expressing 'They didn't fit'
• This sentence give us the impression that whatever S.Ḥ.Z said about sitting behind him is a good thing, since the whole village wants to do it now.

yəxammə́m əlfəkr-ə́nnəs sídi Ḥamə́d, illúff dbăš-ə́nnəs w-in-isín ġallíx a-ʕddaáx (sic) fḥáli.

'Sidi Ḥamə́d thought to himself (litt: he thought his thinking), packed up his stuff and he said to them: I want to go by myself.

•  yəxammə́m 'to think' 3sg.m.pf. I know this word, it's also found in Riffian and MA Berber. The x would suggest an Arabic origin, but the word itself doesn't look particuarly Arabic due to a double m. [Lameen:] from Arabic xammana 'to guess, conjecture, surmise' [Phoenix:] That's why it didn't look Arabic! Assimilation of nasals accross syllables. Cool.
• əlfəkr-ə́nnəs 'his thinking' < Ar. fikr 'thinking, cognitation, reflection'
• illúff 'to pack up' 3sg.m.pf. < Ar. laffa 'to wrap up'
• dbăš-ə́nnəs 'his stuff' < Ar. dabaš 'stuff'. This is interesting, this word is different from dbúš 'clothes'. This is odd, we had previously identified the latter word to be from dabaš. But this seems to be a different lexeme then. the shape of dbuš suggestes that it comes from an Arabic word da/ubūš, this derivation doesn't exist in classical Arabic at least. Is it the plural? Classical Arabic has ʔadbāš, but maybe one of the two Arabic dialects that influence Aujila Berber had dubūš as the plural instead?
• ġallíx 'I want'
• aʕddaáx (sic) This must be a typo for a-ʕaddə́x. 1sg.aor. 'to go'
• fḥáli 'by myself', I don't know this expression, but I assume it's from fī ḥālī 'in my condition/state'.

Nan-ís ləhál n ašál: a-nnaʕádd kull-ídd-ək

'The people of the village said to him: We will all go with you'

• ləhál n ašál 'people of the village' <  Ar. ʔahl 'relatives, folks, family; kin; people, members, followers etc.'
• a-nnaʕádd 1pl.aor. 'to go'
• kull-ídd-ək 'all-with-you' Maybe this sentence is better translated as 'we, and you, will go together'. That'show pronoun-idd-pronoun is usually translated (look back at previous texts with nəkk-ídd-ək for example), but it doesn't sound that good in English.

U škíyən ídd-əs ir a-hlə́bən ašál u baʕadén in-isín:

'And they left with him until they would leave the village and he said to them'

• U škíyən 'and they left'
• ídd-əs 'with him'
• ir 'until'
• a-hlə́bən 3pl.m.aor. of a root hlb, I can't find this, I only find Ar. haliba 'to be hairy', which is definitely not what we're looking for. Paradisi translates oltrepassrono.

ṣbərát, a-nna-kím nək d-awíl tláta marrát ílă yom əlqiyáma;

'Wait, I will tell you I will be an Aujilan three times until the Day of Judgment;'

• ṣbərát imperative 2pl.m. 'to wait' <Ar. ṣabara 'to bind, to be patient'
• a-nna-kím 'I will tell you (pl.)
• nək d-awíl 'I am an Aujilan'
• tláta marrát 'three times'
• ílă 'until', notice not Berber ir is used by Arabic ʔilā
• yom əlqiyáma < Ar. Yawm al-Qiyāma(t)

wása anzurrán s-ġar-kím a-ʕəyyə́ṭən-dîk təláta marrát; a-ušázd w-a-ffukkə́x-t

'One of you who will be suffering, will call me three times; I will come and alleviate it.'

• wása longer form of wa
• a-nzurrán ptc.aor. 'to suffer', I can't find this word anywhere. But should mean 'to suffer'
• s-ġar-kím 'from amongst you (pl.)
• a-ʕəyyə́ṭən-dîk 'they will call me'
• təláta marrát 'three times'
• a-ušá-zd 'and 'I will come for him'. A lot of assimilations have made this word incredibly unclear. It is like this: a- aorist particle uš 'to come' -a- 1sg. ending losing it's -x because the following 3sg.m. indirect object suffix -ís which in turn loses its i due to the a in the previous ending, and assimilates the s in voicing to z for the following -d direction particle.
• w-a-ffukkə́x-t 'and I will alleviate him' < Ar. fakka 'to seperate, disjoin, disconnect'

Undú ušiġd-ká, a-nmḥásəb nək ídd-əs yom əlqiyáma

'And if I don't you, I and him will be held responsible (on) the Day of Judgment'

• ušiġd-ká 'I don't come' 1pl.m.pf.-negative
• a-nmḥásəb 'to be held responsible' 1pl.aor. This would mean it's from Ar. ḥāsiba 'to hold responsible' which probably took the meaning 'to be responsible for s.th.' in Aujila. [Lameen:] The m- must be a passive formation here.

Nək wa d-əlá slíxa s-ar míddən lə́wwəl kəddímən ḥəkkán-dík af sídi Ḥáməd əz-Zarrúq.

'This is what I heard once from old people, who used to tell me about S.Ḥ.Z.'

• wa d-əlá slíxa 'what I heard'
• s-ar míddən 'from people'
• lə́wwəl 'first' < al-ʔawwal 'first', I don't really understand the meaning of this word in this context. [Lameen:] in Siwi this means "once, formerly" as well as "first", which would work here.
• kəddímən 'old' m.pl adj. This is a very odd form, this is clearly from Ar. qadīm 'old'. Why does it have k for q?
• ḥəkkán-dík 'they have told me' < ḥakā 'to tell'. Surprising that it has a kk, there's no reason why that geminate arises here. [Lameen:] the geminate suggests rather "they used to tell me" (imperfective.)" [Phoenix:] Imperfective can have a habitual meaning.

[Lameen:] The point of the story is the idea that God has granted Sidi Hamed the privilege that anyone whom he's led in prayer will be spared the Fire (of Hell) - even a piece of meat he's "led in prayer" can't be burned. On top of that, before leaving he agrees to intercede for any Awjili in trouble. Such ideas of intercession would now be seen as unacceptable by most Muslims (the last sentence suggests they were already starting to be seen that way), but they're fairly representative of the attitude people used to have towards saints in many regions.

999 Pounds

Žḥá yətəmə́nna s-ġar rə́bbi a-yəfk-ís əlf zní

'Žḥá wished of god to give him 1.000 Libyan pounds'

• yətəmə́nna 'to wish' 3sg.m.pf. < Ar. tamannā 'to desire, wish'
• s-ġar rə́bbi 'from god'
• a-yəfk-ís 'to give' aor.3sg.m. + 3sg. indirect object
• əlf 'thousand' < Ar. ʔalf 'id. '
• zní 'pound' < Ar. žunaih 'id.' but obviously the dialectal pronunciation, not the MSA pronunciation. MSA changed it to look/sound more classical. [Lameen:] From the British coin Guinea

u yəqqím iṭə́lləb u itənn-ís: Ya rắbbi undú yənqíṣa iwínan a-uġə́x-tenet-ká!

'and he started requesting/begging and saying: 'O god! If even one is lacking, I will not take them!'

• yəqqím inchoative 3sg.m.pf.
• iṭə́lləb 3sg.m.impf. 'to request' <Ar. ṭalaba 'to look, search; to request etc.'
• Ya rắbbi  'o god!' with an Arabic vocative particle ya instead of a/ay that we've seen so far.
• undú 'if'
• yənqíṣa 'to be lacking' 3sg.m.pf. < Ar. naqaṣa 'to decrease, become less, diminish, be diminished, be reduced'
• iwínan 'one'
• a-uġə́x-tenet-ká aor.1sg. 'to take' + 3pl.f. direct object suffix + negative.

Baʕadén yəslí-t əlhúdi aringí-nnəs

'Then his Jewish neighbour heard him'

• Baʕadén
• yəslí-t 'to hear' 3sg.mp
• əlhúdi 'Jew'. Clearly from Arabic, but my MSA dictionary wasn't of much help. I could find al-hūd 'Jewry' and yahūdiy 'jewish'. This word seems to be a mix between these two.
• aringí-nnəs 'his neighbour'. This final í is remarkable. In Paradisi's wordlist we find arîng,  anîrg pl. ringîn, nirgîn 'neighbour' (notice the interesting metathesized forms). The í is written with a short accented í, which you never really see, so this seems to actually be the schwa of the suffix, but I'm not sure what motivated it to be colored to an i-like sound for Paradisi to transcribe it as í.

Baʕadén in-ís lhúdi y-imán-nəs wăllahi qáma a-ḥdəfá-s tə́səʕa míyət id təsəʕa u təsʕín žní w-a-ggulə́x-t a-yuġí-tənət naġ wə́la

'Then the Jew said to himself: I swear, now, I will throw him 999 Libyan pounds and I'll see (if) he takes them or not.'

• in-ís 'he said'
• y-imán-nəs 'to himself'
• wăllahi 'by god'  means 'I swear to god'
• qáma 'now'
• a-ḥdəfá-s 'I will throw to him' 1sg.aor. This should be from Arabic, but I can't find the word in Wehr. Apparently Siwi has ʕadef with the same meaning.
• tə́səʕa míyət id təsəʕa u təsʕín 999 litt: nine hundred with nine and ninety.
• žní 'pound' Notice that this time the word has the expected ž
• w-a-ggulə́x-t 'and I will see him' 1sg.aor.
• a-yuġí-tənət  'he takes them' 3sg.m.aor.
• naġ 'or'
• wə́la < Ar. wa-lā 'nor, neither', a bit of an odd construction 'or nor' for 'or not'. Dialectal use of wa-lā? [Lameen:] cp. Siwi ula "no" (and discussion of its origins in my thesis...)

w-igi-tə́nət arəqúṭ-i u yəḥdəf-tə́nət dit nə Žḥá s-ar ərróšən.

'And he put them in rags and he threw them in front of Zha from the window'

• w-igi-tə́nət 'and he did/put them'
• arəqúṭ-i 'in rags'. For some reason this word was ringing a bell. Not sure why.
• u yəḥdəf-tə́nət 'and he threw them'
• dit nə Žḥá 'in front of Žḥá'
• s-ar ərróšən 'from the window' <Ar. rawšan 'skylight, scuttle; peep-window, peep-hole, spy-hole'

 U yətuár arăqúṭ u yəḥsəb-tə́nət yufi-tə́nət ənqəṣníta iwínan u baʕadén in-ís:

'And he opened the rags and he counted them and found them to be missing one (pound) and then he said:'

• yətuár 'to open, to be opened' 3sg.m.pf. Actually a passive of yar/ar 'to open', but can have an active meaning as well.
• yəḥsəb-tə́nət 'he counted them'
• yufi-tə́nət 'found them'
• ənqəṣníta 'to be lacking' 3pl.f.pf.

wa yəfki-díka lʕə́lla a-yəfki-dík iwínan

'He who gave me many (pounds), may he give me one (more pound)'

• wa yəfki-díka 'he who gave me'
• lʕə́lla 'much, many'
• a-yəfki-dík 'may he give me', Aorist used to express a wish here.

baʕadén igi-tə́nət ṣṣundúq-i u nəttín yəfríḥa

'Then he put them in a box and he was happy'

• igi-tə́nət 'he put them'
• ṣṣundúq-i 'in a box' < Ar. ṣundūq ‘crate, box; chest’. Here a short u is retained from Arabic. It seems that, the q-dialect Aujila loaned from retained the opposition of u and i, and Aujila simply took it over. This makes the short vowel system in Aujila oddly versatile considering that only one short vowel is present in the Berber lexicon.
• u nəttín 'and he'
• yəfríḥa 'he was happy, satisfied'

baʕadén yəškí əlhúdi w-in-ís yə-Žḥá: aġíd əlbarat-ənnúk.

'Then the Jew cam out and said to Žḥá: bring (me) my money!'

• yəškí 'to come out'
• aġíd Imperative 'bring'

In-ís Žḥá: rə́bbi yəfki-díka albəṣírət u ku tġallít a-ttuġát-tət s-ġár-i?

'Žḥá said: god gave it to me willingly and you want to take it away from me?'

• albəṣírət < Ar. baṣīra(t) 'insight, penetration, discernemnent, understanding' but the phrase ʕan baṣīra(t) means 'deliberately, willingly', which makes more sense to be the meaning here. Does dialect arabic drop the ʕan?
• u ku 'and you'
• tġallít 'to want' 2sg.pf.
• a-ttuġát-tət 'to take' 2sg.aor. + 3sg.f. DO
• s-ġár-i 'from me'

In-ís əlhúdi: nək wa əḥdəx-títa i-kú dax a-gguláx-ku a-ttuġa(t)-tə́net ənqəṣníta naġ wə́la

'The Jew said: I was the one who threw it to you in order to see (if) you would take them with (one pound) missing or not.'

• nək wa əḥdəx-títa 'I'm the one who threw it'
• i-kú 'to you' [Lameen:] All Berber languages, as far as I know, use the full forms of pronouns after i, not the prepositional ones.
• dax 'so, in order to'
• a-gguláx-ku 'to see you'
• a-ttuġa(t)-tə́net 'you would take them'

yəlġóm a-yəfki-tə́net y-əlhúdi

'he refused to give them (back) to the Jew'

• yəlġóm 'to refuse' 3sg.m.pf.
• a-yəfki-tə́net 'to give them' 3sg.m.aor.

Baʕadén in-ís əlhúdi: nək a-nṭárəx-ku-ká ir a-nnaʕádd nək-ídd-ək ar əlqə́ḍi.

'Then the Jew said: I will not let you go until me and you have gone to the Qadi'

• a-nṭárəx-ku-ká 'I will not let you go' 1sg.aor. with 2sg. object. + negative. An Arabic word must lie at it's origin but naṭara means 'to watch, to guard', which seems like the exact opposite of the word that I'd expect.
• ir 'until'
• a-nnaʕádd 'to go' 1pl.aor.
• nək-ídd-ək I with-you
• ar əlqə́ḍi 'to the qadi'

In-ís Žḥá: nək ʕəyyánəx u zmərx-ká a-ʕaddə́x af šərin-núk w-ərwə́ḥa s-ġar šmuṭ u ġar-i-ká təlába

'Žḥá said: I am sick and I am unable to go by foot(?) and I am afraid of the cold and I don't have a Barracan'

• ʕəyyánəx 'I am sick'. Arabic must underly this, I could not find the word. [Lameen:] < Ar. ʕayyān "exhausted, tired, ill".
• zmərx-ká 'I am not able' The form looks like a Perfect, but that is not the tense I would expect, I'd rather expect an Imperfect.
• a-ʕaddə́x 'to go' 1sg.aor.
• af šərin-núk 'on my legs'?? translation by Paradisi as 'a piedi', I can't find the word that this represents. Is it Arabic?
• w-ərwə́ḥa 'and I am afraid' I think this, again, is a Perfect, while I'd expect an Imperfect here. Am I thinking of Perfect too much as a past tense? Or am I not regarding Imperfect enough as a way to express habitual action?
• s-ġar šmuṭ  'of the cold'
• u ġar-i-ká təlába 'and I don't have a barracan' Litt: and to me is not a barracan'

In-ís əlhúdi: afká-k aẓíṭ id təlába ġar ʕádd ídd-i ar əlqə́ḍi

'The Jew said: I will give you a donkey and a Barracan but you must go with me to the Qadi'

• afká-k 'I will give you' 1sg.aor.+2sg.DO
• aẓíṭ id təlába 'donkey and barracan'
• ġar 'but'
• ʕádd 'go' imperative singular
• ídd-i 'with me'

Baʕád uṣə́llən ar əlqə́ḍi išərw-ís ləhúdi s-alá ṣarána.

'Later they arrive at the Qadi, the Jew tells him about what happened'

• Baʕád 'later'
• uṣə́llən 'to arrive' 3pl.pf. I think this is from Arabic waṣala 'to join, unite', possibly form II.
• išərw-ís 'to tell' 3sg.m.pf. + 3sg.IO.
• s-alá s- 'from' but also apparently 'about' + alá 'that wich', a sort of relative pronoun.
• ṣarána 'to happen' 3pl.m. < Ar. ṣāra 'to happen'

In-ís əlqə́ḍi i-Žḥá: ṣəḥíḥ žlan-íya?

'The Qadi says to Žḥá: Trutful is this story? (The force, you must use!)'

• ṣəḥíḥ 'truthful, healthy'
• žlan-íya 'this story', I'm a bit surprised by the Yoda-like flipping of the word order from what I would expect. Is ṣaḥíḥ being used as a stative verb? That would not require endings, but I'm a bit surprised that Arabic adjectives can be employed like that.

Baʕadén in-ís: lhúdi iqə́lləz fəll-íwi u nək rwíḥḥa a-yə́qləz w-a-in-ís ḥə́tta tlaba-nnúk ta tənní fəll-íwi t-tá-nnəs nəttín id aẓiṭ-ənnúk a-in-ís d-wá-nnəs nəttín.

'Then he said: The Jew is lying about me! And I am afraid that he will lie and say that even my Barracan that I'm wearing is his and that he will say that my donkey is his!'

• iqə́lləz 'to lie' 3sg.m.impf.
• fəll-íwi 'on me' to mean 'about me'
• rwíḥḥa 'to be afraid' 1sg.pf. (once again a, unexpected to me, Perfect. It seems like I really can't get my head around this use of the Perfect, isn't he currently in the state of being afraid, thus not having finished it?)
• a-yə́qləz 'to lie' 3sg.m.aor.
• w-a-in-ís 'and he will say'
• ḥə́tta 'even' < Ar. ?
• ta tənní fəll-íwi 'that is on me' Pronoun + 'to be, exist' 3sg.f.pf. + on-me
• t-tá-nnəs nəttín assimilation of d- 'copula' + tá-nnəs 'that of his', nəttín 'he' apparently placed after it to emphasize the possesive relationship. 'He will even says that this Baraccan is his'
• a-in-ís d-wá-nnəs nəttín. 'he will say is his!' once again with extra pronoun 'he'.

Baʕadén iʕáyəṭ əlhúdi w-in-ís y-əlqə́ḍi: wălláhi t-ti-nnúk

'Then the Jew scremead and said to the Qadi: I swear they are mine!'

• iʕáyəṭ 'to scream' 3sg.m.pf., now that I have confirmed what the perfect looks like, I can be sure that the form in the previous text was indeed an imperfect.
• t-ti-nnúk copula d- + fem plural pronoun ti-nnúk 'those of mine'

In-ís əlqə́ḍi: ku təqəllət ṣəḥíḥ ʕádd ʕádd rə́waḥ

'The Qadi said: You really are lying! Go! Go! Go home!'

• təqə́llət 'you are lying' 2sg.impf.
• ṣəḥíḥ 'truthfull, healthy' used as an adverb here. I like how nicely it translates to English.
• rə́waḥ 'go home' imperative singular.

U yəʕádd Žḥá yəfríḥa yəḥəṣṣíla ədbúš id aẓíṭ id əlbarát

'And Žḥá went (away) happy, having obtained clothes, a donkey and money'

• yəḥəṣṣíla 'to obtain' 3sg.m.pf. < Ar. ḥaṣṣala 'id.'
• ədbúš 'clothes'
• əlbarát 'money'

Waiting for the Police

Today, a rather morbid story.

Gan iwínan yərfíʕa af təgilí-nnəs əlḥə́ml ẓẓák s-ġár ləmluxíyət

'There was a (man) carrying a heavy load of Mulukhiyah on his head.'

• Gan 'There is, there was'
• iwínan 'one'
• yərfíʕa 3sg.m.pf. 'to carry'
• af təgilí-nnəs 'on his head'
• əlḥə́ml 'load' < Ar. ḥiml/ḥaml 'burden, load'
• ẓẓák 'heavy' m.sg.
• s-ġár 'from' <- Odd use of this ablative particle (which I still can't distinguish from s- 'from'), as here it shows a definition of material. This is similar to the Dutch word van 'of, from' 'een lading van Mulukhiyah' litt: 'a load from Mulukhiyah'
• ləmluxíyət 'mulukhiyah'

W-ižigáž žənqín-i n ašál u nəttín iʕəyáṭ w-itənn-ís: ləmluxíyət báhəyət!

'And he was walking in the streets of the village and he yelled and was saying: Good Mulukhia!'

• W-ižigáž 3sg.m.impf. 'to walk' looks like a causative of a root gž, not sure what that root would mean. Not attested in Aujila at any rate.
• źənqín-i 'in the streets' < Ar. zinaq 'narrow street' Wehr says that in Moroccan Arabic this is the normal word for 'street'. This word has been Berberised, the plural is also spelled dzanqín and the singular is tzenắqt.
• n ašál 'of the village' note how an interpretation 'country' would not make much sense here.
• u nəttín 'and he'
• iʕəyáṭ <Ar. ʕāṭa ‘to yell, scream’. In my notes I have written down that this is a 3sg.m. perfect. But I think an interpretation as imperfect would make more sense here. Sadly without any other forms, it is difficult to determine what the perfect would have looked like. This form also arose in the story about the donkey at the auction, where it also preceeded w-itənn-ís (an imperfect). 
• w-itənn-ís 3sg.m.impf. 'to say'. This is a pretty clear example of proof that the -ís is an obligatory indirect object, because in this case, there definitely isn't a 3rd person indirect object being adressed to.
• báhəyət 'good' < dial. Ar. bāhi, with feminine conjugation.

W-ižigáž ggút išfíwan u ižíž lúda

'And he was walking for many days, and he sold nothing'

• ggút 'many'
• išfíwan 'days'
• ižíž 'to sell' 3sg.m.pf.
• lúda 'nothing' Does this come from Arabic?

u tafút təḥmáya fəll-ís u taqqím təttá afiš-ə́nnəs id əlgəfán-nəs u təqqím am əžžaḥím fəll-ís

And the sun shone upon him and it was starting to eat his face and back of the head and it continued (to shine) like hell was upon him.

• tafút 'sun', this word mysteriously lacks a k as its final consonant (which it has in common with Ghadamès) the Proto-Berber word is *taʔfuḱt (Ghd. tofətt).
• təḥmáya 3sg.f.pf. 'to burn, scald'. Could be from two Arabic words: Ar. ḥamma 'to heat, make hot' or Ar. ḥamiya 'to be or become hot; to glow (metal)'. These two seem etymologically related in Arabic, though I've never heard of an -y intransitivising suffix.
• təqqím 3sg.f.pf. 'to stay' as an inchoative auxiliary verb.
• təttá 3sg.f.impf. 'to eat'. A well known irregular verb in Berber. Aor. əčč Pf. əčči/a Impf. tətti/a. The 3sg.f.impf. has lost its feminine prefix. If you imagine that word would have become **tətəttá, it doesn't take much imagination to figure out why it was simplified. Kossmann wrote an interesting article on the history of this verb. The verb 'to eat' is used symbolically here in the sense that it means that the sun is starting to cause a sunburn.
• afiš-ə́nnəs 'his face'
• id 'with' but here should be understood as 'together with, and'
• əlgəfá-nnəs 'his back of the head' < Ar. qafa(n) 'nape; occiput, back of the head; back; reverse'
• tăqqím 'she remained'. I think this should be read as '(the sun) continued (to shine)'
• am 'like'
• əžžaḥím 'hell' < Ar. žaḥīm 'fire, hellfire, hell'

u nəttín lləkká ižigáž u iʕəyáṭ: əlmluxíyət báhəyət

'and he was still walking and yelling 'Good mulukhiyah!'

• u nəttín 'and he'
• lləkká 'still'

U taxzə́r ġár-əs təmígni w-ətn-ís: i-kí a-wá-n ləmluxíyət

'and a woman looked at him and said to him: hey you with the mulukhiyah!

• U taxzə́r 'and she looked'
• ġár-əs 'to him'
• təmígni 'woman'
• w-ətn-ís 'and she said to him'
• i-kí this should mean 'hey you!', but I don't really understand how it should be interpreted. Surely Paradisi had a reason to supply this word with a dash, so what's i? and what's kí?
• a-wá-n ləmluxíyət 'o that one of the Mulukhiyah' a- 'vocative' wá 'dummy pronoun' n 'possesive'

U yəxzə́r nəttín ġár-əs w-igá əlḥəməl-ə́nnəs dít n təmígni u tənšə́d-t: z-díwa?

'And he saw her and he put his load in front of the woman and she asked him: how much?'

• U yəxzə́r nəttín ġár-əs 'and he saw her'
• w-igá 'and he put'
• dít n təmígni 'in front of the woman'
• u tənšə́d-t 'and she asked him'
• z-díwa? 'how much?' litterally this is s- instrumental suffix + díwa 'what' =  'with what?'

In-ís: s-sə́bəʕa n millím.

'He says to her: 7 miliemes'

• s-sə́bəʕa 'seven' < Ar. sabaʕa(t) 'id.' [Lameen:] s- is "for" with prices (at least in Siwi), so "for 7 milliemes".
• n millím 'of milliem' [kato:] this wouldn't be 'thousand' i don't think, as in "7,000 dinars". at that time in libya, the millim was the subunit of the girsh which was the subunit of the dinar, so there were 100 girsh in a dinar, and 1000 milliemes in a dinar. him selling molokhiya for ~ 7 milliemes sounds about right for the 1960s.

Tn-ís ənnát: s-ə́rbəʕa bə́ss

'She says to him: (I'll do it) for 4 (milliemes) only'

• Tn-ís ənnát 'she says to him'
• s-ə́rbəʕa literally 'with four'. < Ar. arbaʕa(t) 'four'
• bə́ss 'only, just'. [kato:] common in many dialects.

In-ís: ə́rbəʕa ḥə́nṭā.

'He says to her: 4 is (too) little'

• ḥə́nṭā 'a bit, a little'

Tn-ís: nək ná-ka s-ə́rbəʕa n millím, yəʕžəb-kúya naġ yəʕžəb-kúya-ká?

'She said: I have told you, for four thousand, does it please you or doesn't it please you?'

• nək ná-ka 'I have told you'
• yəʕžəb-kúya 3sg.m. '??' + 2sg. direct object. I'm not sure what this verb means. Ar. ʕažaba 'to wonder, marvel, be astonished' makes very little sense semantically. Apparently this word has undergone a semantic shift in Libyan Arabic. Maybe it means 'to please' now, 'does it please you, or doesn't it please you?' makes sense. [kato:] also common in dial. Ar. as a way of saying "to like" e.g. yǝ‘žǝb-ak "does it please you? (do you like it?)". it takes the suffixed obj. pronouns. [Lameen:] The relevant form is ʔaʕjaba "to please".
• naġ 'or'
• yəʕžəb-kúya-ká negative of above.

In-ís: yəʕžəb-íya-ká.

'He said: it does not pleace me'

Tn-ís: blaš, u təqqə́š əlbáb fəll-ís

'She said: "too bad", and closed the door in front of him (litt. on him).'

• blaš surely dialectal Arabic, don't know what it means. [kato:] blaš - dial. Ar. in Libya meaning 'free' usually. i can't think of using it to say 'too bad', but a shift wouldn't be implausible. [phoenix:] Paradisi translates it as non fa niente 'it doesn't matter' which still makes little sense to me in the meaning 'free'. Do you mean 'free' as in 'not locked up' or 'free' as in 'for no money'?
• təqqə́š 3sg.f.pf. 'to close', it is curious to note that according to Paradisi this verb literally means 'to close from the outside', which would make no sense in this context.
• əlbáb 'door' < Ar. bāb 'id.'

u baʕadén yəxzə́r af əlḥəməl-ə́nnəs wa-ẓẓákən dax a-yərfə́ʕ-t

'And then he looked at his load that was (too) heavy to carry'

• u baʕadén 'and then'
• wa-ẓẓákən wa+ptc. of stative verb ẓẓák 'to be heavy'
• dax 'for, in order to'
• a-yərfə́ʕ-t 3sg.m.aor. + 3sg.m. direct object 'to carry'

lákən yəkrí u yə́vdəd u baʕadén yəmmúdd afus-ə́nnəs yəddúgg af əlbáb

'but he returned and stopped and then he extended his hand and knocked on the door'

• lákən 'but' I'd expect a word that means 'so, therefore' here, not 'but'
• yəkrí 'he returned'
• yə́vdəd 'to stop, stand' 3sg.m.pf.
• yəmmúdd 'to extend' 3sg.m.pf. <Ar. madda 'to extend, distand, expand, dilate; to stretch, stretch out'
• afus-ə́nnəs 'his hand'
• yəddúgg 'to knock' 3sg.m.pf. <Ar. daqqa 'to knock, rap, bang (on the door)' it's interesting to see that Aujila seems to consistently render C1aC2C2a verbs as C1C1úC2C2 (Paradisi writes it as ú not as û, which may point to an accented short u or schwa instead).

u təškí təmígni ġár-əs w-ətn-ís: amúr džižít s-alúwəl axér-lək-ká

'The woman came out to him and said to him: if yo had sold (it) in the first place, would it not be better?''

• təškí 'to come out, to leave' 3sg.f.pf.
• amúr 'if'
• džižít 'to sell' 2sg.pf.
• s-alúwəl 'at first' maybe 'at the first time' < Ar. ʔawwal 'first, first time'
• axér-lək-ká 'would it not be better?' axér = better. -ká = negative. But what is lək? < Ar. la-ka 'to you' ? [kato:] you've analyzed it correctly. it's Ar. xēyr l-ak "better for you (m.)" [phoenix:] interesting that this phrase switches to full-on Arabic. Also not here that axér has vowel e instead of i that we've seen so far.

Yəkri-ká fəll-ís žlán

'He didn't answer a word to her'

• Yəkri-ká + af/fəll 'to answer' 3sg.m.pf. + neg.
• fəll-ís 'to her'
• žlán 'word, speech, story'

u baʕadén yərfə́ʕ əlmizán-nəs wa yəḥməl-tíya u yəxbóṭ-ṭ af təgíli n təmígni

'and then he lifted up his scales that he was carrying and he smashed them on the head of the woman'

• yərfə́ʕ 'to lift up' 3sg.m.pf.
• əlmizán-nəs 'his scales'
• wa yəḥməl-tíya 'that he carried'
• yəxbóṭ-ṭ 'and he smashed it (scales is singular)' <Ar. xabaṭa 'to beat, strike'
• təgíli 'head'

tíva təmígni təmmút

'the woman fell and died'

• tíva 'to fall' 3sg.f.pf. This verb has a very strange imperfect form inə́vva with a prefix n- and gemination of v, which is usually avoided in Berber. [Lameen:] no, biconsonantal verbs of the form *ənC become iC in Awjila, so the original form of this verb is *ənv. Cp. iž impf. nəžža "to be sold", iš impf. nəšša "sleep".
• təmmút 'to die' 3sg.f.pf.

u yəqqím yəvdída agur-ə́nnəs ir a-yúš əl-bulís

'and he stayed standing in her vincinity until the police came'

• yəqqím 'he remained'
• yəvdída 'to stop, to stand' 3sg.m.pf.
• agur-ə́nnəs 'near her'
• ir 'until'
• a-yúš 'to come' 3sg.m.aor.
• əl-bulís < Ar. būlīs 'police' < Fr. police 'id.' Arabic and Berber don't have p so they settled for the next best thing. [kato:] iIthink it could be English or Italian as well, since the French were never really in Libya and I don't see why the libyans would have borrowed the word from another arabic dialect. [Phoenix:] Italian seems unliekly polizia would probably be rendered different in Arabic. English is of course possible.

I am not really sure what invokes the rage of the man.

A deceptively small country

While reading through the Aujila texts I have noticed something about the word ašal usually glossed as 'country' or, in French 'pays'. This suggests that it is related to PB *aʔḱal 'earth' > Kb. akal, MA akal, ašal Tashl. akal, Ghd okăl. okal , Zng aʔgäy To. ăkal, which all have the meaning 'earth'.

The correspondence is problematic neither *k nor *ḱ ever correspond to Aujila š except for this word.

If we take a look at the actual entry in Paradisi's wordlist we learn that ašal is glossed 'paese, villaggio'. This could mean 'country, village'; A bit of a wide range of semantics don't you think?

It just so happens, that Italian paese has a second meaning 'small village'. I think it's much more likely that this is the meaning Paradisi is aiming at with paese, since this would explain why he casually wrote 'paese, villaggio' as the gloss without pointing out the huge disparity in meaning and that the entry 'villaggio' has 'ved. "paese" ' as it's only comment.

I believe that we are dealing with a Aujila word ašal '(small) village' that has absolutely no relation to PB *aʔḱal 'earth'. A mistake that has gone unnoticed for years because of the obvious phonetically similar French pays 'country' (obviously, also etymologically related).

Of course, it is still possible to argue that '(small) village' could have been derived from 'ground' as well. After all, the word paese underwent a similar semantic shift. But add to it that the correspondence is imperfect and that even 'country' isn't quite 'ground' yet, I think this word probably isn't a cognate.

Proto-Berber C1C1C2 verbs

Me and Glen Gordon have been brainstorming furiously about some of the bigger issues in Proto-Berber reconstructions on his blog.

A lot of the discussion ended up coming down to the prefixes of Berber nouns, and specifically the prefix *u- as in *ulβ 'heart' (or *ulv, when accepting Glen's suggestion for a changed PB orthography, which I won't yet, as long as I haven't discussed it on this blog).

Some evidence leads me to believe that this prefix *u- is originally from a *w-, that came to be vocalised as *u- in some way. We see this because Iwwellemeden əwəl 'heart'. Other evidence presents itself in *urə/ăγ (*urə/ăḳ) 'gold' which seems to be derived from a root *wrγ 'to be yellow'.

While this is promising, we do not yet understand the deeper underlying ways that PB *w interacts with other consonants and vowels. To get a sense just how difficult it is to understand what exactly is going on, I would like to discuss a certain class of Berber verbs today, the so C1C1C2 verbs. Such verbs have a geminated initial consonant and a second consonant. Prasse (1972-1974) believes such verbs to have originally been wC1C2, and he has a point. Because of the complex ablaut patterns of the verbal system, we get a a good view of the many different interections in different positions of *w with surrounding vowels and consonants.

But, before we start looking at a difficult class like C1C1C2 verbs, let' s first establish a general reconstruction of the C1C2C3 verbs the most 'unmarked' form. It should be noted that, in general, the verbal system of Berber is quite irregular, so it is not always easy to establish a regular system of reconstruction.

Let us use the root *KRZ 'to cultivate; to acquire'

For Proto-Berber we can (probably) reconstruct four tenses: Aorist, Perfect, Imperfect, Negative Perfect and Negative Imperfect.

For triradical roots the form is usually:

Aor. ăC1C2əC3 Perf. əC1C2ăC3 Impf. C1ăC2C2ăC3, Neg.Perf. əC1C2eC3, Neg.Impf. C1əC2C2əC3

So for the root KRZ its respective forms will look as follows:

Aor. *ăkrəz Perf. *əkrăz Impf. *kărrăz Neg.Perf. *əkrez Neg.Impf. *kərrəz

Even within this elegant system, we sometimes find irregularities in some languages. For example in the verb LMD 'to learn'  (a loanword from Punic) we usually find this system as above, but the Middle Atlas variety described by Amaniss we find a different form for the Imperfect.

Aor. əlməd Perf. əlməd Impf. tt-ələmad Neg.Perf. ləmid (MA doesn't have a Neg.Impf.)

In Mali Touareg (Heath 2006) this verb behaves as expected.

The imperfective prefix *tt- often shows up in verbs that have two radicals, or whose first radical is somehow 'weak'. The root *ʔKR 'to steal' for example conjugates as follows:

Aor. *ăʔkər Perf. *əʔkăr Impf. *tt-ăʔkăr Neg.Perf. *əʔker Neg.Impf. *tt-əʔkər

Word with an initial long vowel (<*h?) behave in the same way.

 So, since C1C1C2 verbs were originally *wC1C2, they technically have three radicals as well, so formally you would expect them to behave in the same way as a triradical root.

Let's take the root *KKS (<*WKS) 'to take away'

Aor. *ăkkəs Perf. *əkkăs Impf. *tt-ăkkăs Neg.Perf. *əkkes Neg.Impf. *tt-əkkəs

As you can see *w too is treated as a weak initial radical. Berber seems to have some avoidance of initial *w, *ʔ and long vowels (< *h?), but from these examples we can at least probably conclude that *ăwC and *əwC yield *ăCC and *əCC respectively.

Watchful readers may have noticed that there is not much evidence for a third consonant *w at all in the examples above. This is true, this suspcion is based on both the Verbal Noun formation and the verbal derivatons.

Verbal derivations work as follows *ss- is the causative prefix (no doubt familiar to those who have worked with Semitic) and *mm- is the reciprocal prefix. *ss- assimilates to any fricative in the root, and *mm- dissimilated to *nn- if there is an *mm- in the root. The passive prefix *tt-/*ttw- is also common and probably related to the *tt- prefix that you see in the weak-initial roots.

*ss-LMD 'to teach'

Aor. *əss-əlməd Perf. *ăss-əlmăd Impf. *əss-ălmad

*mm-LMD 'to get to know each other'

Aor. *ənn-ălmăd Perf. *ănn-əlmăd Impf. *tt-ănn-ălmăd

*ZZY 'to get better'

Aor. *ăzzəy Perf. *əzzăy Impf. *ətt-ăzzăy

*ss-ZZY 'to cure'

Aor. *əzz-uzăy Perf.  *ăzz-uzăy Impf. *əzz-uzuy/*əzz-izuy

The variant *əzz-izuy as the impf. is quite common, and may be a dissimilation.

As you can see, the geminate is gone, and instead a u appears. This is the main reason why an initial radical *w is assumed (besides the verbal noun scheme of this type of verbs is uC1əC2).

So assuming that the *u is a regular reflex of *w in this context, can we understand why in the non-causative form we find a geminate and in the causative form we find a *u + single consonant?

Imposing the regular triradical vowel scheme of *ss-LMD on the once of *ss-WZY we would expect a hypothetical Pre-Proto-Berber paradigm:

Aor. **əss-əwzəy Perf. **ăss-əwzăy Impf. **əss-ăwzay

These forms definitely do not wrok perfectly. The aorist of *ss-WZY clearly has a final *ă. This in itself is quite surprising as ə-ə-ă is not a typicaly vowel pattern for the aorist.

As Kossmann shows in his Ayer Touareg grammar (pg. 57/58) Perf. vowel pattern *ă-ə-ă correspondes to *ə-ə-ə. While these vowel patterns are not directly assigned by him to causatives, the form of *ss-LMD seems to suggest that they do apply to to these derivations. (I'm pretty sure Kossmann wasn't the first to suggest this vowel pattern, nor is this the first publication to print this table, but I'm not sure who first proposed it).

The Perfect corresponds perfectly with *ss-LMD. The Imperfect differs quite a bit, as the 'full vowel' inserted before the last consonant in the WCC verbs is an *u not an *a. This almost looks as a sort of progressive vowel harmony. Let's, for the sake of the argument assume this is true, and let's assume that The Aor. form somehow innovated the final *ă, and that the Proto-Berber pattern followed the regular Aorist pattern *ə-ə-ə.

We are left with yet another problem, it would mean that əwC and ăwC yield *u, while earlier we established that these sequences yield əCC and ăCC.

The difference is, that this time there is a causative suffix *ss- in front of it. I don't find *ssə/ăwC > *ssuC a particularly convincing sound law, since nothing about the sound *ss motivates vocalisation. moreover the *mm- reciprocal causes the same vocalisation.

The thing *mm- and *ss- do have in common is that they're both geminate consonant prefixes (although this is not always the case in all Berber languages). Since these prefixes are geminate, and the person prefixes are not, in some way syllable weight or shape may play a role here.

It's almost too obvious to state, but because of the prefixes, the sequence ə/ăwC is always in a CVC context while in the non causative/reciprocal form it is sometimes in a VC context (1sg., 3pl.m./f. and the imperatives). But we would expect that the non-prefixed forms would then have an alternation between *ə/ăwC and *uC. Analogy could obviously play a role, but are the far more marginal 1sg., 3pl. and imperatives enough to influence, for example the commonly occuring 3sg.m. and 3sg.f.?

Another issue with this hypothesis is the question why the *tt- imperfective prefix doesn't cause vocalisation.

Also, if this hypothesis is correct (and I'm far from certain it is), we get in trouble with the verbal nouns of this type whose shape is *uCəC, this would mean we'd need an extra consonant prefixed (*w??) to get the desired syllable shape *wə/ăwCəC is not particularly elegant. Once again we can pull out the 'analogy' get-out-of-jail-free-card and say that the the original form was *awCəC and that the EA (état d'annexion) was *w-ăwCəC, and that the EA spread to the EL (état libre). But this is not entirely in line with what we know about triradical verbal nouns either which tend to have the shape *aC1C2aC3 and *aC1əC2C2uC3.

A final point that casts some doubt on the C1C1C2 verbs = wC1C2 verbs, is the fact that Proto-Berber has no issue with the sequence *ə/ăw or *ə/ăw in final position at all, where, without fail it is found in a CVC shape, without changing into *u (although it does in a great deal of Berber languages, comparative evidence shows this wasn't the case in Proto-Berber). But we should not that there indeed is a eerie absence of wC1C2 and yC1C2 roots.

To sum up:

While it is tempting to think that C1C1C2 verbs originally had an initial semi-vowel. It is almost impossible to find a defining context which motivated the assimlation of wC > CC on the one hand and a vocalisation of *ə/ăw > *u on the other hand. The Verbal nouns present a whole extra issue as their phonetic context seems more similar to the unprefixed forms of C1C1C2 verbs, while their reflex of *ə/ăw looks like a prefixed form.

Maybe the solution to this problem will not be available through internal reconstruction. Maybe, or knowledge of Semitic vocalic patterns can give us more insight. It is not outside of the realm of possibility that the Berber vocalic patterns underwent metathesis. Let's say if the Perfect triradical form actually corresponds to the Proto-Semitic perfect *C1aC2aC3 it would mean that *əC1C2ăC3 was originally *C1əC2ăC3 (vocalic problems aside). Then the causative form would correspond to the Proto-Semitic causative *S-vC1C2vC3, Proto-Berber *ss-əC1C2ăC3

So in a Pre-Proto-Berber stage we would have:

Stage 1: *wəzăy Causastive *ss-əwzăy.

Shift one: *əw > *u occurs.

Stage 2: *wəzăy Causastive *ss-uzăy

Shift two: vocalic metathesis occurs

Stage 3: *əwzăy Causastive *ss-uzăy

Shift three: assimilation of semi-vowel occurs

Proto-Berber *əzzăy Causastive *ss-uzăy

Would that work? Could it also account for the fact that the *tt- prefix does not cause this vocalisation of *u? I believe Proto-Semitic had *t-C1vC2vC3 > Arabic form VI t-CāCaC with a prefix *t.

Form VI is problematically reciprocal while form VII (*n- prefix ~ Berber *mm-?) is passive, which is the exact opposite of the Berber semantic distribution.

The solution is elegant, but I know nowhere near enough about Semitic vocalisation (and Proto-Afro-Asiatic for that matter) to confirm this hypothesis.

I hope this article hasn't become too rambly, and that, those who are interested can make sense of it. It's a difficult subject with a lot of different factors. Please let me know if anything is unclear and I'll attempt to rewrite it. I don't want you to break your head over some impossible formulation.

[EDIT] I just realized that in Semitic Linguistics, the Perfect is usually considered a secondary development from an original past participle, which explains that the suffix conjugation that goes with the perfect looks like the possesive conjugation. The 'true' form is the imperfect aC1C2iC3 with prefix conjugation (like Berber). So the clever scenario I came up with doesn't hold much ground, unless you want to assume a Berbero-Semiticsub-branch of Afro-Asiatic that had already developed this past participle class, after which Berber lost its suffix conjugation.

While possible, this is certainly a few too many steps with too little proof to feel entirely comfortable positing this. Alternatively it maybe be possible that the root of the perfect is similar to the Arabic form III, but I don't see the relation between transitive and perfect, nor do I think Arabic form III is reconstructable for Proto-Semitic.

Buying your own donkey, and losing your Bracelet.

Another story, that is actually two stories in one.

Yəškí Žḥá ssúq-i u ídd-əs aẓiŏṭ-ə́nnəs iġə́lli a-ižiž-ít

'Žḥá left to the market, and with him (was0 his donkey, he wanted to sell it.'

• Yəškí 'to leave' 3sg.m.pf.
• ssúq-i 'to the market'
• u ídd-əs 'and with him'
• aẓiŏṭ-ə́nnəs 'his donkey'
• iġə́lli 'to want' 3sg.m.pf.
• a-ižiž-ít 'to sell' 3sg.m.aor. +  3sg.m direct object suffix. žiž is actually a causative formation of iž 'to be sold' with the common Berber assimilation of -ss- caustive prefix to the following sibilant.

Baʕadén yəfk-ít əddəllál dax a-ižiž-ít

'Then he gave it to the Auctioneer so that he would sell it.

• Baʕadén 'then'
• yəfkí-t 'he gave it' Notice that the giving construction her is a double transitive, rather than a accusative-dative construction.
• əddəllál 'auctioneer' < Ar. dallāl 'id.'
• dax 'for'

Baʕadén yəqqím əddəllál iʕəyáṭ w-itənn-ís:

'Then the auctioneer started to shout and he said:'

• yəqqím 'to stay' used inchoatively.
• iʕə́yáṭ 'to shout' 3sg.m.impf. < Ar. ʕāṭa 'to yell, scream'
• w-itənn-ís 'and he was saying' 3sg.m.impf.

Wáya d-aẓíṭ fəšúš qə́wi u d-aẓíṭ mrí

This is a agile and strong donkey, and it is a beautiful donkey.

• Wáya 'this'
• d-aẓíṭ 'is a donkey'
• fəšúš 'agile', this word wasn't in Paradisi's wordlist. [Lameen:] fəšúš is pan-Berber (ifsus, etc.) = "light, fast"
• qə́wi 'strong' < Ar. qawīy 'id.'
• mrí 'beautiful'

Baʕadén əqqimán míddən rənníyən (or rənnán) fəll-ís s-ar žlan-íya

'Then the people started bidding on it because of this story (the Auctioneer's)'

• əqqimán 'to stay' 3pl.m.pf. Once again used inchoatively.
• míddən 'people'
• rənníyən ptc. of 'to increase' which here means 'to bid', Paradisi also says rənnán is an option, which would be the 3pl.m.pf.
• fəll-ís 'on it'
• s-ar 'because of'
• žlan-íya 'this story'

Baʕadén yəslá Žḥá žlan-íya w-in-ís y-imán-nəs:

'Then Žḥá heard this story and said to himself:'

• yəslá 'to hear' 3sg.m.pg.
• y-imán-nəs 'to himself'

la-búdda aẓiŏṭ-ənnúk kuwə́yəs ṣəḥíḥ

'Certainly, my donkey is the most healthy!'

• la-búdda 'certainly' looks like something Arabic, but I don't know the expression. [Kato:] Ar. la budda 'certainly'
• aẓiŏṭ-ənnúk 'my donkey'
• kuwə́yəs should mean 'the most', but I don't recognize the word. [Lameen:] 'good' (Egyptian Arabic, Siwi...)
• ṣəḥíḥ 'healthy' < Ar. ṣaḥīḥ 'id.'

Baʕadén yqəqím irə́nni id míddən.

'Then he started to bid (along) with the people'

• yqəqím must be a typo for yəqqím
• irə́nni 'to bid' 3sg.m.pf.

Baʕadén yəškí əlbarát u yəfkí-tənət y-əddəllál u yuġá aẓiŏṭ u yəʕádd yəfríḥa.

'Then he left the money (?) and he gave it to the auctioneer and he bought the donkey and he went home happily.'

• yəškí 'he left'. I don't understand what this verb is doing here. I'd expect a verb that mean 'he took'
• əlbarát 'the money', This is a plural noun.
• u yəfkí-tənət 'and he gave them', them referring to 'the money'
• y-ăddəllál 'to the auctioneer'
• u yuġá 'and he took'
• u yəʕádd 'and he went'
• yəfríḥa 'he was happy' < Ar. fariḥa 'to be glad, happy, delighted'

Baʕadén əlʕəqqəáb-i yəqqím nəttín id-dməgní-s (or təmigní-nnəs) u yəḥk-ís af ləḥkáyət n-arənní

'Then in the evening he was with his wise, and he told her the story of the auction'

• əlʕəqqəáb-i 'in the evening', obviously an Arabic word underlies this, but I could not find it. [Lameen:] Derived somehow from ʕaqb "after".
• yəqqím not inchoative this time, here it's a locative verb 'to stay; to be at a place'
• nəttín 'he'
• id-dməgní-s 'with his wife', this noun takes the kinship term possesive -s rather than possesive -(ə)nnəs. But Paradisi points out that təmigní-nnəs is possible too.
• u yəḥk-ís 'and he told her', not that the indrect object pronoun of the third person does not distinguish in gender. This is surprising as the 2nd person is distinguished by gender, and the direct object pronoun is too. In verbal agreement, 2nd person singular doesn't distinguish gender, while 2nd person plural and 3rd person singular and plural both do. < Ar. ḥakā 'to tell'
• af ləḥkáyət literally 'on the story', but means 'about the story' < Ar. ḥikāya(t) 'story, tale'
• n-arənní 'of the bidding' 'bidding' here is the verbal noun of the rənni 'to increase, bid'.

Tn-ís ənnát: qáma annák laḥkáyət axír n-tánnək.

'Then she said: I will tell you a story better than yours'

• Tn-ís 'to say' 3sg.f.pf. + 3sg. indirect object.
• ənnát 'she'
• qáma 'now'
• anná-k  'to say'1sg.aor. + 2sg.m. indirect object suffix.
• axír 'better'
• n-tá-nnək 'of yours', with 'dummy' feminine pronoun ta+possesive suffix. 'of yours' to mean 'than yours' surprises me. Is this common in Berber? [Lameen:] 'of yours' to mean 'than yours': n is commonly used as "than" with Arabic comparatives loaned into Berber, even in languages like Kabyle that otherwise prefer the "on"-structure (see discussion in my thesis.)

Tn-ís: yəxə́ṭṭəm síla wa žižán ləmluxíyət,

'She said to him: Someone who sells 'mulukhiya' came by,'

• yəxə́ ṭṭəm 'to drop by someone' 3sg.m.pf.
• síla wa žižán 'someone who sells'. The 'who sells' part is obvious, wa + ptc. 'to sell', but I have no idea what síla means. Is it Arabic?
• ləmluxíyət 'Mulukhiya' < Ar. muluxiya(t) 'id.'

yəqqím itazən-dík

'He started weighing it for me'

• itazən-dík 3sg.m.impf. 'to weigh' < Ar. wazana 'to weigh'

Baʕadén gíx əddə́mləž n əddə́həb əlkə́ffət ta-gán əlmizán

'Then I put (my) golden bracelet on the plate of the scales'

• gíx 'to do, put' 1sg.pf. There is no putting on preposition. Can this word be used as a double transitive in the same way as 'to give'?
• əddə́mləž 'bracelet' < Ar. dumluž 'id.'
• n əddə́həb 'of gold' < Ar. dahab 'gold'
• əlkə́ffət 'scale/plate of a scales' < Ar. kiffa(t) 'id.'
• ta-gán ??? ta- feminine pronoun + gán 'there is' particle? or gán as the ptc. of 'to do, put'?
• əlmizán 'scales' < Ar. mīzān 'balances, scales' <- all together 'plate that is on the scales'?

Baʕadén yəkkə́mməl uzún-nəs

'Then he finished his weighing'

• yəkkə́mməl 'to finish' 3sg.m.pf.
• uzún-nəs 'his weighing' uzún is the verbal noun of zan 'to weigh'

ənṭə́rx əddəmləž-ənnúk əlmizán-i rwíḥḥa a-ini-dík: kəm ttákə rt s-ġár-i.

'I had left my bracelet on the scaled and I was afraid that he would say to me: you are stealing from me!'

• ənṭə́rx 'to leave' 1sg.pf.
• rwíḥḥa 'to be afraid' 1sg.pf. (This assimilation has been mentioned before. *ʕx > ḥḥ
• a-ini-dík 'he would say to me'
• kəm 'you'
• ttákərt 'you are stealing',  2sg.impf. of akər 'to steal'.
• s-ġár-i 'from me'

Baʕadén in-ís Žḥá: báhi, šuġarx s-əlhəwáyyi u kəm s-ar-əzgíg u tudík a-ttaʕmír təfilli-nnáx

'Then Žḥá said to her: 'very well, I have lost something outside (of home) and you inside (of home), and so (we lose) our house's prospering.'

• báhi, 'very well' [Kato:] dial. Ar. bāhi 'good, ok, fine'
• šuġarx 'to lose' 1sg.pf.
• s-əlhəwáyyi 'from outside'
• u kəm 'and you'
• s-ar-əzgíg 'from inside'
• u tudík 'and so'
• a-ttəʕmír 'to prosper' 3sg.f.aor. < Ar. ʕamura ‘to thrive, prosper, flourish, flower, bloom’ [Lameen:] If attaʕmír really is "prosper" (ʕmaṛ in Siwi just means "do, be"), then it's intended ironically. [Phoenix:] It is not actually in Paradisi's wordlist,so maybe the meaning related to that of Siwi.
• təfilli-nnáx 'our house'

I'll admit that this last sentence with attaʕmír isn't flowing too well. Maybe I'm missing out on something. Paradisi translates: Le disse quindi Žḥâ: bene, io perdo da una parte e tu dall'altra, così funziona la nostra casa.

impf.

The origin of PB Noun prefix *e

Proto-Berber nouns have 4 types of prefixes. *a, *e, *i and *u.

Feminine nouns circumfix nouns with an additional *t-...-t (in several cases the final *-t is dropped).

• *afuʔs 'hand'
• *esen 'tooth'
• *udəm 'face'
• *ils(əʔ?) 'tongue'
• *ifaʔssăn 'hands'

All these words can appear in a 'État d'Annexion' or EA 'annexed state'. This state is used when a word is the subject of a verb, when directly following that verb, or when it comes after (most) prepositions.

Tha annexed state is forms by shortening the prefix vowel and adding a prefix *w-. For words with initial *i-, this *w- is changed to *y-.

This shortening is not universal, some nouns retain their initial vowel even in the EA. *u- is always stable. Initial *i is very commonly stable, but when it is a plural prefix of *a and *e, it's generally unstable.

Touareg has this shortening, but lacks the prefix *w-. Ghadamès, Zénaga and Aujila Berber lack an EA alltogether.

• *w-ăfuʔs
• *w-ăsen (We cannot say for sure that this form had *w-, as only Touareg retains this EA which lacks *w-)
• *w-udəm
• *y-ils(əʔ?)
• *y-əfaʔssăn

As you can see, in the EA *a- and *e- are indistinguishable from eachother. The basis of this reconstruction is a bit slim. Only Touareg retains this EA for *e-, in all other languages with an EA *e- has shifted to i- and subsequently as i- in the EA. But as there is no reason to Believe that Touareg innovated, while for the other language there is clear reason to assume an innovation, it would seem that *ă- as the EA of *e- is original.

This reflex is not exactly obvious. *ă is the short reflex of a low vowel, while *e- is rather a high vowel.

So why is this? While I don't exactly have the answer, nor definite proof for this, I think I have some evidence to show that *e- might have been *a- in Pre-Proto-Berber.

As my MA Thesis shows (check out Academia.edu) ,  nouns that have an *e in the root will invariably have an initial *e. That's an odd distribution. no *aC(C)eC words exist. Since *a- is by far the most common prefix, I think it is safe to assume that some kind of 'umlaut' has taken place that shifted *a- to *e- in front of another *e. Especially considering their identical EA I think this is a likely scenario.

So is that it? That seems simple enough. Sadly, not it isn't. In my small sample, there are no *aC(C)əC nouns and there are *aC(C)ăC and also *eC(C)əC . This is once again a similar 'umlaut'-like situation. But my sample is much too small (and too specifically about biradical nouns) to confirm this distribution. Also, if we expect that *ə and *e cause umlaut, why not *i? There are definitely *aC(C)iC nouns, and I can't think of any *eC(C)iC only the fly.

Then, there are some formations that do overlap, even in my small sample. Both *aCăC and *eCăC nouns exist, both with proper reconstructable attestations.

So is my hypothesis wrong? Maybe. It could also be 'a little bit correct'. The complementary distribution of *e and *a in *aC(C)eC nouns is striking. Maybe the initial *e- has two origins. 1. Umlaut of *a 2. An original initial *e-.

At my MA Thesis defense, Maarten Kossmann suggested that maybe the *eCeC type nouns may be a 'biradical root only' formation (a very popular formation for biradical verbal nouns by the way). While my thesis mentions one word with direct counterevidence *ezyeḍ 'donkey', this word is problematic. The *e is based on Touareg and Ghadamès, which both have an emphatic *ḍ which would color a root *i to e. So this word may have been *ezyiḍ instead.

More thorough research will have to be done on the exat distribution of *e as opposed to *a, but I've made a first step in the direction.

Count it!

Yəškí iwín əlʕáləm iddəhwár lhəwáyyi u yəqqím iddəhwár kkúll inə́ššəd əlʕulamá u iġəllə́b-tən

'A scholar left and went around out (in the world) and was going around everywhere asking scholars (questions) and he bested them all.'

• Yəškí 3sg.m.pf. 'to leave'
• iwín 'one'
• əlʕáləm 'scholar' < Ar.ʕālim 'id.'
• iddəhwár 3sg.m.impf. 'to go around' < Ar. dahwara 'to hurl down; to tear down, topple, overthrow'. ??? [kato:] ELA īdahwar 'to take a tour, to walk around'
• lhəwáyyi 'outside, outdoors, out, abroad' < Ar. hawāʔī ‘airy, breezy; aerial’ (?)
• u yəqqím 'and he remained', used as an auxiliary verb here to express a continuous action
• kkúll 'all, each' but also 'everywhere' in this context apparently < Ar. kull 'all, each'.
• inə́ššəd 3sg.m.impf. 'to ask', an imperfect form of a verb we've seen earlier < Ar. našada 'to ask'.
• əlʕulamá plural of 'scholar' < Ar. ʕulamāʔ ‘scholars’
• u iġəllə́b-tən 'and he defeated them all'. I expect a Perfect here, but the form looks like an imperfect. Alternatively it's form II of Ar. ġalaba ‘to subdue, conquer, vanquish, defeat’, but that has a passive meaning in MSA (how about dialectal Arabic?). Even if it was form II, I'd expect the form to have short a's in the root. so, the question remains: why is this form in the imperfect?

u nəttín iddəhwár ir iṭúg Žḥá.

'and he went around until he reached Žḥá.'

• u nəttín 'and he'
• ir 'until'
• iṭúg 3sg.m.pf. 'to reach'. [Kato:] this vb. means 'to reach, to get to' in ELA. īṭugg (geminate g)

Baʕadén inšə́t-t w-in-ís: mag ínni ənnúṣṣ ən təmúrt kkúll?

Then he asked him and said: Where is the center of the whole world?

• Baʕadén 'then'
• inšə́t-t 3sg.m.pf. 'to ask' with 3sg.m. direct object suffix which causes an assimilation of the d of the verb to t.
• w-in-ís 'and he said to him'
• mag 'where?'
• ínni 3sg.m.pf. 'to be, exist', we usually find this word as yə́nni. Quite surprising that the Schwa is not found in this form.
• ənnúṣṣ < Ar. nuṣf 'half' but here it means 'center, middle'
• ən 'of' in a vocalising position.
• təmúrt 'earth'
• kkúll 'whole'

In-ís Žḥá: amakán wa ínni aẓiŏṭ-ənnúk w-əndú tṣəddəqt-íya-ká əžvóṭ təmúrt

'And Žḥá said to him: On the place where my donkey is, and if you do not believe me, mesure the ground!'

• amakán wa ínni aẓiŏṭ-ənnúk 'place where my donkey is', a relative sentence construction with wa.
• w-əndú 'and if', it should be noted that Paradisi transcribed this word here as w-öndú, which is sufficiently exotic to mention, as I had not yet seen him transcribe ö for anything before.
• tṣăddăqt-íya-ká 2sg.pf. 'to believe' + direct object 1sg. + negative particle < Ar. ṣaddaqa 'to believe'.
• əžvóṭ imperative sg. 'to measure'. The first word with a v! Aujila Berber takes up a special position in Proto-Berber reconstruction because it retains this sound, it is therefore quite funny that we only run into it now. To make matters worse, I don't think this word can be reconstructed for Proto-Berber.

Itəḥáyyar əlʕáləm w-inšə́t-t: ammágwa əlʕə́dəd nə-nǧúm?

'The scholar is confused and he asks him: What is the number of stars?'

• Itəḥáyyar 3sg.m.pf. 'to be confused' < Ar. taḥayyara 'to become confused'
• ammágwa 'what, which'; Paradisi lists this word as mágwa in his wordlist, is this a variant, or is the am- part a morpheme?
• əlʕə́dəd 'number'< Ar. ʕadad ‘number, numeral’
• nə-nǧúm? 'of stars' < Ar. nužūm 'stars', this word seems to lack the almost always universally present Arabic definite article a(l)-, I would expect the form to be **ənnəǧúm or something, but maybe phonetic environment did not allow for this article to be clearly reflected.

In-ís Žḥá: d-əlʕə́dəd n ažə́vu n aẓiŏṭ-ənnúk, undú tṣəddəqt-íya-ká əḥsə́b ənǧúm w-əḥsə́b ažə́vu n aẓiŏṭ-ənnúk

'Žḥá said to him: (it) is the number of hairs of my donkey, if you don't believe me count the stars and count the hairs of my donkey'

• d-  copula
• ažə́vu 'hair'. Another word with v and once again without a good Proto-Berber etymology!
• aẓiŏṭ-ənnúk 'my donkey'
• undú 'if', this time transcribes with u initially
• əḥsə́b imperative sg. 'to count' < Ar. ḥasaba 'id.'

Baʕadén yənšə́t-t amédən u in-ís: ammágwa elʕə́ded n ažə́vu n tamirt-ənnúk

'Then the man asked and said to him: What is the number of hairs of my beard'

• amédən 'man' <- Interesting how the text switches from 'scholar' to 'man' here.
• tamirt-ənnúk 'my beard'. A very well attested Berber word < *taʔmart, surprisingly has a vowel i in the root. This could be an analogy from the plural *tiʔmiraʔ, but may point to a greater tendency of PB *a to shift to i in Aujila Berber. I have looked a little bit into it, and it seems that *a shifts to i before *r, *l, *m, *n. But further research needs to be done.

In-ís Žḥá: əlʕə́dəd n ažə́vu n tamirt-ə́nnək am əlʕə́dəd n azíf n aẓiṭ-ənnúk, w-undú tṣəddəqt-íya-ká əḥsə́b.

'Žḥá said to him: The number of hairs of your beard is the same as the number of the tail of my donkey, if you dn't believe me, count.

• am 'like, as'
• azíf 'tail'

Baʕadén ittəwíyəl amédən u yəkrí yəndíma.

'Then the man was amazed and he returned home being sad.'

• ittəwíyəl 'to be dumbfounded, amazed'. I could not find this in Paradisi's wordlist. [Kato:] ELA yetwayyel 'to be dumbfounded, to wonder at'.
• yəkrí 'to return' 3sg.m.pf.
• yəndíma 'to be sad, regretful' 3sg.m.pf. Paradisi translated 'pentito'.  Seems to be a derivation from Ar. nadima 'to repent, rue, regret'.

Come up with me

Žḥá yəskína lʕəllí-y-i,

'Žḥá lived high-up'

• yəskína 3sg.m.pf. 'to live' Probably from Ar. sakana 'to be or become still, tranquil, peaceful; to calm down, repose, rest' which underwent a semantic shift in dialectal Arabic. Can somebody confirm this semantic shift? [bulbul:] SKN means "to live, to reside" in Classical Arabic already, it shows up in Lisan al-Arab.
• lʕəllí-y-i < Ar. ʕallīy 'high' + locative particle i  so 'in a high place'.

baʕadén yušád iwínan isása af təfillí-nnəs

'then, a beggar came to his house'

• baʕadén 'then'
• yušád 'came' 3sg.m.pf.
• iwínan isása 'one' + 3sg.m.pf. 'to beg' 'one that begged' a subordinate clause.
• af təfillí-nnəs 'to his house'

Baʕadén yə́xzər Žḥá s-ar-ẓúṭ w-in-ís: ddíwa tġallít a-y-amédən?

'Them, Žḥá watch (him) from above and said to him:'

• yə́xzər 'to watch' 3sg.m.pf. Can't find it in Paradisi's wordlist, which probably means it's dialectal Arabic. [bulbul:] There's a Moroccan verb xzer meaning "to look threatingly at so/st".
• s-ar-ẓúṭ 'from above'
• w-in-ís 'and he said to him'
• ddíwa 'what'
• tġallít 'to want' 2sg.pf.
• a-y-amédən 'o man'

In-ís amédən: yi-d-díla

The man said to him: come here

• yi-d-díla imperative singular 'to come' + díla 'here'

yuš-íz-d Žḥá.

'Žḥá came to him'

• yuš-íz-d 3sg.m.pf. 'to come' + íz for 3sg.m. indirect object.

Baʕadén in-ís amédən: fki-dík s-ar/s-ġár əlḥə́qq ər-rə́bbi

 'Then, the man said to him: give to me (something) for the God's sake.'

• fki-dík imperative 'to give' + 1sg. Indirect object
• s-ar/s-ġár 'from'
• əlḥə́qq 'rightness', we've seen this word earlier where it meant 'one's rightful payment', that translation doesn't quite cut it here. Something like 'from that what belongs to me in the name of God'. I'm probably butchering some Islamic formula for beggars here. Anyone with a less terrible understanding of the Islam help me out here. [bulbul:]The formula the beggar uses is a classic, you will find variations on it in, say, The Arabian Nights translated as "give me something for God's sake", "out of that He has given you" or, probably closest to what we have here, "so that God will reward you".
• ər-rə́bbi 'of god' with an interesting assimilation of nr to rr. This assimilation is quite common in other Berber languages too, such as Middle Atlas Berber (where also rn yields rr).

Baʕadén in-ís Žḥá: yíd ẓóṭ.

'Then, Žḥá said to him: come up'

• ẓóṭ 'up', interestingly in earlier texts we found this word as ẓuṭ. In Paradisi's transcription: ẓūṭ, ẓôṭ. This time the word is stressed, maybe this has something to do with it.

U yəlbúb nəttín ídd-əs ar-ẓúṭ w-in-ís: ġar-i-ká

'and he climed upward with him, and he said to him: 'I don't have anything!''

• U yəlbúb 'and he climbed', this is the same word that meant 'to ride' in an earliers text I translated. So it's more 'to climb onto a donkey' and 'to climb up'.
• nəttín 'he' independent pronoun
• ídd-əs 'with him'
• ar-ẓúṭ 'upward'
• ġar-i-ká 'to me-not' a way of expressing possesion with ġar and the negative. This is interesting, because in the story with the red shirt, I translated fəll-əs to express 'to have', in retrospect it would be better to translate that sentence as 'on him was a red shirt'. This is apparently the most common way to express 'to have' as this construction is mentioned in Paradisi's wordlist under avere 'to have'.

Baʕadén in-ís wa sásan: af-íwa tnəd-dík-ká žlan-íya u nək nníx ar-úṭā?

'Then the beggar said to him: Why did you not tell me this story when I was below?'

• wa sásan 'dummy' pronominal wa + a derivation of the verb 'to beg' with a final n, a similar construction was found in the previous story with wa nníyən 'the one who is'. I think this confirms it that this is indeed the participle construction. 'the one who begs' > beggar.
• af-íwa 'why'
• tnəd-dík-ká 'did you not tell me'
• žlan-íya 'this story'
• u formally 'and', but in this context, should mean 'when' [bulbul:] "u" as "when" is interesting, could it be the same structure as the Arabic circumstantial clause?
• nək 'I'
• nníx 'to be' 1sg. same verb as previous story's participle nníyən. Found it in Paradisi now, under essere, trovare. Not sure why I didn't find it before.
• ar-úṭā 'below'

Baʕadén in-ís Žḥá: w-af-íwa tənnəd-(d)ík ə́rəš ar-úṭā?

'Then Žḥá said to him: and why did you tell me 'come down'

• ə́rəš 'to descend' imperative.

 

 

They boy in the red shirt

yəškí Žḥá s-ar (s-ġár) təfillí-nnəs w-ídd-əs iríw fəll-ís təkabə́rt təzwáġt

'Žḥá left his hous, and with him (his) boy who had a red shirt.'

• yəškí 'to leave' 3sg.m.pf.
• s-ar (s-ġár) 'from', again not s- is used but the compound preposition s-ġár, this time definitely not in front of a living object, so Lameen Souag's theory to compare it's use to Arabic min ʕand does not seem to hold up.
• təfillí-nnəs 'his house'
• w-ídd-əs 'and with him'
• iríw 'boy', I would expect a possesive suffix.
• fəll-ís 'on him' used to form possesive sentence. Similar to the way 'to have' is expressed in Japanese, Sanskrit and Arabic.
• təkabə́rt 'shirt'
• təzwáġt 'red'

u yərfaʕ-tíya af agərəṭ-ə́nnəs.

'and he carried him on his neck.'

• u 'and'
• yərfaʕ-tíya 'to carry him'
• af agərəṭ-ə́nnəs 'on his neck'

Baʕadén ittuwí-t u yəqqím idəwár fəll-ís u nəttín af agərəṭ-ə́nnəs.

'Then, he forgot it, and he continued looking for him, and he was (still) on his neck.'

• Baʕadén 'then'
• ittuwí-t 'he forgot it' 3sg.m.pf.
• yəqqím 'still be' literally 'to remain' used as an auxiliary here.
• idəwár 3sg.m.impf. [kato:] in ELA īdǝwwǝr 'to look for, to search'. īdūr is 'to revolve, turn, etc.' i wonder if yəqqím idəwár means something more like a past continuous, 'he was looking for it, he went about looking for it'. the construction would parallel ELA gǝ‘ad īdǝwwǝr 'he remained looking (for it), he looked about'.
• u nəttín 'and he'

u itənn-ís i-kúll amédən: dgulít-ká iríw fəll-ís təkabə́rt təzwáġt?

'and he was saying to every man: 'have yo not seen a boy who has a red shirt?''

• u itənn-ís 'and he was saying', interesting switch of tense here to the imperfect to express and itterative (maybe inchoative: and he started saying) meaning.
• i-kúll amédən 'to every man' < Ar. kull 'every'
• dgulít-ká 'have you not seen' 2sg.pf. + negative particle

Baʕadén in-ís iwín n amédən: 'bálək nəttín wa nníyən af əlkətf-ə́nnək?'

'Then a man said to him: 'perhaps he is the one that is on your shoulders?''

• in-ís 'he said to him'
• iwín n amédən 'a man'
• bálək 'perhaps' [kato:] bálək: 'maybe' in ELA. see also belkī 'id.' i think it's of Turkish origin. [Phoenix:] Looked it  up, indeed of Turkish origin belki 'maybe'.
• nəttín wa nníyən 'he who is the one that is'. wa functions as a sort of relative pronoun introducing the rest of the phrase nníyən should be the verb form here, but I can't find it in the dictionary, nor do I understand the form. Is this a participle? That would be the first time that I see one in Aujila.
• af əlkətf-ə́nnək 'on your shoulder'

Baʕadén yərfə́ʕ Žḥá təgilí-nnəs ar-ẓúṭ igúl iríw w-in-ís y-iríw:

'Then, Žḥá raised his head upward and saw the boy and said to the boy:'

• yərfə́ʕ 'to raise', this is the first time that we find this verb in its classical Arabic meaning 'to lift', so far it meant 'take away, take along, carry' etc.
• təgilí-nnəs 'his head' I'm tempted to wonder if this word is derived from the root GVL 'to see'. Making it sort of 'a thing you see with', which are your eyes and by extension your head.
• ar-ẓúṭ ar- = 'to' maybe a simplification of ġar 'at, on' + ẓúṭ 'above'
• igúl 'he saw'
• y-iríw 'to the boy' dative i- changed to y- in front of a vowel.

ur-náka-ká baʕád atqimát idd-í təʕaddít-ká?

'Didn't I tell you that when you are with me to not go (anywhere)?'

• ur-ná-ka-ká A typical northern Berber double negation! You do not see this much in Aujila Berber. ná-ka 1sg.pf. +2sg. indirect object.
• baʕád 'after', I would expect this word to mean 'earlier', 'before' in this context.  Would that be possible?
• atqimát 'to stay, remain, be at' 2sg.aor.
• idd-í 'with me'
• təʕaddít-ká 'to go' 2sg.pf. + neg