Berber-Tuaregs-Canarians.doc

Alexander Militarev

Libyo-Berbers – Tuaregs – Canarians
(Tamâhaq Tuaregs in the Canary Islands in the Context of Ethno-Linguistic
Prehistory of Libyo-Berbers: Linguistic and Inscriptional Evidence)[1]

There are a few striking lexical parallels between the extinct languages of
the Canary Islands aborigines and one of the Saharan Tuareg languages
belonging to the Tamâhaq dialect cluster (Ahaggar, Taitoq and some others).
Apart from the fact that all these Canarian idioms belong to the Berber-
Canarian[2] linguistic family (whose position within the
Afrasian/Afroasiatic macrofamily/phylum and reconstructed prehistory are
presented in the author's interpretation), and, hence, have an inherited
common lexicon, the Canary-Tamâhaq parallels feature common phonetic
developments which are so nontrivial and unique that the only explanation
for them is ethno-cultural contacts, namely, lexical borrowing from Tamâhaq
in Canarian. A series of such loan-words quoted and analyzed below gives
ground to the hypothesis of a Tamâhaq speaking Tuaregs' migration to the
Canary Islands, roughly datable between the 7th and 14th centuries CE.
Basing his ideas on this hypothesis, the author endeavors to decipher some
inscriptions in the Island of Ferro (Hierro) written in tifinaɣ, the only
one extant of the Libyan script varieties, with the help of the Ahaggar
dictionary demonstrating that they were composed in Tamâhaq Tuareg. Aside
from that, reconstruction is attempted of the linguistic and ethno-cultural
prehistory and history of the Berber, Tuareg and Canarian speakers.

1. The Canarian language group or dialect cluster is genetically
related to the Berber languages of North Africa. This relationship is
sustained by a series of morphologic, phonetic and lexical isoglosses, many
of which were registered and analyzed by the main figure of the Canarian
studies D. J. Wölfel and his precursors.[3] Whether it should be classified
as a group of closely related languages or a dialect cluster is not clear
to say nothing of a conventional distinction between the two taxonomic
ranks. On the one hand, forms of the same lexemes recorded in different
islands often fully coincide or feature close similarity. On the other, in
view of a presumed time distance between the first settlers of the islands
(dated by various authors from the 1st or even 2nd millennium BCE to the
turn of the eras) and the recording period as well as the notorious lack of
water vehicles[4] implying lack of contacts between the islands, the
difference between the island idioms is expected to have been much greater
than the one following from this similarity.
1.1. The Canarian and Berber linguistic unities are regarded by the
author as two taxonomically equal branches of the Berber-Canarian family
which also includes Epigraphic Libyan[5] and is, in its turn, part of the
Afrasian macrofamily.[6] The Berber-Canarian genetic dichotomy is
sustained, besides a few archaic morphologic features in Canarian lacking
in Berber and explained from common Afrasian patterns, by a number of
inherited Afrasian lexemes absent in Berber (see 1.1.1.) and presumed
loanwords in Proto-Canarian from various Afrasian and non-Afrasian
languages, unattested in Berber, which seem to have been very old
'continental' borrowings synchronous to the Proto-Berber period or, in
other words, prior to Proto-Berber splitting into dialects (see 1.1.2.)
dated to the very end of the 2nd millennium BCE.[7]
1.1.1. Examples of Canarian words likely inherited from Afrasian[8]
and not preserved in Berber:
1.1.1.1. Palma aganeye 'brazo cortado' [Wölfel, 416] ~ Afrasian *ginaʕ-
'arm, hand' : Chadic W.: Bokos kôŋ, Sha ngaan 'upper arm', E.: Tumak gēny
'hand'; Cushitic N.: Beja ganaʔa id., C.: Khamir ägäna 'palm of hand', E.:
Saho ginaʕ 'hand', Somali gaʕan (met.) 'arm, hand', Oromo gana 'palm of
hand', Warazi kanɛɛʔ, Yaaku kinnɛʔ 'hand' (k < *g); Omotic: Basketo ganná
'shoulder', etc.
1.1.1.2. Gran Canaria cuna 'chien', Tenerife cancha 'perros
pequeños', 'perro', 'little
dogs', cuncha 'chien' [Wölfel, 483-4], to be reconstructed as *kʷan-č/t- (-
č/-t probably a diminutive suffix) ~ Afrasian *kʷahin- 'dog' : Chadic W.:
Fyer kʷéeŋ, E.: Dangla kànyà, Migama kânnyà, Jegu kány, Birgit káyàŋ; E.
Cushitic: Yaaku kwehen; Omotic *ka/u(H)n- 'dog': N.: Gamu, Zala, Kafa,
Mocha kunano, Sheko keanu (<*keHan-), Mao (Hozo) kana, etc., S.: Dime kɛnɛ,
Galila kani. Cf. Nostr. *ḲüjnA 'wolf, dog'.
1.1.1.3. Gran Canaria lia 'soleil d'été', Lanzarote alio, Gomera,
Ferro lion 'sun' [Wölfel, 583] ~ Afrasian *lVʕ(lVʕ)-, var. *lVʔ(lVʔ)-
'light, luminary' : Semitic: Arabic lʔlʔ 'to shine', laʕlaʕ- 'mirage in the
afternoon heat'; W. Chadic: Kirfi lii-n 'dawn', Daffa-Butura leʔ-
'anbrechen (Tag)'; Cushitic: E.: Saho, Afar lāʕ 'be warm, hot', Saho laláʕ,
Afar laʕō 'day', Gawwada leʕayyo, Tsamai lɛʔo (ʔ < *ʔ) 'moon', Yaaku l'ɛɛʔ
(ʔ < *ʕ) 'moon', S.: Iraqw lôʕa 'sun, God'; S. Omotic: Ongota walʕ-ana
'star'.
1.1.1.4. Palma abora 'dios…que estaba en el cielo', 'god', 'Dieu, ou
le régulateur des astres' [Wölfel, 432]; two different etymologies can be
proposed:[9]
1.1.1.4a. Afrasian *bVr(ʔ)-: Semitic *bVrVʔ- 'to create': Hebrew
bārāʔ, Mandaic bra, Arabic brʔ; S. Omotic: Ari bérri 'Gottesname:
Erdgottheit', Hamar bary-o 'creator'.
1.1.1.4b. Afrasian *bVry- 'evil spirit, sorcerer': Semitic *barVy-:
Akkadian (Old Babylonian) bārû 'diviner', barīrī-tu, a female demon,
Mandaic baraia 'exorciser', Neo-Aramaic būrā 'fou, insensé ', Geʕez bāryā
'one who is in the service of a demon; epilepsy, a spirit that brings an
epilepsy'; Berber: Ahaggar būri 'attaque de nerfs', E. Tawllemmet bor-t-ən
(pl.) 'génies, esprits'; Chadic W.: Hausa bóorii, a form of supposed devil
possession, Tangale bụra, name of a bad spirit (cf. also Bokos mburú, Daffo-
Butura mbúrú 'medician'), C.: Mbara mbré 'sorcerer', E.: Mokilko bírre 'se
promener (mauvais esprit) pour manger les âmes des gens'; (?) E. Cushitic
*habār- (with a non-etymological h-): Saho abaar, Afar abaare 'curse',
abaaro 'evil, curse, a devilish act', Somali, Boni habaar, Oromo, Rendille
abaar- 'curse'.
1.1.1.5. Tenerife achaño 'año', Gran Canaria acano 'el año lunar'
[Wölfel, 480] (if *a-san-, not *a-kan) ~ Afrasian *san- 'year': Semitic
*šan-at- 'year'; W. Chadic *wa-sun- 'year; rainy season'; E. Cushitic:
Sidamo sané 'year', etc.
1.1.1.6. Palma hirguan '…demonio in figura d'uomo lanuto', irvene
'el demonio en figura de perro lanudo', irvene 'apparitions', yrvene 'el
Diablo' [Wölfel, 484, entry "Der Hund"] – if, in fact, meaning 'dog' and
reconstructed as *hirgʷan-,[10] cf. the following series of metathetic
Afrasian terms: Semitic: Syrian harnāgā 'golden jackal' (cf. also Arabic
ǯāriḥat- 'bête ou oiseau de chasse (chien, guépard, faucon)'); Chadic C.:
Hildi gərhaŋ 'jackal', Musgu gúrnai, ŋúrnai, Masa ŋurnaita, Banana gɔnira
'hyena', E.: Kera gòrnòy id.
1.1.1.7. Tenerife haña, jana 'la oveja', ana 'sheep', 'mouton',
'carnero' <*Han(y)a [Wölfel, 485] ~ Afrasian *(ʔa)w/yVn-/*(ʔa)nay- 'k. of
small bovid' : Semitic *ʔawn-: Akkadian unû 'k. of meat' (?), Gurage: Chaha
onā, Eža onnā 'young male goat or sheep'; Egyptian (New Kingdom) wny 'calf
(as a representation of Osiris)'; Chadic W.: Siri yáàní 'she-goat', E.
*ʔaynV: Migama ĩ:ná, Jegu te-ené (pl. ʔeen), Birgit ʔayney 'goat'; Cushitic
N.: Beja ano 'sheep', E.: Afar anaʕ-to 'lamb (female)' (-ʕ is non-
etymological), Somali wan, pl. wanan 'ram', Rendille onó 'sheep', Burji
ayáan-e 'gazelle'; S. Omotic: Dime iin, (?) Ongota hoona (h- is non-
etymological) 'sheep'.
1.1.1.8. Tenerife afaro 'grano', 'grain', ofaro 'grain' [Wölfel, 502] ~
Afrasian *pi/ar- 'fruit, corn, seed': Semitic *piray- 'fruit; sprout;
offspring, flower' (including Harari fəri 'seed'); Egyptian (Old Kingdom)
pr.t 'fruit, crop, seed'; W.Chadic: Hausa fár-fárā́ 'k. of guinea corn';
Cushitic N.: Beja far 'seed' (also 'flower, blossom, bloom'), C.: Qemant
fir 'corn' ('fruit', in other Agaw languages), E.: Saho fire 'flowers,
fruit', Oromo firi 'fruit'.
1.1.1.9. Tenerife beñesmer '…sementera…en el mes de Agosto', venesmer
'the month of August', benismer 'le mois d'août', benismen 'la saison des
récoltes', benesmen[11] 'la situación de las cosechas' [Wölfel, 481]
(convincingly reconstructed as *we-n-esmer 'der (Zeitraum) der Ernte'
[ibid., 482], i.e. 'that of harvest') ~ Afrasian (or only Semitic) *čVmVr-:
Semitic *ṯmr 'to ripen, produce a (good) crop': Arabic ṯ̲mr 'fructifier,
donner des fruits (arbre)', Sabaic ṯmr 'to produce crops (land); crops
(spec. cereals)', Geʕez samra 'to flourish, be fruitful', Tigrai sämärä,
Amharic sämmärä 'to be a bumper crop', Soddo asmara 'season for
agricultural activities, harvest', Mehri ṯmər, Jibbali ṯǖĩr 'to come to
fruition, ripen', Mehri šə-ṯmūr 'to have a good crop', Jibbali sə-ṯmír 'to
get a good harvest (from a palm-tree)', Soqotri tmr 'to bear fruit';
(Cushitic C.: Khamir asmárā 'Herbst, Erntezeit', E.: Oromo asmara 'season
for agricultural activities, harvest' are likely Ethiopian loanwords),
perhaps, also E.Chadic: Lele sómrō `dry season'.
1.1.2. A few examples of presumed loanwords in Canarian absent in
Berber:
1.1.2.1. Gomera, Ferro jubaques, juvaque 'ovejas gordas', tabaque 'reses
gordas, fubaque 'a fat sheep', juhaques 'brebis grasses', etc. [Wölfel, 489-
90]; the consonantal variants j-/t-/f- and -b-/-v-/-h- together with the
clear vocalic pattern -u-…-a- reflect the efforts to transcribe the
original word sounding like [SubaK],[12] looking very much like a loanword
from E.Cushitic: Saho, Afar subaḥ 'butter', Somali subag (with irregular -g
instead of the expected ḥ) 'clarified butter or animal fat', Rendille
subaḥ 'butter; fat, clarified animal fat' (also Jiddu sooh 'fat', Baiso
suba 'butter', Elmolo sííḅiʔ 'fat') < Afrasian *ĉVbVḥ- 'fat': Semitic: Geez
ŝəbḥ, Tigre šəbeḥ, Tigrai səbḥ-i, Amharic säb 'fat' (n.), Harari säbaḥ
'animal fat', Mehri ŝabḥ 'fat' (n.), Jibbali ŝabḥ 'fat on meat'; W. Chadic:
Diri ŝəḅu 'fat' (adj.)'
1.1.2.2. Gran Canaria atazaykate 'gran cuore', Atacaycate, athacaite
'gran corazon', atacayte 'stout heart', atacaicate 'le grand coeur',
altacaite 'el valiente', altaycayte 'el valiente', 'le brave', etc. (in
Wölfel, 413-14, compared with Hausa 'heart'), to be most probably
reconstructed as *a(l)-ta-Zaykay-te with the article-like a- prefix[13] and
ta…ta confix occurring in Canarian and highly characteristic of Berber ~ W.
Chadic: Hausa zūčìyā, pl. zukata, zukoči 'heart' (zùčiyà 'get into a
temper') with a Chadic etymology (W.: Diri aǯukwa 'heart', C.: Mafa me-zek
'soul') and, perhaps, an Afrasian one; the likeness of the Hausa plural
form to the Gran Canaria word is too striking to be haphazard.[14]
1.1.2.3. There are two parallels with Hebrew, most likely lookalikes,
however, worth mentioning:
1.1.2.3a. Gran Canaria belingo 'diversion, fiesta, jolgorio' [Wölfel,
483] ~ Hebrew blg (hif.) 'become cheerful, rejoice, brighten up' (also
'cause to flash', Arabic blǯ 'be cheerful, happy; to break (dawn)' <
Afrasian *bVlVg- 'shine': E. Cushitic: Oromo balag 'shine', Somali bilig
'sparkling'; N. Omotic: Mao (Bambeshi) bɛ̀lgɛ́ 'star'.
1.1.2.3b. Tenerife chacerquem '…la miel', chacerquen 'miel', 'honey, or
melasses made of mocanes',[15] Tenerife and Gran Canaria tacerquen 'cerveza
ó vino', Gran Canaria azarquen 'arrope de mocanes' [Wölfel, 525] ((all
<*(t)a-SVrḳ-Vn) ~ Hebrew ŝōrēḳ, a valued, bright-red species of grape' <
Semitic *ŝrḳ 'be red'.
1.2. It will be relevant here to make a brief survey into the author's
reconstruction of the Berber-Canarian ethno-linguistic prehistory.
1.2.1. The split of Proto-Afrasian in the early 11th mill. BCE,
synchronous with the Younger Dryas end and climatic amelioration in South-
West Asia, into South Afrasian (Proto-Cushitic-Omotic) and North Afrasian
(Proto-Semitic-Egyptian-Berber-Chadic) presumably took place in the Levant,
where I identify the Proto-Afrasian speakers as the creators of the
Natufian and Post-Natufian archaeological cultures.[16]
1.2.2. The Proto-North Afrasian split (ca. 9,300 BCE) into Proto-Semitic
and Proto-North African Afrasian (Proto-Egyptian-Berber-Chadic), whose
speakers at some point migrated to North Africa; they must already have had
farming experience judging by a number of cognate agricultural terms their
languages share with Semitic. Therefore, they could hardly migrate to
Africa prior to the beginning of North African Neolithic revolution (ca.
6,000 BCE.), which obviously was initiated by them. By that time,
separation between Proto-Egyptian and Proto-Berber-Chadic speakers dated to
ca. 7,800 BCE had already taken place implying their synchronous migration
(or two closely spaced migration waves) to Africa marked with the
appearance of Neolithic settlements in Egypt (proto-Egyptian speakers), the
replacement of the Mesolithic Capsian culture in the Maghreb by Neolithic
cultures around 6,000 BCE, coinciding with the glottochronologically dated
split between Proto-Berber-Canarian and Proto-Chadic, and their further
spread throughout the Sahara (proto-Berber-Canarian speakers)[17] and to
the Sahel (likely proto-Chadic speakers) starting from the 6th millennium.
1.2.3. It would be tempting to presume the migration route to the
African shore via Cyprus with its population's Y-chromosome haplogroups
showing strong Anatolian and Levantine influence[18] and its Neolithic
culture (e.g., round houses constructed in mudbrick) and subsistence
system: the insipient agriculture and domestication of the same animals as
in the Levant: dogs, goats, sheep, cattle, pigs (plus the earliest known
domestication of cats) equal to Post-Natufian Pre-Pottery Neolithic (late
PPNA, EPPNB).[19] However, Cyprus could have been only a relatively short-
time "traffic terminal": staying there for centuries would have made the
migrant group(s) lose ("forget") part of their Common Afrasian lexicon
conveying certain realia of the Levant (in the first place, names of
predators absent in the island, but retained in the new African
environment) by the time of their settlement in Africa. There is also a
remote possibility of their maintaining constant contacts with the Levant
and its realia. Anyway, the coincidence between both the island settlement
and divergence of North Afrasian timing, the abrupt end of the Neolithic
aceramic civilization of Cyprus around 6,000 BCE and the synchronous rise
of North African Neolithic revolution are intriguing.
1.2.4. After the split of Proto-Berber-Canarian, impossible to determine
glottochronologically (only a few Canarian words from the diagnostic list
were recorded) but evidently prior to the Proto-Berber split, both the
Proto-Canarian and Proto-Berber speaking groups seem to have stayed for
some more time in the vicinity of Egypt, which would account for the
following cases:
1.2.4.1. Tenerife axo '[el cadáver] seco y miriado, que llamaban ~',
xaxo, haho 'cuerpo muerto', 'momia' [Wölfel, 462] to be reconstructed as
*Haḫw strikingly resembling Egyptian (from Old Kingdom on) ꜣḫw 'spirit;
deceased',[20] very likely a source of the Tenerife term raising the
possibility of close cultural contacts of Proto-Canarian speakers with
Egypt and indicative of the Egyptian origin of mummification in the Canary
Islands.[21]
1.2.4.2. Proto-Berber *Hirab̲ 'letter, message' and *Harab̲ 'to write',
Gran Canaria tara, tarja 'rayas en tablas, pared o piedras', 'señal para
recuerdos', tarha 'marque pour les souvenirs', 'señal para recuerdos' ~
Egyptian (Pyramid texts) hꜣb (< *hrb), to send (a letter or message inter
alia), to write a letter, (Middle Kingdom-New Egyptian), letter, message'
(for further details see 3.4.3.).
1.2.4.3. Proto-Berber *ta-matar-t 'sign of acknowledgement,
recognition, testimony, etc.' – a well adopted (with a specifically Berber
ta-…-(V)t confix) borrowing from Egyptian (Middle Kingdom) mtr 'testimony,
witness, instruction, etc.'[22]
1.2.4.4. Berb. *sVlm-ay, pl. *(i/a-)salm-an 'fish' (the pan-Berber
generic term): Ahaggar a-sûlmey, pl. i-sûlmäy-än, Snus, Semlal, Izdeg,
Qabyle a-sləm, etc. ~ Canarian (various islands) salema 'Sparus
Cantharus', 'pez de la clase de los torácicos' ([Wölfel, 563] compares it
with the Snus form only mentioning that it is common Berber, and with Latin
salmo-onis). The Latin term was obviously borrowed from the Berber pl. of
the *(i/a-)salm-an type.[23]
However, there are two more metathetic forms in Berber: Ghat a-ləmšay
and Ghadames u-lisma. The latter by its order of root consonants especially
resembles Egyptian (New Kingdom) nšmw.t (with the hieroglyph for n
conveying both n and l, in this case < *lVšm-Vw-t) 'k. of fish'.[24] This
Berber-Egyptian term might have been taken for a common heritage from
Afrasian if not for the fact that the New Egyptian nšmw.t evolved from the
meaning 'scales of fish' attested as early as in the Middle Kingdom period
(and, hence, having the "native" Egyptian etymology) is identifiable as the
source of the Berber term to be interpreted as *lVsm- or, rather, *lVŝm-
with the relative adjectives suffix *-ay, probably reflecting the original
Egyptian meaning 'scaly'. The fact of borrowing the New Kingdom term speaks
for the presence of Proto-Berber (or Proto-Berber-Canarian) speaking
population in the vicinity of Egypt in the second half of the 2nd mill.
BCE. I have little doubt that the Latin and Celtic terms for 'salmon' (note
its playing such an important role in the Celts' mythology) were borrowed
from the Libyo-Berber plural form *(i/a-)salm-an.
1.2.4.5. Egyptian mšwš, name of a Libyan tribe in the late 13th
century BCE texts likely from Proto-Berber *a-maziɣ, a "Berber" man (see
3.2); if valid, this identification implies the pre-Proto-Berber (a period
predating the disintegration of Proto-Berber into early dialects) *-maĉiɣ
<*-maĉiḫ[25] with *ĉ regularly corresponding to Egyptian š- (in the Inlaut)
while the Auslaut -š (<*ḫ) vs. Berber -ɣ could be due to assimilation.
1.2.4.6. Egyptian ꜣbyḳwr, a greyhound (and the name of a Libyan
chief) in a Middle Kingdom (11th Dynasty - 2118-2069 BCE) text matching a
common Tuareg term *abaykur 'dog of bad race, cur' but also
'greyhound':[26] Ahaggar abaykôr, Ayr, E.Tawllemmet abǎykor, Adghaq abəykor
'chien de mauvaise race', Ghat abaykur 'lévrier' (I analyze it as *ab-
aykar, a composed word, something like 'non-dog, under-dog': *a/iba,
negation, 'ne pas y avoir de', etc., and *a-ykar 'young dog').[27]
1.2.5. Relying on these significant if scanty lexical evidence I
conjecture that the split of Proto-Berber-Canarian[28] happened prior to or
during the last period of the New Kingdom, probably due to the defeats
inflicted by Ramesses III in the 1170s BCE on the coalition of Libyans and
the Sea Peoples. This coalition seems to have left a couple of lexical
traces in Proto-Berber, perhaps, hinting at the Hurrians' presence among
the Sea Peoples. They are:[29]
1.2.5.1. Proto-Berber *a/isi, pl. iss-an 'horse' < Hurrian es̲s̲i
'horse'.[30]
1.2.5.2. Proto-Berber *(a/i-)birgan 'tent' (including Zenaga tə-burgən
'tente dressée sur des piquets') unequivocably from Hurrian-Urartian
*burgana 'fort' (note the meaning shift).
1.2.6. Following the split of Proto-Berber falling on the late II
mill. BCE,[31] part of Libyan tribes likely started migrating westward
occupying the Mediterranean coastal areas while another part (Proto-East
Berber speakers) settled — or stayed — in the Libyan desert oases. This is
also the period when Garamantes whom I, like some other authors, identify
with Proto-Tuaregs emerge on the historical arena, first in the Fezzan.
1.2.7. Besides the famous Garamantes' westward trade routes (see, for
instance, [Trans-Saharan Trade, Wikipedia]) marked with numerous paintings
and engravings during the so-called "Horse Period" dated roughly from the
late 2nd mill. BCE on, of horse-drawn chariots, there are lexical evidence
for another, south-eastward route to Nubia[32] and even Ethiopia.[33]
1.2.8. Returning to the Canary Islands: if until recently, there has
been no archaeological traces of human presence prior to the turn of the
eras, today we have evidence of human activity from Lanzarote and the
adjacent small island of La Graciosa dated to the 11th-10th centuries
BCE.[34] In Lanzarote, "organic remains associated to Phoeno-Punic
materials (pottery, copper and bronze objects and a glass seed) have been
found."[35]
2. There are certain facts which at first glance seem to contradict a
postulated dichotomy within Berber-Canarian family. These are separate
Canarian—non-Tuareg Berber and Canarian—Tuareg lexical and phonetic
isoglosses. Whereas the former ones are but a few cultural terms not
relevant for our main subject (cf. point 4), the latter share certain
phonetic peculiarities proper, of all the Berber subgroups, to the Tuareg
one only, and thus need to be explained.
2.1. Partly, these Canarian-Tuareg isoglosses are accounted for by a
reputed conservatism of Tuareg languages owing to their relative isolation
from an all-penetrating impact of Arabic. They have retained so many common
Berber features lost by other Berber languages, that the late Prof. K.-G.
Prasse, a leading world expert in Tuareg languages, even averred that the
Proto-Tuareg reconstruction should not differ in principle from the Proto-
Berber one (the opinion I can accept only to a point).
3. Apart from common Canarian-Berber and separate Canarian-Tuareg
lexical items, there is a series of isoglosses linking various Canarian
idioms with one North Tuareg dialect cluster known as Tamâhaq.
3.1. That could be, and partially is, explained by a simple fact that,
thanks to the first-class most comprehensive dictionary by Ch. de
Foucauld[36], as far as I know, unparalleled in African lexicography, a
student may operate with a huge amount of words of the most important
Tamâhaq dialect, that of Ahaggar (Tahaggart). Since the number of Ahaggar
words available surpasses by far that of any other Berber language, the
probability of chance coincidence with the Ahaggar lexemes is the highest
for every possible language drawn to comparison.
3.2. There is a series of Canarian and Ahaggar words, though,
revealing such a striking phonetic affinity that it cannot be elucidated by
the above reason. The affinity in question concerns, above all, a phonetic
phenomenon quite unmotivated in Canarian, but historically well explained
in Ahaggar. What is meant here can be clearly illustrated by the word used
by most Berber-speaking communities to designate their mother tongue. In
non-Tuareg Berber, it is ta-maziɣ-t (from a-maziɣ 'a Berber', pl. i-maziɣ-
ən); in various Tuareg languages, tǎ-mašǎq, tǎ-mažǎq, tǎ-maziq; and in
Ahaggar and several closely related dialects, ta-mâhaq in Foucauld's or tǎ-
mahǎq in Prasse's notation.[37] The phonetic correspondences based on this
comparison are as follows: non-Tuareg Berber z ~ Tuareg (except Ahaggar and
the other Tamâhaq dialects) š - ž - z - ǯ ~ Ahaggar (and other Tamâhaq
dialects) h.
3.2.1. The corresponding Proto-Berber phoneme is reconstructed as *z.
This consonant usually (but not always – see 3.2.2) yields a voiced
sibilant z in Canarian, corresponding to several Proto-Afrasian affricates:
*ʒ/*c/*ĉ/*ǯ [Prasse, #520].[38] Examples:
3.2.1.1. Lanzarote, Fuerteventura azeca 'muraille', 'muralla' [Wölfel,
558] ~ Berber *t-ā-zaqqāw: Ghadames ta-zəqqa, pl. ta-zəqw-an 'wall', Snus a-
zəqqa 'house', Mzab ta-ẓəqqa 'fence' (ẓ < *z assimilated to q) ~ Ahaggar tǎ-
haqqa, pl. ti-həɣ-în 'storehouse for food supplies' (part of the forms
quoted are compared [ibid., #558]).
3.2.1.2. Palma zeloy 'soleil', 'el sol' [Wölfel, 583-584] ~ Berber *ā-
zayl: Siwa a-zəl 'daytime', Fodjaha azal, pl. a-zalîẉ-ən, Ayr ̣ə-žil, Semlal
a-zal 'sunlight', Izayan a-zil 'heat of the day', Qabyle a-zal, pl. izil-an
'daylight' ~ Ahaggar a-həl 'sunlight' (part of the Berber forms are
compared [ibid.]; for the Tuareg forms see [Prasse, #219].
3.2.1.3. Palma azuquahe 'moreno a negro', azuquache 'moreno',
azaquache 'brun', azuguanche 'le brun', asuquahe 'black or duskey', etc.
[Wölfel, 425-426] ~ Berber *ā-zVwwāɣ 'red':[39] Siwa a-ẓəgg̣əɣ, (-gg- <
*ww), Audjila a-zwaɣ, Ayr a-mi-zwǎɣ, Tadghaq šəggaɣ, Izdeg a-zuggwaɣ,
Djerba a-zuggaɣ, Qabyle a-zəggʷaɣ ~ Ahaggar ihwaɣ 'to be red' (part of the
Berber forms compared [ibid.]; cf. [Prasse, #435]).
3.2.1.4. Tenerife mencei 'nove regni', 'amparo y defensa', menzei 'rey
(i.e. defensa)', mencey 'el rey, rey, soberano', 'le seigneur ou le roi',
etc. [Wölfel, 465-466] ~ Berber *ā-manzuy 'first; firstborn, one who comes
the first' (participle < *nazVy 'to come or go early in the morning, to
come or go the first'; compared with a question mark in [Prasse, #554]):
Ghadames a-mənzu, mənzay 'beginning, start', Rif, Senwa a-mənzu 'first-
born, aged, eldest', Nefusa a-mənzu 'the first, first-born, the eldest',
Qabyle a-mənzu. pl. i-mənza 'one who comes the first, first-born' ~ Ahaggar
e-məñhi 'forerunner, herald' (non-Tuareg forms compared in [Wölfel, 466];
for the Ahaggar and Qabyle ones see [Prasse, #534]).
3.2.2. Besides the Canarian words of the above group with *z
(transliterated as z, sometimes as s between vowels, occasionaly as c or ch
possibly rendering *ʒ, an affricate counterpart of *z) corresponding to
Berber *z, there is a group of Canarian words with *h[40] corresponding to
Ahaggar h, the one[41] proceeding from Proto-Berber *z as described in
point 3.2.1. Examples:
3.2.2.1. Lanzarote hyguyeres (-es appears to be a French plural
ending) 'Euphorbia Canariensis', 'une maniere de bois qui ... ne puct
ardoiren nulle maniere jusques atant qu'il est secq et pouris' (in [Wölfel,
572] compared with Ahaggar eseɣir 'morceau de bois sec', possibly related
to Berber *iɣar/ *yəqqur 'to be dry' but hardly to the Lanzarote term) ~
Ahaggar ǎ-həqqor 'beam made of a palm-tree' ~ Zekkara a-zaqqur 'tree', Snus
ṯa-zəqqur-ṯ 'firewood', Qabyle a-zəqqur 'trunk of a dry tree' (< Berber *ā-
zaqqūr ; cf. [Prasse, #349]).
3.2.2.2. Ferro apio 'fontana', gapio 'fuente', hapio 'fountain', gapo
'fuente' [Wölfel, 590] (no Berber parallels suggested) ~ Ahaggar tǎ-haf-
t[42] 'canal d'arrosage' ~ Ghat ta-zəf-t id., pl. či-zzif [Prasse, #163].
3.2.2.3. Gran Canaria taharenemen 'higos pasos', 'figues seches'
[Wölfel, 504]. This is definitely a composed word most likely to be
analyzed as *tahar-ən-əmən) ~ Ahaggar âhâr, Taitoq ahar 'figue', Ahaggar t-
âhâr-t, pl. t-âhâr-în 'figuier' ~ Ghat azar (Ayr t-ahar-t, Tadghaq a-har
are borrowed from Ahaggar, which explains -h- in place of the expected -z-)
'figue', Ghadames t-azar-t, pl. t-azar-in, Qabyle t-azar-t 'dry figs'
[Prasse, #354] (part of the forms including the Ahaggar one compared in
[Wölfel, 504-505]; other Gran Canaria forms like tahaunenen [ibid.] are
either misheard by the recorders or have a different etymology). As for
Berber parallels to the second part of the Canarian composed word, cp.
Tuareg: Ayr, E. Tawllemmet mǎnna 'sécheresse' [Alojaly, 130], Ahaggar menna
'sécheresse prolongué' [Foucauld, 1206]. I suggest the interpretation of
taharenemen as *t-ahar-(ə)n-(ə)man(na) meaning 'fig of drought'.[43]
3.2.2.4. Gomera tahuyan (to analyze as ta-huy-an, pl.) '… faldetas
de... pieles pintadas', 'enaguas de pieles', 'basquiñas', 'jupe',
'pettycoats of goat skins', (Palma) tahuy [ta-huy] 'piel' [Wölfel, 530]
(compared with the Ahaggar form) ~ Ahaggar te-hayhay-t (< *zayhay-) 'sac
en peau à longues franges' ~ E. Tawllemmet a-šǎyha 'sac en peau special
(pour les vétements)' [Alojali, 186]; for different parallels to the
Ahaggar form cf. [Prasse, #450].[44]
3.2.2.5. Lanzarote, Fuerteventura maho 'a shoe', 'calzado', mahos (to
analyze as maho-s, with a Spanish plural ending) 'calzados de los cueros de
las cabras, el pelo afuera, unos como zapatos', maxo 'zapata', 'bottines',
majo 'el calzado', maohs (maoh-s) 'un pezzo di corame di capra avvolto à i
piedi' ([Wölfel, 530-531; compared with the Ahaggar term) ~ Ahaggar ta-mhi-
t (pl. ti-mhî-t-în) 'sac en peau de chèvre' ~ Tadghaq ta-msi-t id. [Prasse,
#515].[45]
3.2.2.6. Palma ahuar [a-hwar] 'tierra' [Wölfel, 600], benahorare
'patria, mi patria, mi tierra', benahoare 'rni patria... mi tierra',
benehoare 'the name of the Island [of Palma]', banahore 'mi patria… mi
tierra'[46] (< *ben ahʷar[ar]-e < *wə/a n a-hʷar-i 'that of my land': cp.
Ahaggar wa n 'celui de' [Foucauld, 1448-1450] and -i, -e affixed pronoun
'de moi' [ibid., 685]) ~ Ahaggar ǎ-haggar 'central part of the plateau in
the country of Kel-Ahaggar; the Tuareg inhabitants of this country' < *ā-
hawwār). This term borrowed in other Tuareg languages (e.g. in Ayr in the
form ahaggar with -h- instead of the expected -z-) is identical with the
ancient Libyan-Berber tribe name Hawwara cited by medieval Arabic
geographers, which means that a-haggar is originally an ethnonym likely
referring to speakers of what can be called Proto-Tamâhaq rather that a
toponym. In its turn, it seems to be derived from the verb əhwər 'to
precede, to go before, to be older than' from common Berber *əzwər 'to
precede, to march ahead, to be the first': Ghadames ezwər, Ayr əzwər,
Qabyle zwir, etc.; cf. also [Prasse, #437]. The primary meaning of ǎ-haggar
would be then something like 'the one used to precede, to be the first'
from hâggər, a habitative form of əhwər, which fully corresponds to the
traditionally high position of the Kel-Ahaggar people among the
neighbouring tribes.
3.2.3. The only explanation I can give to this phenomenon is lexical
borrowing from Ahaggar or from its ancestor language (Proto-Tamâhaq) in
Canarian, whatever historical, cultural or other implications.[47]
3.2.3.1. The time-span, within which the contacts between the Tamâhaq-
speaking Saharan migrants and Canarian aborigines took place, can be
roughly estimated on the following basis. Since Ahaggar is a h-Ianguage
opposed to the non-Tamâhaq Tuareg š-, z-, ž- languages (see 3.2.), the time
of separation of Ahaggar resp. Tamâhaq from the rest of Tuareg would be a
terminus ante quem non of a presumed Tamâhaq migration to the Canary
Islands. This date has been obtained by the author through
glottochronological calculations of 40 Berber languages (see in Appendix 2
a representative selection of 17 languages) resulting in ca. the 6th-7th
centuries AD.[48] The terminus post quem non can be easily made up. This is
naturally the beginning of the European colonization of the Canary Islands,
after which a migration from Sahara could have hardly remained unnoticed
and unrecorded. However, because there are no traces of Islam in the
Archipelago, the Tuareg migration must have taken place before their
islamization.
3.2.4. In Ahaggar, beside the three series of regular correspondences:
h < *z, h < *h and h <*b̲ or *hʷ (i.e. merger of the three h of different
origin as described in 3.2.1. and 3.2.2.), there is one more process of a
purely morphonological character: the old z is retained when reduplicated
in certain verbal forms, but yields h when not, e.g. ighal (imperfective)
versus gəzzul (perfective) 'to be short' (< *əgzəl). It may be surmised
that this peculiar process is the remnant of the last stage of the *z > h
transition "utilized" by the language to form a morphological opposition.
If this hypothesis holds water, at least two cases of a free alternation of
h and z can be revealed in our presumed "proto-Ahaggarisms" in Canarian,
then pointing exactly to the period of separation of Ahaggar (or proto-
Tamâhaq, whose other daughter dialects are not sufficiently studied to know
if they feature the same process) from non-Tamâhaq Tuareg idioms. They are:
3.2.4.1. Canarian (Ferro) eraoranhan 'un Idolo maschio',[49]
eranoranhan, erahoranhan, eraoranzan 'God of the people of Hierro', 'male
Deity', 'le Dieu des Hommes', orahan 'un Dio', orojan, oranjan 'Dios'
([Wölfel, 437]; no Berber parallels adduced). This is to be analyzed as era
'he who', in one instance followed by the nota genitivi n (eranoranhan,
i.e. era n oranhan), plus *[h]ora[n]h/zan meaning 'recompensing, requiting,
giving (back)', n before h/z probably the result of a secondary positional
development. The first element, era, corresponds to Ahaggar ere 'celui
que', which, when a subject in a sentence, requires a participle after
it[50] represented here by the second element, orahan (oranhan, oranzan,
etc.) exactly corresponding to Ahaggar participle yorəh-ən derived from
arəh 'donner en retour'.[51] This Ahaggar verb is related to Ayr, E.
Tawllemmet ərəz, Tadghaq irəš id. [Alojali, 166; Prasse, #596], the Proto-
Tuareg form being *āraz.
3.2.4.2. Canarian (Tenerife) añepa 'la asta que el rey llevaba del
ante de sí', 'unas varas tostadas de tea y sabinas muy agudas', 'lanza de
tea, que precedía al rey', 'gran lanza de tea fina, con una banderilla de
juncos al extreme', 'lanza ó guión real',[52] anepa 'la lanza que el rey
llevaba delante', 'a Scepter or Spear', 'étendard du roi', (sic!) anzpa
'... una pertica sottile... ben lavorata, la quale era il segno reale'
([Wölfel, 477-478] with the Ahaggar parallel) ~ Ahaggar ǎ-ñhəf 'batôn gros
et long'.
This is the most striking case. As mentioned above, in Ahaggar h, as
shown by K.-G. Prasse [p. 5 seq.], three common Berber consonants have
merged. Besides h < *z (see point 3.2.), this Ahaggar laryngeal may reflect
the old *h yielding h in other Tuareg languages and usually no consonant in
non-Tuareg Berber, and the old *hʷ (or *b̲) yielding h Tuareg and b̲ in
Ghadames and Audjila of the Eastern Berber group. In Ahaggar, the nasal n
gets palatalized before h < *z, but does not before h continuing the two
other protophonemes, e.g., Ahaggar a-nhel 'autruche' (a-nhil, a-nǎhil, etc.
in other Tuareg languages) [Prasse, #544] pointing to h < *h, but Ahaggar
te-ñhǎr-t 'narine/nez' (~ Ghat ta-nzər-t, Ayr čə-nžar-t) pointing to h < *z
[ibid., #548]. Thus, in Ahaggar, the development is as follows: *nz > (*nh,
an intermediate stage) > ñh. Hence, Ahaggar ǎ-ñhəf is from *a-nzəf
corresponding to Ghat a-nžəf 'tison' [ibid., #541], while Canarian añepa is
from *a-ñhepa with a variant form anzpa representing one more instance of
the alternation h/z.[53] Taking into account the semantic affinity of the
Ahaggar and Canarian terms',[54] Wölfel was quite right to regard them as
related:[55] the development *nz > ñh in Ahaggar is so unique that a
similar case in any non-Tamâhaq language, Canarian in our case, cannot be
explained as an independent parallel development. Therefore, the only
plausible interpretation would be a borrowing of the Canarian word from
(Proto-)Ahaggar.
3.3. There is also a series of Canarian words containing *h which
correspond to Ahaggar words with h reflecting Common Berber *h. These
examples can be interpreted, on a fifty-fifty basis, either as inherited
Tuareg-Canarian isoglosses (not borrowings) reflecting Common Berber-
Canarian *h (see point 2.) or as more Tamâhaq/Ahaggar loans in Canarian. A
few examples:
3.3.1. Canarian (Gran Canaria) fayahuracan 'capitan', fayahuracanes
'...como capitanes, eran caudillos en la guerra de mucho respeto, que se
elegían por su nobleza, fuerzas y destreza para el ejercicio de su empleo,
obedeciendo los vecinos de sus pueblos a su llamamiento, y de sus
bocinas...'. [Wölfel, 470]. This is a composed word, the first element of
which is attested in Gran Canaria as faya 'hombre poderoso', 'reste ó
mencey de aquella parte'[56] ([Wölfel, 469]; no Berber parallels suggested)
comparable with Ahaggar ufu, Ayr afu (reciprocal nə-yufu) [Alojali, 36],
Qabyle if, etc. 'to be best, better than', whereas the second, huracan, is
from *hurak-an (-es in huracanes is a Spanish plural ending) with the
participle marker -an, to be compared with Ahaggar hərəkk-ət 'respecter'
[Foucauld, 654] (Ayr, E. Tawllemmet hərəkk-ət 'respecter, avoir peur de'
[Alojali, 80]). Therefore, the meaning of fayahuracan can be surmised to
be somewhat like 'best (and) respected'.
3.3.2. Canarian (Gran Canaria) guaires 'capitanes de los más
esforzados y valientes', guaire 'el noble', guayre 'el consejero', guayres
'consejeros de guerra' (also cuayres), gayres 'poderosos', 'consejeros de
la guerra', 'the members of the Privy Council', gayre 'noble', etc.
[Wölfel, 470-471].[57] This can be reconstructed as *gʷayre, possibly
reflecting *hʷayre and as such is comparable with Ahaggar t-ihôrar 'fait
d'être très respecté' [Prasse, #390] ~ Ayr, E. Tawllemmet ihar, perfective
yǎhor, ihor 'mériter' [Alojali, 79] (< Common Tuareg *ihwar or *ihʷar).
3.3.3. Canarian (Gomera, Ferro) aala 'eau', 'agua' ~ Ahaggar tǎ-hala
'petite source', Snus t-ala 'mare alimentée par une source', Qabyle ṯ-ala
'source', etc. – all quoted in [Wölfel, 513-514]. The double a in the
Canarian form serves to convey either a long vowel, or, more likely, a
laryngeal in between, i.e. the word should have actually sounded like
[ahala].
3.4. Finally, there is a series of Canarian examples with *h
corresponding to Ahaggar h or 0 (no consonant) and to Ghadames and Audjila
b̲, thus reflecting common Berber *hʷ ( according to K.-G.
Prasse, *h2 or *h3) or *b̲, which may have been a positional variant of *b
in the vicinity of a laryngeal; there are comparative data, if scarce, from
other Afrasian languages in favor of both hypotheses. Such cases, just as
in point 3.3., can be interpreted either as inherited common Berber-
Canarian terms or as (Proto-)Ahaggar borrowings in Canarian; the second
explanation is more valid at least in one case (see 3.4.3.).
3.4.1. Canarian (unspecified dialect) güiro 'Las relaciones entre los
dos sexos...,tienden a rodearse de misterio y de reserva, ...Hay quien
tiene a orgullo el poner de manifiesto los güiros de sus contemporáneos
...' Wölfel reads this word as [gʷiro] (the real pronunciation can be
[hʷiro] or [wiro] as well, considering Spanish transliteration) comparing
it with Ahaggar ər 'aimer, vouloir' and Ghadames eb̲ri [Wölfel, 410-411].
The Proto-Berber form would be, in Prof. Prasse's opinion (personal
communication), *îrHîh, perfective *yǔrHǎh, obtained on the basis of
Ahaggar, Ayr ǎr, E. Tawllemmet iru, perfective yəra, yǎra, Tadghaq ǎrh,
intensive imperfective[58] iharr, Ghadames äb̲r, perfective ib̲ro, intensive
imperfective ib̲ärr, Tashelḥait iri, verbal noun t-ayri 'to love'; for all
these I should rather reconstruct *ihʷar/*yahʷir (*yarihʷ in some of the
forms due to morphonological metathesis) comparable with Canarian güiro
([hʷiro] or [wiro] in real pronunciation, see 3.3.2.)
3.4.2. Ferro hero '... fuente De que le dieron título a la Isla', 'la
cisterna', 'la citerne', hera 'la arena donde el agua estaua', hieri,
jierro, hero 'Hierro' ~ Ahaggar ahir 'source d'une débit extrêmement
faible, alimentée par une ou plusieurs veines imperceptibles' ([Wölfel,
511]; compared with the Ahaggar form). K.-G. Prasse wonders, whether the
Ahaggar term could be a verbal noun from əhər 'boucher une ouverture, une
chose ouverte' [Prasse, 65], which corresponds to Ayr, E. Tawllemmet əhər,
Ghadames eb̲ər (äb̲ər, in Prasse's notation). While this is possible, though
somewhat strained semantically, there is a more direct parallel to the
Ahaggar ahir in Ghadames īb̲er (Prasse: eb̲är) 'canal, rigole, sequia
d'irrigation'[59] with the regular Ahaggar h ~ Ghadames b̲ correspondence.
3.4.3. Gran Canaria tara, tarja 'rayas en tablas, pared o piedras',
'señal para recuerdos', tarha 'marque pour les souvenirs', 'señal para
recuerdos' ~ Ahaggar t-êrəw-t, pl. t-êra 'lettre (missive); amulette
consistant en un écrit; dessin d'ornement (consistant en lignes, points,
figures géometriques)', Ayr t-irǎw-t, pl. t-ira 'letter (message), amulet
with an inscription on it', Ghadames ūrəb̲, verbal nouns īrəb̲, urrab̲, Zenaga
arha, Semlal ara, Qabyle aru 'to write', t-ira 'writing', etc. (most of the
forms compared in [Wölfel, 461-462]). The Proto-Berber form is *Harab̲
(verbal noun *Hirab̲) rather than *Harahʷ in the light of epigraphic East
Numidian (Thugga) ttrb and trbthn [ibid., 462] derived from the verb 'to
write' based on the consonantal root *rb. The weakening of the expected h >
w in Tuareg is not unusual, the interesting fact being an old laryngeal
retained in Zenaga where it normally falls out. Anyhow, for Ahaggar the
form *t-erahu-t, pl. *t-erah is reconstructable, the latter being a
plausible source of the presumably borrowed Canarian [t-arha] with a very
close meaning.
This is exactly the case mentioned in point 3.4. (b̲ < *b with a
laryngeal in the same root), which is corroborated by the East Numidian -b
and external comparison: Berber *Harab̲ 'to write' ~ Egyptian (Pyramid
texts) hꜣb (< *hrb) 'to send (a letter or message inter alia), to write a
letter', (Middle Kingdom-New Egyptian) 'letter, message' ~ Chadic (West):
Hausa rubū̀tā (compared in [Wölfel, 511]), Ngizim tàr̃nu 'trace, etch, draw;
write', (Central) Bura rubwa, rubuta, Buduma rebōde, Afade ohárbotú, etc.
'to write'. Since the meaning of 'writing' in Egyptian hꜣb obviously
developed from 'sending a message/letter' from 'sending' and, hence, is
secondary, the Berber and Chadic parallels must be early loanwords from
Egyptian either borrowed independently or, rather, by various Chadic
languages from Proto-Berber.[60]
3.5. Once the hypothesis of Ahaggar/Tamâhaq borrowings in Canarian is
taken into consideration, one can trace more highly probable cases, not
based on the above phonetic developments. A few examples:
3.5.1. Canarian (Tenerife, Gran Canaria) taxaraste 'panadero',
tajaraste 'Reigentanz' (in Canarian Spanish) [Wölfel, 483; no Berber
parallels quoted] < *ta-haras-te ~ Ahaggar ɣəryas 'quête accompagnée de
cris de joie et aubade faite par un certain nombre de négresses chez un
personnage considérable' [Foucauld, 1774) ~ Ayr, E. Tawllemmet ɣərias
'sorte de dance' [Alojali, 73].
3.5.2. Canarian (Ferro, Tenerife) heres 'pozas, charcas', 'albercas en
el Hierro' (this could be otherwise understood as the form identical to
hero in point 3.4.2. with the Spanish plural ending if not for -z in
herez), 'la citerne', herez 'la cisterna', eres 'hoyo o poceta formado en
las rocas impermeables del álveo de los barrancos, donde se acumula con el
agua de lluvia arena fina y limpia. Cuando se quiere extraer el agua se
forma un pequeño hoyo en la arena, hasta que aparece el agua ...' (Wölfel,
511-512 compares Afaggar ires 'puit') ~ Ahaggar îrəs 'puit profound, petit
trou creusé dans le sol pour en tirer de l'eau, ayant plus de 2 m de
profondeur' [Foucauld, 1670], Ayr, E. Tawllemmet, Tadghaq erəs id.
[Alojali, 163].
4. The known Canarian lexicon contains several Arabisms, which, in
theory, can be explained as direct borrowings from Arabic,[61] but are,
more likely, loans through the mediation of Berber.
4.1. Canarian (Palma) badanas '...los cueros [de las ovejas de pelo
liso] muy gruesos y sueltos', (Gran Canaria) badanas '... pieles adobadas
de color acanelado' ([Wölfel, 533]; compared with the Semlal/Sus and Ndir
forms) ~ Semlal a-bəḍan, Ndir, Seghrusen, Qabyle a-bəṭṭan 'peau d'animal
(mouton, chèvre, etc.)' [Dallet, 56], Ghadames ta-bəṭṭan-t 'peau de mouton
avec sa laine' [Lanfry, 33]. This term, which I could not find in Tuareg,
is a loan from a Maghrib dialect matching Classical Arabic biṭān-at-
'tanned sheepskin used as lining'; the primary meaning of the Arabic noun
is 'lining, inner part' derived from baṭn- 'belly, maw' (< Semitic *baṭn-
'(big) belly' [Militarev-Kogan, #42]).
4.2. Canarian (Gran Canaria) sabor 'consejo', 'consulta y
congregación', 'consejo de guerra', 'cortes o dieta general', 'une espèce
de diète', 'the Privy Council' ([Wölfel, 474-475][62] also quotes sambor
and tabor, the latter appearing a variant of a different term for 'council'
– see [ibid., 475]) looks very much like a widely spread Arabic term šūr-ā,
dialectal šawr 'advice, council' borrowed in most Berber languages:
Ghadames šawər [Lanfry, 41], Ahaggar šawer ([Foucauld, 133]; marked as
rare), Ayr, E. Tawllemmet šǎwwǎr 'consulter', šǎwwǎra 'consultation'
[Alojali, 186], Baamrani šawr, Qabyle eššwer 'conseil' [Dallet, 117], etc.
Canarian -b- may render *w in a Spanish transcription (then repeated in
French and English records where otherwise -b- would not be used to render
w) or *ww yielding *bb like, say, in Qabyle.
4.3. Canarian (Gran Canaria) taifa 'concurso o reunion' ~ Arabic ṭāʔif-
at- 'Stamm, Familie, Bande, Gesellschaft, Gefolge' (compared in [Wölfel,
423-424]). The Arabic term passed into various non-Tuareg Berber languages
(cp., e.g., Qabyle ṭṭayfa 'génération; societé; group' [Dallet, 846],
Ghadames ta-ḍəyyəf-t 'réception organisée d'hôtes, d'invités, parents et
amis, à l'occasion d'une fête de famille' [Lanfry, 85], etc., all marked as
Arabisms) which must have been an immediate source of the Canarian word.
4.4. Canarian (Tenerife?) arba 'cuatro', arbago, arbiago '40' [Wölfel,
626-636]. Wölfel suspects the identity of this numeral practically
rejecting the possibility of an Arabic loan in Canarian and only wondering
how this Arabic term for 'four', ʔarbaʕ(at)-, passed into the so-called
'second list' of Canarian numerals. Still the derived forms arba-go, arbia-
go '40' (cp. linago '20' < liin or lini '2', cansago '50' < cansa '5',
etc.) cannot be haphazard. The Arabic ʔarbaʕ- is a well-known loan in non-
Tuareg Berber languages (borrowed in Ayr and E. Tawllemmet as ərrəbu, but
in the meaning 'quart' [Alojali, 156], not 'four'), cp. Ghadames ərbəʕ
[Lanfry, 308], Qabyle rebʕa [Dallet, 703] 'quatre', etc., any of which
could have been an immediate source of the Canarian term.
4.5. Canarian (Tenerife, Gran Canaria and other islands) támaras
'frrutos', 'cuando los dátiles están en el rácimo se llaman támaras, voz
arábiga que significa también dátil', támara '…[palma = Phōnix
canariensis]…' [Wölfel, 507-8]. This Arabic word was borrowed in various
non-Tuareg Berber languages, e.g., Rif, Shawiya, Qabyle tmər and could pass
to the Canary island through them (recorded with the Spanish plural
ending); the pleophony of the Canarian loanword vs. nonpleophonic source
forms, though, remains unclear.
5. The distribution of Arabisms attested in Canarian within the Berber-
speaking area testifies against an intermediary part of Tuareg; it was
supposedly played by a non-Tuareg Berber language, whose speakers must have
constituted another wave (or waves) of migrants to the Canary Islands. This
is corroborated by a series of cultural terms common of Canarian and non-
Tuareg Berber (i.e. unattested in Tuareg), which, for some reason or other,
cannot belong to the old Proto-Berber-Canarian stock, e.g.: Canarian
(Tenerife) xercos 'un calsado hechura de abarcas', 'zapatos', 'espèce de
bottes', also jercos[63] ~ Snus a-herkʷâs, a-herkûs 'des sortes de
pantoufles de cuir, sans talon, en cuir de mouton ou de chèvre', Ntifa
(pl.) hirkas-in, Rif a-harkus, Imesfiwan i-hirkas, etc. Wölfel compares
these forms with Latin hircus 'goat', undoubtedly the source of Berber
*hirkʷas/*hirkus 'shoes made of goat skin' [Wölfel, 531-532].
5.1. This presumed non-Tuareg Berber migration wave, which brought a
few Berber and Arabic terms to the Canary Islands (but obviously no other
traces of Islam), should have taken place after the contacts of the
supposed source Berber language speaking group with Arabs in the continent
had reached the stage of adoption of some Arabic cultural realia by Berber
tribes prior to their conversion to Islam and while retaining their mother
tongue – otherwise we must suppose that Berbers who reached the Canary
Islands were Arabic speakers which would contradict the above
argumentation. Therefore, late 7th-8th centuries are probably to be
considered terminus ante quem non for this migration.
6. Compared with the non-Tuareg Berber wave postulated in point 5, the
Tamâhaq Tuareg one seems to have influenced the Canarians much more as seen
from the analysis of the presumed Ahaggarisms in Canarian. Their number and
a touch of prestige about at least part of them suggest the Tamâhaq-
speakers' impact should have been sensible enough. It is also indicative
that in the list of these Ahaggarisms, there are neither Arabisms nor
Islamic impacted realia which seems to restrict the dating of the Tamâhaq
speakers' migration to the period between the Proto-Tamâhaq separation from
the other Tuareg stock and the adoption of Islam by the Tamâhaq-speaking
Tuareg tribes. But the big question remains: are there any facts outside of
linguistics that could serve as evidence of Tuareg contacts with the
indigenous Canarian cultures?
6.1. Before trying to answer this question, we have every reason to
put a counter-question: what could the distinctive features of Tuareg
culture be that are expected to have been manifested in the Canary Islands
as traces of the Tuareg migration? Extremely scanty material culture and a
Spartan way of living of the Tuaregs in the Sahara leave little opportunity
of revealing such features.
6.1.1. One of them is a vestige of Christianity in some of the pre-
Islamic Tuareg traditions[64] and lexicon[65] and vague traces of
monotheism probably tinged with Christianity in the Canarian[66]
traditions. For the Tuareg one, the only plausible explanation is a
cultural and linguistic heritage of those Libyan tribes, the Garamantes[67]
and the associated ones, which once came in contact with the Christian
(Roman, Byzantine) world lost later in the process of Islamization. For the
Canarian tradition, I am inclined to identify that hypothetic vestige with
the Tamâhaq migration thus dating the latter to the period when the Tuaregs
were not yet influenced by Islam at all; taking all this into account I
would bet on the 7th century CE.
6.1.2. Still one feature of the Tuareg culture can be justly regarded
as both distinctive and prominent, that is the Tuareg script, tifinaɣ, one
of the varieties, and the only surviving one, of what is known as the
Libyan script.[68] There are several slightly differing variants of tifinaɣ
employed by various Tuareg tribes, and one considerably differing from the
others published by Foucauld without indicating the source; this latter was
no longer in use at the beginning of our century and could be thus called
"Old Tuareg". The most wonderful thing about tifinaɣ is that, as far as I
know, it is the only script used for a pretty long period of time by the
ethnic groups standing far in their social and economic development from
the stage of early or proto-civilization typical of cultures based on early
written traditions (like Sumerian, Egyptian, Chinese, Phoenician and so
on). It would be more understansable if tifinaɣ had been one of those short-
lived ephemeral alphabets which have been invented here and there, or if it
had been recently borrowed from a more developed culture, or if, at least,
it had been an esoteric cryptography preserved through centuries. None of
these is the case. One of the two bilingual (Libyan-Punic) inscriptions in
the East Numidian script related to tifinaɣ, is dated to the 2nd century
BCE; as for the social characteristics of tifinaɣ, according to H.Lhote
[Lhote], in the middle of the past century every second woman and every
third man could easily write and read short inscriptions in it made on
rocks, vessels, shields and bracelets (nowadays also in paper) conveying
dedications, love messages, marks of property, dictums, etc.
6.1.3. As for the origin of the Libyan script, I. Gelb [Gelb]
was definitely wrong to qualify it as a script whose symbols were
arbitrarily invented. I have no doubts that it is related to the Semitic
script – rather derived from it than going back to some common source. The
interesting thing is that, apart from the nine symbols I identify in the
Libyan and Semitic (common to both Phoenician and Arabian) scripts as
having the same or alike shape and render the same or congenerous sounds,
there are six more idenifiable symbols common to the Libyan and one or more
of Southern varieties of the Semitic script[69] and only three common to
the Phoenician-Punic ones, all looking exactly Neo-Punic (see Appendix 1).
To explain that especial similarity between the Libyan and Arabian scripts
by borrowing by the former from the latter one has to answer a catch
question: what could the presumed contact's routes have been and where?
6.1.3.1. Curiously, the answer appears to come from the origin
of the common Berber term for 'iron': Old Lybian zl, Shilh uzzal, Wargla,
Mzab uzzəl, Qabyle uzzal, Ghadames wazal, East Tawllemmet uzzel, Ahaggar,
Ghat t-az/ẓuli, etc. The protoform should be reconstructed as
*(H)uzzāl. This word does not have a "native" Berber etymology (such as a
meaning shift from some qualitative adjective, e.g., a color name). It does
not have cognates in other Afrasian languages, either.[70] Instead,
traditional scholarship suggests its origin from Phoenician brzl (the
corresponding Hebrew term is barzäl).[71] This etymology, though, faces the
problem typical of any unreliable and dubious etymology: inability to
explain phonetic changes accompanying a postulated borrowing, in this case,
elision of the first syllable bar- or transformation of the initial Semitic
b- into Berber w/u[72] with a full assimilation of the second Semitic
radical -r-.
6.1.3.2. What passed unnoticed, to my knowledge, is the
connection between Berber *(H)uzzāl (to be clearly reconstructed as
*ʔuzzāl) and a passage from Ezekiel (27:19): wə-dān[73] wə-yāwān mə-
ʔūzzāl... nātānnū barzäl ʕāšōt...[74] "and Dan (or Wedan) and Yawan from
Uzzal paid (you with) worked iron…" That is how ʔūzzāl is commented in
HALOT 21: "...in Arab. trad. the pre-Islamic name of s anʕa, capital of
Yemen."
And that is what says the online Encyclopedia of The Bible:

Uzal
(אוּזָ֖ל), a son of Joktan in the Table of Nations (Gen 10:27; 1 Chron 1:21).
The two sons of Eber, Peleg and Joktan, represent the two main divisions of
the Sem.-speaking people. Joktan is considered by some scholars to be the
founder of the Arab nation. There is an Arab. tradition that Uzal was the
original name of Sanaa, the capital of Yemen in SW Arabia in both
Himyaritic and modern times. Bright (wrought) iron is mentioned as one of
the exports to Tyre... Another suggestion for Uzal is Azalla in the
vicinity of Medina. Ashurbanipal, the Assyrian, records in his annals the
capture of Azalla in his account of the expedition against the Nabateans…
In the LXX rendering of the passage concerning Tyre's trading with many
countries (Ezek 27:13-23) Uzal is included (27:19). The reading of the RSV
follows the LXX: "...and wine from Uzal they exchanged for your wares"…
Since the Azalla mentioned by Ashurbanipal is located in this region, it is
best to identify the Biblical Uzal with a city near Medina.

Full coincidence of the Berber *ʔuzzāl 'iron' with the Biblical
toponym ʔūzzāl, associated with iron at that, leaves little doubts about
the source of borrowing of the Libyo-Berber term; naming a product or
merchandise after the locality where it is manufactured or delivered from
is quite common.
6.1.3.3. There are two plausible routes of this term's (and
the corresponding artifact) conveyance from the South Arabia to the Libyo-
Berber speaking area: across the Bab-el-Mandeb strait and further via Nubia
northward and northwestward; and via the Phoenician seaports to the African
Mediterranea coast. While early Libyan-Nubian contacts are confirmed by the
Berber loanwords in Nubian (see Appendix 5), ʔūzzāl is mentioned in Ezek 27
known, together with Ezek 26, as Lamentation for Tyre, which implies the
Phoenician mediation. As for the possible conveyance time, cf. Wikipedia,
Iron Age: "The widespread use of iron weapons which replaced bronze weapons
rapidly disseminated throughout the Near East (North Africa, southwest
Asia) by the beginning of the 1st millennium BC". This dating corresponds
to the one obtained glottochronologically by the author for the Proto-
Berber split (see Appendix 1).
6.1.3.4. This 'iron-clad' connection between South Arabia,
whatever city, Sanaa or Azalla, is Uzzal identifiable with, and North
Africa may well account for the role of the South Semitic script in
formatting the Libyan one.
6.1.4. Recent investigations reveal an increasing number of rock
inscriptions in Libyan characters found in various Canary Islands, but
mostly in the island of Ferro (Ierro). In the light of our hypothesis of a
Tuareg migration to the Canary archipelago it would be only natural to try
and read them with the help of the Ahaggar dictionary. As a first step, I
chose the inscriptions containing characters having the same graphic form
and phonetic value in most varieties of the Libyan script (the main
difficulty for a decipherer is certain similar graphemes with quite
different reading). They are as follows: [75]

1) Canarian y Old Tuareg tifînaɣ y
tifînaɣ of Ahaggar y Tripolitanian y

(all are variations of East Numidian y ,

Old Tuareg y tifînaɣ of Tadghaq y
Tripolitanian y
)

2) Canarian w Old Tuareg w
Ahaggar w (late development from )

Tripolitanian w East Numidian w

3) Canarian m Old Tuareg m
Ahaggar m

Tripolitanian m East Numidian m

4) Canarian n Old Tuareg n
Ahaggar n

Tripolitanian n East Numidian n
(and ?)

5) Canarian l Old Tuareg l
Ahaggar l

Tripolitanian l East Numidian l [76]

6) Canarian r Old Tuareg r
Ahaggar r

Tripolitanian r East Numidian r

7) Canarian h Old Tuareg h
Ahaggar h

Tripolitanian h
[77]

8) Canarian ɣ Old Tuareg ɣ Ahaggar ɣ
(developed from, cp. q

developed from : dots from lines)

Tripolitanian ɣ East Numidian
(reading uncertain, perhaps h)
[78]

9) Canarian g (reading uncertain) Old Tuareg g

tifînaɣ of East Tawllemmet and Tadghaq (read as ž, z, probably <*g; not
attested in Ahaggar)

Tripolitanian g
East Numidian g

10) Canarian s
Old Tuareg
š s

Tripolitanian (reading uncertain)
East Numidian s or š

11) Canarian s1
Ahaggar š

tifînaɣ of Ayr š Tripolitanian
(reading uncertain)

East Numidian s (?)

12) Canarian s2[79] tifînaɣ of Adghaq š East Numidian
(probably also )[80]

13) Canarian ṣ (?) tifînaɣ of
Ayr
[81]
Tripolitanian ṣ East
Numidian ṣ

14) Canarian z (?) Old Tuareg
Ahaggar (a sign for rendering Arabic ṣ)
z ẓ

tifînaɣ of Ghat ṣ Tripolitanian (reading unknown, see #17)
East Numidian ẓ or z

15) Canarian z or ž (reading uncertain) Ahaggar ž Tripolitanian
z East Numidian z (?)

16) Canarian t Old Tuareg t Ahaggar t
Tripolitanian t East Numidian t

17) Canarian ṭ (or less probably, ḍ) Old Tuareg ṭ Ahaggar ḍ
tifînaɣ of Ghat, Tadghaq ṭ, ḍ

Tripolitanian (reading unknown, see #14)
East Numidian ṭ

18) Canarian d Old Tuareg d
Ahaggar d

Tripolitanian (reading uncertain, see #9)
East Numidian d
(and, probably, )

19) Canarian b Old Tuareg b
Ahaggar b

Tripolitanian b
East Numidian b
(and ?)

6.1.4. A 'Tripolitanian' inscription from Ghirza (Wadi Zemzem) in
Lybia (the reading and interpretation by the writer) is taken for
comparison with the Canarian inscriptions. The former had been scratched on
the plaster wall of a house rebuilt on the ruins of a pagan temple and
occupied, according to O. Brogan [Brogan] who discovered it, in the 10th
century C.E., probably by Hawara (Hawwara, see 3.2.2.6.), a tribe with the
ethnic name identical with that of Ahaggar. It is to note that in Libyan
inscriptions made into vertical columns, the usual direction of reading is
from bottom to top. As for the direction of successions of 'words', i.e.
columns of characters, in some inscriptions they are read from left to
right, in others vice versa (in the Ghirza inscription it does not matter,
for the 'words' are not syntactically linked).

Transliteration:

line (A) ( (B) m ṣ w r n
line (C) ( (D) l gg[82] w t n
line (E) ( (F) z n g n lt[83] n s m n

Reconstructed vocalized form:

line (A) ( (B) miṣūr-an
line (C) ( (D) laggʷa-t-an
line (E) ( (F) zanag-an (or i-znag-an) ələt naṣūm-an

Interpretation:

Each of the first two lines (A) ( (B) and (C) ( (D) represents a noun with
the typical Berber plural suffix -n. In the line (E) ( (F) three words are
distinguished, the first and the third also representing nouns with the
plural -n suffix, and the second, ələt, is a kinship term meaning
'daughters' in Ahaggar (the term is known to other Tuareg and non-Tuareg
Berber languages, but in a different form). The four nouns in plural
coincide exactly with four well-known ethnonyms of Libyan and Berber
tribes: Miṣṣuran;[84] Leuathae (of Procopius) or Laguanten (of Corippus);
Nasamones (of Herodotus and Strabo); and Zenaga (cp. the same name of
modern Berber-speaking group of Mauritania).

Translation: Miṣṣuran, Lawwatan, Zanagan daughters of Nasamones.

6.1.5. Here are several rock inscriptions found in the island of
Ferro. Some of them have been published, while the others reached the
author through the kind collaboration of Prof. H. Nowak who was the first
to find and copy them.
6.1.5.1. Inscription published by H. Biedermann [Biedermann]. It was
found by H. Nowak on the coast of La Caleta:

Transliteration:

line (A) ( (B) w d. s n[85]
line (C) ( (D) t s d
line (E) ( (F) m t

Reading in the Ahaggar language[86] and the English translation:
line (A) ( (B) wad əssən 'hey, you! know (or beware):'
line (C) ( (D) təssəda 'sinking'
line (E) ( (F) mət '(is) death'

Comments:

line (A) ( (B) wad (< wa di) is 2nd person singular pronoun; essen is
imperative of the verb 'to know'
line (C) ( (D) t-əssəd-a is a verbal noun of əssəd 'to plunge, to sink, to
immerse'
line (E) ( (F) mət is a derivative of the verb əmmət 'to die'; it has no
syntactic relations in a sentence and is always placed at its end conveying
the idea of a sudden or violent death.

Interpretation: a warning on a coastal rock against getting into water?

6.1.5.2. Inscription found in El Julan by H. Nowak (personal
communication):

Transliteration:

line (A) ( (B) y m r n
line (C) ( (D) y n m s. d l

Reading in the Ahaggar language and the English translation:

line (A) ( (B) yəmməran '(a) passing (one)'
line (C) ( (D) i n mess ədəl 'anyone of (the) master (or the Lord): come
(and stay) overnight!'

Comments:

line (A) ( (B) yəmmər-an is a masculine singular form of a participle of
the verb əmmər 'to pass (by)'
line (C) ( (D) i is an indefinite pronoun; n is nota genitivi; mess means
'master, owner' and 'the
Lord' when used with enclitic possessive pronouns of the 1st person
singular (Mess-i) and plural (Mess-ineɣ); edel is imperative of the verb
'to come for a night, to stay overnight at'.

Interpretation: a sort of invitation or advertisement for passers-by of a
certain category in a local
inn?
6.1.5.3. Inscription published by H. Biedermann (op. cit. in 6.1.5.1.)
found by H. Nowak in the mouth of the Barranco de Tejeleita:

Transliteration:

line (A) ( (B) y s n y n
line (C) ( (D) ( (E) n m d r s n


line (A) ( (B) yssən yən 'knows one'
line (C) ( (D) ( (E) n imâdrâsən 'of (the) few'.

Comments:

line (A) ( (B) yəssən is imperfective 3rd person singular of the verb essen
'to know'; yən is a
numeral 'one'
line (C) ( (D) ( (E) n is nota genitivi; i-mâdrâs-ən is plural of a-mâdrâs
(these forms are not
attested in Foucauld's dictionary, but are easy to reconstruct after
regular morphological patterns), verbal noun from ədrəs 'to be not
numerous, to be few'.

Interpretation: a notice or a maxim (kind of Canarian-Tuareg gnosis?).

6.1.5.4. Inscription published by H. Nowak [Nowak]. It was found by
him in Hoyo Blanco.

Transliteration:

line (A) ( (B) w r d l
line (C) ( (D) t d n t
line (E) ( (F) n m r


line (A) ( (B) ur ədəl 'don't beg (or hope) for'
line (C) ( (D) tâdənt 'fat' (or 'lard')
line (E) ( (F) animir (probably an-imir, from êmir 'moment') is an adverb
meaning 'still, yet, not yet, more, no more, etc.')

Comments:
line (A) ( (B) ur (or wər) is a negative particle used prepositionally with
verbs; ədəl is imperative of the verb 'to beg (or hope) for'
line (C) ( (D) tâdənt is a feminine singular noun meaning 'fat', 'lard'
line (E) ( (F) animir (probably an-imir, from êmir 'moment') is an adverb
meaning 'still, yet, not yet, more, no more, etc.')

Interpretation: this trivial household message looking odd enough on a rock
is not unusual for
the genre of tifînaɣ inscriptions.

6.1.5.5. Inscription found in La Caleta by H. Nowak (personal
communication).

Transliteration:

line (A) ( (B) n n m r y n


line (A) ( (B) ənnən émir yən 'read sometimes'

Comments:

line (A) ( (B) ənnən is imperative of the verb 'to spell (out), to read
letter by letter' conveying the idea of reading only in tifinaɣ (for əɣər,
a borrowed general term for 'to read', see Appendix 4 #7); émir is a noun
meaning 'moment' and, followed by yən 'one', 'one moment, a moment' and, in
some contexts, 'sometimes'.

Interpretation: all the three lines are identical and arranged in such a
way that the lower symbol of the upper line is on the same level as the
upper symbol of the middle line, while the lower symbol of thlatter is on a
level with the upper symbol of the lower line; the whole group looks like
an exercise in reading or a slogan calling the viewers to reading in
tifinaɣ (at a kind of school?).

Appendix 1

Libyan and Semitic scripts affinity

Appendix 2

A hundred word (Swadesh's) list of Berber languages[87]

1. all
1) Siwa kôma; (2) Ghdm ak, Ghat ik-it, Zng äkki 'every' (with pl. 'all'),
Sml akkʷ, Izd akk̥ȯ,
Mz aš (<*ak), Nfs ak, Qbl aḵʷ; (3) Ghdm imda, Ahg emdu, Tai emda-n; (4)
Ahg, Ayr a-ɣrud [] No Auj, Sok, Rif, Shw, Tim.

2. ashes
1) Siwa iɣid, Ghdm išəd (-š- < *ɣ; cf. tašarḍemt 'scorpion' < *-ḳarḍem-
t), Sml iɣd, Izd, Mz
iɣəd, Shw, Nfs, Qbl iɣəḏ; (2) Auj áġb̲ẹl; (3) Sok t-nîfes-t, Rif ṯ-nifə's-
ṯ; (4) Ghat iẓəd, Tai īẓed, Ahg êẓəd, Ayr eẓəd; (5) Zng ti-ʔyif-t (<*-gif-
); (6) Tim ləflaf.

3. bark (of tree)
1) Siwa ā-sān (b. of palm-tree, no other word for 'bark' in the oasis);
(2) Ghdm ti-fra
(takuka quoted by Bl. is glossed 'écorce de un courge'); (3) Ghat ta-šə-
nǯəf-a, Tai ta-se-nǯef-a, Ahg ta-sə-ñḡəf-a; (4) Ayr t-yəšše; (5) Ayr t-
yəŋw-ǎt (related to Ahg t-iñe-t 'morceau d'écorce'), Zng tanwiʔḏ (<*ta-
nwiʔ-t); (6) Sml ti-fərk-it, Izd i-fərḵ-i, Shw ta-fərš-iṯ; (7) Mz t-ilm-
ay, Nfs t-iləm-it, Qbl i-šləm [] No Auj, Sok, Rif, Tim.

4. belly
1) Siwa ǯar, Auj ažâr, Sok žār (all <*gar); (2) Ghdm ta-dis-t, Rif, Mz,
Shw a-ʕddis (though ʕ
usually points to Arabic loanwords, this must be an authentic Berber term
in view of no ʕ in other Berber idioms), Tim ta-dis-t, Nfs ti-ddis-t;
(3) Tai t-asa, Ahg t-esa; (4) Ayr ta-rfa; (5) Zng ta-ḫsäh; (6) Izd a-
ḫdil; (7) Qbl a-ʕəbbuḏ̣ (like in the case with a-ʕddis, ʕ does not signal
an Arabism, as there are forms without it elsewhere, cf. Ntifa a-buḍ id.)
[] No Sml, Ghat.

5. big
1) Siwa a-zwār, Ghat zwer, Tai i-huhar-en (<*-zuhar-), Ayr a-mi-zwăr; (2)
Auj móqqar, Sok
måqqår, Ghdm məqqur-ən, Ghat məqqər, Tai moqqeren, Ahg a-mɣar, Sml moqqōr-
ən, Izd, Nfs a-moqr-an, Rif a-m'qr-an, Shw a-məqr-an, Tim a-məqq-ān (with
assimilation of -r-), Qbl a-məqʷr-an; (3) Zng ə-kkušb-ih, yo-ḵšb-i (<
*kəsw-; hardly < Arab ka/isb- 'gain, profit' as claimed by some authors
in view of Izayan a-kəsw 'big', obviously related); (4) Mz a-zâluk.

6. bird
1) Siwa a-ǯṭiṭ, Auj a-gaḍîḍ, Sok a-ždît, Ghdm a-gaḍiḍ, Ghat a-ǯəḍiḍ, Tai
i-geṭeṭ, Ahg e-
gəḍiḍ, Ayr a-gəžiḍ, Zng a-ɣḍuḍ (<*-gḍuḍ), Sml a-yḍīḍ, Izd a-gḍiḍ, Rif a-
žḏiḍ, Mz a-žəddid, Nfs a-žəṭiṭ; (2) Tim zukk [] No Shw, Qbl.

7. bite (v.)
1) Siwa addad, Auj ěddéd, Ghat, Ahg, Ayr ədəd, Tai added; (2) Ghdm embər
(<*ən-bər),
Rif brʔm (<*bər-m); (3) Tai, Hgr ekš; (4) Zng a-ʔmoḍ̲, umaḏ, (hab.) ita-
mazz (<*-maẓ); (5) Sml əbbi; (6) Izd ḳərrəš, Qbl kʷərrəš (not < Arabic,
as claimed by some authors); (6) Shw ḏərrəm, Nfs ədrəm; (7) Tim ɣəzz []
No Sok, Mz.

8. black
1) Siwa a-zəṭṭāf, Auj sḍẩf, Sok sěṭṭof, Ghdm ə-ẓdef, Ghat səṭṭaf, Tai yu-
ẓẓaf, Ahg te-ẓẓef-e,
Nfs a-ẓəṭṭaf; (2) Tai i-kawal-en, Ahg te-kkewel-t, Ayr e-káwəl (i-ẓḍaf is
rare); (3) Zng əḏəǯ (<*ədəl); (4) Sml a-ssgan; (5) Izd a-bəḫḫuš (<*bərk-
uš, with assimilation of -r-), Rif a-b̲ārš-an (<*bark-an), Mz a-bərš-an, a-
brč-an, Shw a-bərk-ən, Tim a-bəḥk-ān (-ḥ- <*r), Qbl a-b̲ərḵ-an.

9. blood
1) Siwa i-damm-ən, Auj dímm-en, Sok i-dämm-ẹn, Ghdm demm-en, Zng dəmm-
ən̥ḣ, Sml,
Izd i-damm-ən, Rif i-ḏam-'n, Mz i-dam-ən, Shw i-ḏamm-ən, i-ḏəm, Tim, Nfs
i-dəmm-ən, Qbl i-ḏamm-ən; (2) Ghat a-zəni, Tai a-henî, Ahg a-hni, Ayr a-
zni.

10. bone
(1) Siwa î-ɣəs, Auj a-γẩs-t, Sok ī-γöss, Ghdm ɣəss, Ghat, Tai i-ɣəs, Ahg
e-ɣəs, Ayr e-ɣăs, Zng i-ssi (<*i-ɣs-), Sml i-ḫss, Izd i-ɣs, Rif ɣsu, Mz,
Shw, Nfs i-ɣəs, Tim ī-ɣəṣ, Qbl i-ɣəss.

11. breast
1) Siwa i-fifen, Ghdm, Ghat, Tai i-fef, Ahg é-fef 'sein', Zng uff-än
'sein' and 'pis', Sml ti-
ffī-t, Izd i-ff, Mz i-f; (2) Auj zzô, pl. zzíw-en 'breast'; (3) Auj a-
dmār, Sok a-dmár, Ghdm a-dmār, Ghat, Tai a-dmar (Ahg i-dmâr-en 'partie
basse de la poitrine chez les quadrupèdes'), Sml i-dmar-ən, Izd, Nfs a-
dmər, Rif ṯa-ḏmaa-ṯ, Mz i-dmar-ən, Shw i-ḏmar-ən, Tim i-ḍmān (loss of
"rhotic" -r accompanied by "emphatization" of -d), Qbl i-ḏmar-ən; (4) Ayr
a-hǎror; (5) Zng a-gärgur 'poitrine'; (6) Sml ti-bbi-t, Izd ṯ-bəbbi-š,
Rif a-bbi-š, Shw ṯa-bbu-š-t, Nfs bibbi-š, Qbl ṯa-bbuš-ṯ.

12. burn (trans. v.)
1) Siwa éqqed, Tai, Ahg, Ayr əqqəd; (2) Ghdm ss-erγ (əqqəd 'cautériser,
allumer'), Ghat,
Mz, Nfs ərɣ, Tim saɣu (< *sa-rḳu), Qbl ss-ərɣ; (2) Ghat entes; (3) Zng
yäss-əndär; (4) Sml ždər (<*gdər or z-der?); (5) Izd kməḍ, Rif šm'ḍ [] No
Auj, Sok, Shw.

13. claw/fingernail
1) Siwa a-čir, Auj í-šker, Sok i-ššέr, Ghdm a-škar, Ghat i-škər, Tai i-
sker, Ahg ê-skər, Ayr
a-škar, Zng skər, Sml a-šḫar, Izd i-skər, Rif i-ššaā (-ā <*r), Mz a-ššar,
Shw i-ššər, Tim i-šša, Nfs a-ššər, Qbl i-ššər.

14. cloud
1) Sok tě-žnâu, Ghdm ti-žnew-t, Ayr a-gnaw, Sml ta-sə-gnū-t, Izd i-s-
ignu, Rif a-s-ignu, Mz
ta-žənnu-t, Shw ṯə-zənnaw-ṯ, Tim ti-žnu-t, Qbl a-s-iḡna; (2) Ghat či-
ǯrək; (3) Tai a-ẓyar, Ahg aẓiar 'nuage isolé' (a-ḡənna 'ciel, pluie,
nuage'); (4) Ayr i-fəssay-ǎn 'nuages légers'; (5) Zng ta-biɣ-d (<*ta-
biḳ/g-t); (6) Nfs i-blem [] No Siwa, Auj.

15. cold (adj.)
1) Siwa a-ṣemmaṭ, Auj a-šemmáḍ, Sok ta-ṣẹ́nḍ-i (all three are nouns), Ghdm
semməḍ (v.),
Ghat sammiḍ (adj.), Tai ī-smaṭ (v.), Ahg e-samiḍ (adj.), Ayr ismaḍ (v.),
Zng tə-šmuǯ -i (-ǯ <*d <*ḍ), Izd a-səmmaḍ, Rif ə'sm'ḍ, Mz səmməd, Shw a-
səmmiḏ̣, Nfs i-semoṭ, Qbl a-səmmaḏ̣; (2) Ghdm e-grəs (v.); (3) Sml a-
kərram; (4) Tim i-qarāf-ən (n. pl.)

16. come
1) Siwa us, ūs-éd, Auj yuš-âd, Sok yus, yus-əd, Ghdm ās, asū-d; Ghat as,
as-ed, Tai as-ed,
Ahg as, Ayr asu, Zng yəssä, Rif, Mz, Shw, Tim, Qbl as, Nfs as, as ed; (2)
Auj yid, Izd addu (3) Auj woḍ, Ghat awoṭ; (4) Sml ašk.

17. die
1) Siwa immūt, Sokna, Ghdm, Ghat, Ahg, Mz, Nfs əmmət, Tai yəmmat, Ayr
ămmat, Zng
əmmih, Sml ĕmmət, Izd mmət, Rif əʔmmeʔṯ, əmmət, Qbl əmməṯ; (2) Ahg kətiy-
ət; (3) Zng äddiih (with part. näh: yaddā-näh 'il est mort') [] No Auj,
Shw, Tim.

18. dog
1) Siwa a-gurzn-î, a-qurzin-i, əl-gurazən, lu-grāzén, Auj a-ɣzin, gzîn;
(2) Sok i-yidi, Ghdm
īḍi, Ghat, Tai a-ydi, Ahg e-ydi (pl. i-yăṭâ-n), Ayr iži, idi, Zng iḏi,
Sml a-ydi, Izd i-ydi, i-gdi, Rif, Mz, Tim, Qbl a-yḍi, Shw a-yṭi, Nfs yudi
(pl. iṭa-n).

19. drink (v.)
1) Siwa su, swi, Auj šu, Sok sū, Ghdm əsw, Ghat, Tai, Ahg əsu, Ayr ašu,
Zng ešbi (<*ešwi),
Sml, Izd, Rif, Mz, Shw, Tim, Nfs su, Qbl səw.

20. dry (adj.)
1) Siwa qqor, Auj qqūr, Sok əqqār, Ghdm əqqār, Ghat, Tai əqqor, Ahg, Ayr
iɣar, Zng
yaʔwur (<*yaɣur), Sml i-qqur-ən, Izd qqar, Mz əqqor, Shw qorr, Nfs iqqer;
(2) Qbl a-kiw-an [] No Sok, Rif, Tim.

21. ear
1) Siwa ta-məzzuḫ-t, Sok ta-məzzuḫ-t, Ghat ta-məzzuk, Tai tamẓuk, pl. ti-
mẓuǯ-in, Ahg,
Ayr ta-məẓẓuk, Zng ta-mazgu-ḍ, Sml a-məẓẓuɣ, Izd a-məzzuɣ, Rif a-mʔzzuɣ,
Mz ta-məzuɣ-t, Shw i-məžži, Tim ta-məžž-ət, Nfs ta-məǯ-it, Qbl a-məẓẓuɣ;
(2) Auj í-sem, Ghdm ī-səm; (3) Ayr ta-n-ḍərəq.

22. earth
1) Siwa i-ždi (<*i-gdi); (2) Auj tä-mûr-t, Ghdm ta-mmur-t ('terre, sol'),
Rif ṯa-mur-ṯ, Mz ta-
mur-t, Shw ṯa-mur-ṯ, Nfs ta-mur-ṭ; (3) Sok u-mlâl; (4) Ghat, Tai, Ahg,
Ayr a-maḍal; (5) Ghdm u-kal terre (matière)', Zng aʔ-gäy (<*-kal), Sml a-
kǟl, Izd a-šal (š<*k), Qbl a-ḵal [] No Tim.

23. eat
1) Siwa əč, Auj ečč, Sok əčč, Ghdm əšš, eč, Ghat əč, Ayr ătt-šu, Zng
yitYšä, Sml əšš, Izd t-š,
Rif eʔšš, Mz əč, əš, Shw, Nfs əč, Qbl əčč (all <*it-š); (2) Tai, Ahg əkš
[] No Tim.

24. egg
1) Auj t-isĩ (isolated term, perhaps, < Chadic); (2) Sok ta-s-adəl-t, ta-
z-dâl-t, Ghdm ta-s-
adəl-t, Ghat, Ahg, Ayr ta-s-adal-t, Tai ta-s-edal-t, Mz ta-z-dəl-t (all
<*dal 'cover, lay down'); (2) Ayr ta-kəki-t (otherwise < Arab. kayk-at-
); (3) Zng to-ʔž̲ih (as Zng ʔ < *ḳ, ž̲ can render *l, h can render any
pharyngeal or laryngeal, the most probable origin is <* ḳVlḥ- < AA
*ḳa(w)ḥil- ~ *ʔVn-ḳulaliḥ- 'egg'); (4) Sml ta-gläy-t, Izd ta-glay-t; (5)
Rif ṯa-mžač (<*ta-mlal-t), Swy ṯa-məllal-t, Qbl ṯa-məllal-ṯ [] No Siwa,
Tim, Nfs.

25. eye
1) Siwa ṭ-əṭṭ, Auj a-ṭi, Sok, Ahg t-iṭ, Ghat č-iṭ, Tai t-èṭ-t, pl. t-iṭa-
t-win, Ayr t-yeṭṭ, Zng t-uḍ,
Sml ṯ-iṭṭ, t-iṭṭ, Izd t-iṭ, Rif ṯ-iṭ, Mz t-iṭ, ṭ-iṭ, Shw a-ṯ-iṭ, Tim, Nfs
ṭ-īṭ, Qbl ṯ-iṭṭ. (2) Ghdm a-wəll.

26. fat
1) Siwa ta-dəm-t (not < *-dən-, cf. Berb.: Sened ta-dum-t 'fat'; Semitic:
Arab. ʔidām-
'condiment'; E.Cushitic: Sidamo duʔma 'fat', etc.); (2) Auj d-díen-t, Sok
ta-dẹ́n-t Tai ta-den-t, Ahg t-âdən-t, Ayr t-dən-t, e-dan, Zng tä-ḏūn-ət,
Sml ta-dun-t, Shw a-ddun-t, Tim ta-dun-t, Nfs ta-dûn-t (< Afrasian *duhan-
'fat': Semitic *duhn-, *daha/in-; W. Chadic: Sura ḍɔ́ŋ, Polchi ḍiin,
etc.); (3) Ghdm ta-ssəm-t, Ghat, Tai i-sim, Qbl ṯa-ssəm-ṯ; (3) Mz udi []
No Izd, Rif.

27. feather
1) Auj, Sok, Ghdm ta-ržal-t (<*-rgal-), Zng tä-rgäl; (2) Ghat a-zaǯ (same
as 'hair' <*-zag);
(3) Tai, Ahg a-fraw; (4) Mz, Nfs tə-bulbul-t [] No Siwa, Ayr, Sml, Izd,
Rif, Shw, Tim, Qbl.

28. fire
1) Siwa tə-msi, Sok ti-msi, Ghat či-msi, Tai t-əmsi-n, Ahg, Ayr te-mse,
Rif ṯi-msĩ, Mz
ti-msi, Shw ṯi-məs, Tim ti-msi, Qbl ṯi-məss; (2) Auj a-fîu, Ghdm u-fa,
Ghat ə-fiw, Ayr e-few, Izd a-fā, Nfs tə-faw-t; (3) Zng oʔ-ẓ̲uẓ̲ən; (4) Sml
takǟ-t.

29. fish
1) Ghdm u-lism-a, Ghat a-ləmš-ay, Tai a-sulmaî, Ahg a-sûlmäi, Zng siǯm-
an, Sml, Izd, Qbl
a-sləm, Rif a-srʔm (r <*l), Shw a-səlm; (2) Nfs trabit [] No Siwa, Auj,
Sok, Ayr, Mz, Tim.

30. fly (v.)
1) Siwa əm-fər, Sml firri, Izd afru, Shw fərfər, Nfs far; (2) Ghdm ekkəd
(<*eggəd
<*ewwəd), Ghat əǯǯəḍ (<*əwwəḍ), Tai iggaṭ, Ahg, Ayr əggəd (-gg- <*-ww-);
(3) Tai illaî, Sml, Izd ayll; (4) Zng yunḍar (less likely < Arab.); (5)
Mz bəšš; (6) Shw afəg, Qbl afəḡ; (7) Tim əkka (< *əkkar) [] No Auj, Sok,
Rif.

31. foot
1) Siwa ṭar, Auj, Ghdm, Ghat a-ḍar, Ahg a-ḍər, Ayr a-ḍăr, Sml, Izd a-ḍar,
Rif ḍaā (-ā <*
-r), Mz, Nfs ḍar, Shw ḏ̣ar, Qbl a-ḏ̣ar; (2) Sok ti-škən-t; (3) Tai a-tfar;
(4) Zng T: t-ənš-ih 'pied' (a-ḏ̣aʔr 'pied + jambe'), Tim t-ins-ət.

32. full
1) Siwa čūr (<*t-kur) Sok i-ččūr (v.), Ghdm eṭ-kur (v.), Ghat it-kar-ən,
Tai eṭ-kar-en, Ahg
əṭkər (v.), Ayr əkkər (v.), Rif šuǟ (<*šur <*t-šur <*t-kur), Mz čar (v.;
<*t-šar <*t-kar), Shw šara, Nfs tə-ššar, tə-ččar (v.), Qbl a-čar-an; (2)
Zng oʔvər (v.; < *ʔVfVr), (3) Zng yuḏnug (v.; < *-dnuy) [] No Auj, Sml,
Izd, Tim.

33. give
1) Siwa ūš (<*ūk), Sok aš, yušâ, Izd š, Rif uḵš (<*ukVk), uš, Mz uš, Shw
woš, Tim uš; (2)
Ghdm ekf, Ghat, Tai, Ahg əkf, Ayr əkfu, Zng yukfä; (3) Auj efk, Sml, Nfs,
Qbl əfk (*kVf and *fVk must be considered two different roots on the
Proto-Berber level); (4) Ayr ăffu (in view of əkfu, hardly assim. < *akf,
more likely <*Haff-, cf. Hgr i-s-ûf 'gift', Zng offi 'give!', imp.).

34. good
1) Ghat yulaɣ-ən, Tai yulaγ, Ahg alɔɣ (v.); (2) Ayr mol-ăn; (3) Ayr ifrar
(v.); (4) Zng
ä-gmäh; (5) Sml i-fulk-I; (6) Sml iẓil; (7) Rif rəʔdd [] No Siwa, Auj,
Sok, Ghdm, Izd, Mz, Shw, Tim, Nfs, Qbl.

35. green
1) Siwa a-wrāɣ, Sok wårráγ; Ghdm ta-wərγ-i, arəɣ (v.); (2) Ahg te-ẓẓef-é;
(3) Tai i-dal-in,
Ahg dal-ət (v.); (4) Zng bayḏig (<*bayd/zig); (5) Sml izigziw (v.), Izd a-
zəgzaw, Rif a-zigʔzaw, a-ziza, Mz a-zizaw, Shw a-ziza, Tim, Nfs a-zizaw,
Qbl a-zəḡzaw (all < **zizaw-, cf. Ahg hewhew 'be gray', Ghat zewzew 'be
blue') [] No Auj, Ghat, Ayr.

36. hair
1) Auj a-žeb̲û, Sok zâw, Ghdm a-zaw, Zng a-ʔzbi (<*-zwi), Mz a-zaw, Nfs
zaw; (2) Ghat a-
zəǯ (-ǯ<*g - cf. Hgr azaḡ 'mane'), Nfs zugg, pl. -izûgg-en ; (3) Tai a-
mẓat, pl. a-mẓad-en, Ahg i-mẓad, Ayr ə-nẓaḍ, Qbl a-nẓad; (4) Sml a-zzär,
Izd a-zzar; (5) Shw a-zəkk-uṭ, Tim a-zəkk-uṭ [] No Siwa, Rif.

37. hand
1) Siwa fūs, Auj fuss, Sok fus, Ghdm ū-fəss, Ghat, Tai, Ahg, Ayr, Izd,
Rif, Qbl a-fus, Zng
ä-fuʔš, Sml a-fūs, Mz fas, Shw fus, Tim fūs, Nfs u-fəs.

38. head
1) Sok i-ɣof, Ghdm ī-ɣəf, Ghat i-ɣəf, Ahg e-ɣəf, Ayr e-ɣăf, Zng i-ʔf (<*i-
ɣf), Izd i-ḫf, Rif i-
ḫʔf, Shw i-ḫəf; (2) Sml a-gayyu; (3) Mz ta-bəǯna (-ǯ<*-g); (4) Tim ta-
məgn-a; (5) Qbl a-qarru.

39. hear
1) Siwa, Ghat, Tai, Mz, Qbl səl, Auj esél, Ghdm əsl, Ahg əsəl, Ayr ăslu,
Rif t-sʔra (r <*l),
Shw səll, Nfs asəl; (2) Zng ogrih 'entendre, ouïr'; (3) Sml, Izd s-fəld
[] No Sok, Tim.

40. heart
1) Siwa uli, Auj, Sok, Ghat, Ahg, Sml, Izd, Mz, Shw, Tim, Nfs, Qbl ul,
Tai ul-əm, Ayr əwəl,
Zng ällü-n (pl. of üy <*ul), Rif ur (<*ul); (2) Ghdm uǯum (<*gwV-m-?).

41. horn
1) Siwa a-ččao (<*-skaw), Auj i-šk, pl. škîw-en, Sok, Shw i-š (<*išš
<*isk), Ghdm a-škaw,
Ghat i-šək, Tai, Ahg, Ayr i-sək, Zng t-əskah, Smll i-sk, Izd i-sk, i-šš,
Rif, Mz, Nfs a-ššaw (<*-skaw), Qbl i-šš [] No Tim.

42. I
1) Siwa niš, Auj nek, Sok, Shw nəč, Ghdm nešš, Ghat, Tai, Ahg, Sml nešš,
Ayr năkk, Zng
nik, Izd, Qbl nəkk, Rif nəʔš, Mz nəš, Nfs neš, nič [] No Tim.

43. kill
1) Siwa, Ghat, Ahg, Mz, Shw, Nfs, Qbl ənɣ, Auj áněγ, Sok önγ, Ghdm ənn
(<*ənʔ <*ənɣ),
Tai anɣ, Ayr ănɣu, Zng äʔni (<*aɣni, met.), Sml, Izd nəɣ, Rif əʔnɣ [] No
Tim.

44. knee
1) Siwa, Sok, Mz, Shw fud, Auj, Ghat, Tai, Ahg, Izd a-fud, Ghdm ū-fəd,
Ayr ə-fud, Zng
oʔffuḏ, Sml a-fūd, Rif fuḏ, Nfs u-fəd; (2) Qbl a-ḡʷəšrir [] No Tim.

45. know
1) Siwa, Ghdm, Ghat, Ahg, Ayr, Sml, Shw, Nfs əssən, Auj ššen, Zng əssin,
Izd isin, Rif
ʔssn, ʔssəʔn, Mz əssən, sin, Tim sən, Qbl issin; (2) Sok âgiz [] No Tai.

46. leaf
1) Auj tě-lebbîš-t; (2) Ghdm tə-fra, Ghat, Ahg, Sml a-fraw, Tai ti-fraw,
Izd i-fər, Mz a-friw,
Tim ti-friy-t, Nfs ta-fri-t, Qbl i-fərr; (3) Ayr a-fărkok, Rif ṯi-fāš-ṯ
(<*-fark-); (4) Zng aʔǯə̥ḣ (<*ala, cf. Ayr ăla 'feuille minuscule') [] No
Siwa, Sok. Shw.

47. lie
1) Siwa, Mz, Nfs əṭṭəs, Ghdm, Ghat eṭṭəs, Zng tidẓ̲̌ih (ẓ̲̌ < ẓ < ṣ < *s
assim. to * ṭ; the same
root as in yuḏəš 'fait de dormir, sommeil': ḏ <*d < ḍ <*ṭ); (2) Tai, Ahg
əns, Ayr ănsu; (3) Sml gʷən, Izd gən; (4) Shw, Qbl əẓẓəl [] No Auj, Sok,
Rif, Tim.

48. liver
1) Siwa, Tim t-sa, Auj tí-sī, Sok t-sā, Ghdm tū-sa, Ayr tă-sa, Sml tǟ-sa,
Izd ta-sa, Rif ṯ-sa,
Mz ta-sa, Nfs tu-sa, Qbl ṯa-sa; (2) Ghat, Tai, Ahg a-wsa, Zng a-ʔšä, Shw
a-ssa.[88]

49. long
1) Auj ɣázzef, Zng ō-ʔzzəf (<*u-ɣzəf), Sml ɣzīf, Izd a-ɣuzzaf, Qbl a-
ɣʷəzf-an; (2) Ghdm
zəgr-ət (v.), Ghat zəǯr-in, Tai heǯer-en, Ahg həḡr-ət (h <*z), Ayr zăgr-
at, Rif a-zirāā̄ (< *a-zigrar), Mz a-zǯərar, Shw a-zirar, Nfs a-zəgrār []
No Siwa, Sok, Tim.

50. louse
1) Siwa ta-lšə-tt (š<*k), Auj ṭa-wllək-t, Sok ti-rši-t (r<*l), Ghdm ta-
llək-t, Ghat či-llik, Tai,
Ahg ti-llik, Ayr ty-əllək, Zng ti-llik-t, Sml ti-lkī-t, Rif ṯi-šši-ṯ (-šš-
< -lk-), Mz ti-lli-t, Shw ṯi-li-ṯ, Nfs ti-lši-t, Qbl ṯi-lḵi-ṯ; (2) Izd
ta-ḫḫuy-t [] No Tim.

51. man
1) Siwa a-oggid, Ghdm weggīd, Tim uggid; (2) Auj a-médẹn, pl. mẹdîn-ẹn;
(3) Sokna mār;
(4) Ghdm ūtəm 'mâle', pl. wətm-ān; (5) Ghat, Tai, Ahg aləs, Ayr eləs; (6)
Ghat, Tai ai, pl. ei-an; (7) Zng īǯ, iš, pl. iǯǯən-ən (perhaps, < *iyy,
then to score with 6, or <*igg, cf. closely related Tetserret əgg, pl.
əggay-an 'homme, époux', then < Afrasian *gaw(V)y- '(member of) a
community, group (sharing the same territory'); (8) Sml a-rgäz, Izd, Mz,
Shw, Nfs, Qbl a-rgaz, Rif a-ā̄yaz.

52. many
1) Siwa kum 'many, all'; (2) Siwa waǯ-in (-ǯ<*g), Auj ěgg-ût, Sok ugg-út,
Ghat a-iǯ-in, Tai
a-îǯ-en, Ahg a-yəḡḡ-în, Ayr a-yăgg-ên, Sml kīg-ān, gīg-ān; (3) Ghdm hāl,
hāl-ən, Tai, Ahg hull-an, Ayr hullən, wəllen; (4) Ghdm kill-an; (5) Zng ä-
gūlləž̲, a-gulliš (most likely, < *-gulil); (6) Rif, Tim, Qbl a-ṭas; (7)
Mz mə-nnaw-t; (8) Shw a-nu-kad; (9) Nfs ə-rḫa-n [] No Izd.

53. meat
1) Siwa, Sok a-ksūm, Auj ksûm, Ghdm a-ksəm (pl. īsan), Izd, Shw a-ksum,
Rif a-šsum, a-
ysum, Mz a-ysum, Tim a-ysūm, Nfs u-səm, Qbl a-ḵsum; (2) Ghat, Tai, Ahg,
Ayr i-san (all < *i-sa-n, pl.); (3) Zng t-fiʔžži (<*-fiyy), Sml ti-fiyi.

54. moon
1) Siwa, Mz, Tim ta-ziri, Auj te-žér (a-yûr 'luna nuova'), Sok ta-zîri,
Ghdm ta-zīri 'plein
lune, lumière de la lune' (ū-yer 'lune, croissant de lune, mois
lunaire'), Zng ä-ʔž(ž)ər; (2) Ghat a-yur, Tai e-îor, Ahg e-ôr, Sml a-
yyūr, Rif yuā̄ (<*yur), Shw a-yyur, Nfs u-yer, Qbl a-ggur (<*a-yyur); (3)
Ayr t-əlli-t [] No Izd.

55.mountain
(1) Siwa, Sml a-drär, Ghdm a-durar, Ghat, Tai, Ahg a-drar, Rif a-ḏraǟ,
Shw, Qbl a-ḏrar, Nfs drar; (2) Sok a-m'rru, pl. i-měrr-ân; (3) Tai ta-hor-
t; (4) Ayr a-dɣaɣ; (5) Ahg u-ẓẓaf; (6) Mz a-wrir [] No Auj, Zng, Izd,
Tim.

56. mouth
1) Siwa a-mbu; (2) Auj a-m, Sok i-mî, Ghdm a-mi, Ghat, Ayr, Mz, Tim, Nfs
i-mi, Tai, Ahg
e-mi, Zng i-mmi, Sml i-mi, Shw i-mi, ti-mmi-š-t, Qbl i-mi, ti-mmi-š-t;
(3) Izd a-qmu, Mz a-qmum.

57. name
1) Ghdm, Ayr, Izd, Mz, Nfs ism, Ghat, Ahg, Sml, Tim, Qbl isəm, Tai asam,
Zng ašm [] No
Siwa, Auj, Sok, Rif, Shw.

58. neck
1) Siwa ta-məǯa (<*-mVga); (2) Auj a-gǻråṭ, Zng ə-gärḍ̲, Sml, Izd a-m-
gəṛḍ, Qbl a-m-ḡərḏ̣;
(3) Sok ta-krûm-t, Ghdm ta-kurəm-t, Mz ta-šərum-t (š <*k), Shw ṯa-krum-ṯ,
Tim ta-krum-t, Nfs tu-krim-t; (4) Ghat, Ayr, Rif iri, Ahg êri [] No Tai.

59. new
1) Siwa a-trär, Auj a-trâr, Sok trir, Shw a-ṯrar, Nfs a-trar; (2) Tai i-
naî, Ahg i-nay, Ayr iynay
(v.), Zng ə-yən [] No Ghdm, Ghat, Sml, Izd, Rif, Mz, Tim, Qbl.

60. night
1) Siwa iṭ/ḍ, Auj á-b̲oḍ, Sok iḍi, Ghdm ī-b̲eḍ, Ghat i-həḍ, Tai, Ahg e-hoḍ,
Ayr e-hăḍ, Zng
īḏ̣, Sml īḍ, Izd iḍ, Mz (dəǧ-)iḍ, Shw yəṭ, iṭ, Tim degg-iḍ/d, Nfs yəḍ, Qbl
yəḍ̲ [] No Rif.

61. nose
1) Siwa ta-nzär-t, Auj, Nfs tə-nzər-t, Sok tu-nzär-t, Ghdm tə-nzar-t, Ayr
tyə-nžar-t (ž <*z),
Zng tī-nžär-ən, Izd a-nzar, Rif ṯi-nzaā̄-ṯ, Mz ti-nzər-t, Shw ṯi-nzar, Tim
ti-nzaḥ-t, Qbl a-nzar-ən; (2) Ghat, Tai a-nǯur (-ǯ- <*g), Ahg a-ñḡur; (3)
Shw a-ḫənfuf [] No Sml.

62. not
1) Siwa ula, Sok yul (pres.), Mz, Nfs ul; (2) Sok ngi (past); (3) Ghdm
ak; (4) Ghdm awas;
(5) Tai, Ahg, Zng, Sml, Izd, Shw ur, Ayr wər, Rif wā̄ (<*war), Tim wa <
war, Qbl ur, wər [] No Auj, Ghat.

63. one
1) Siwa éǯǯ-ən (<*iyy-), Auj iwî-n, Sok iǯǯ-ən, Ghdm yū-n, Ghat îyy-ən,
Tai, Sml ya-n,
Ahg yə-n, iyyä-n, Ayr iyă-n, Zng yuʔ-n, əyyuʔ-n, Izd yuʔ-n, əyyuʔ-n, Rif
ižž, ižž-ən, Mz igg-ən (-gg- <*-yy-), Shw iš-t, iǯw-əṭ, yow-ṭ, Tim igg-
ən, Nfs uǯu-n, Qbl yiw-ən.

64. person (human being)
1) Auj a-regâz, Ahg a-raḡah, Zng ä-räġäž̲ 'homme (quelconque), personne,
individu'; (2)
Sok uggîd, (3) Ghdm, Ghat, Tai, Rif, Mz, Nfs, Qbl i-man; (4) Ayr eləs []
No Siwa, Sml, Izd, Shw, Tim.

65. rain
1) Siwa a-nzār, a-nẓâr, a-mẓâr, Sok a-mẓâr (m <*n), Ghdm a-naẓar, Sml a-
nẓar, Izd a-nzar,
Rif a-mzaā̄, a-nzaā̄, Nfs a-nzer; (2) Auj ab̲ûn; (3) Ghat, Tai a-ǯənna, Ahg
a-əḡənna, Ayr a-gənna, Mz ta-ǯənw-it, Tim a-žənna; (4) Zng a-kkanak; (5)
Shw ṯa-məṯna; (6) Qbl a-ḡəffur.

66. red
1) Siwa a-zəggâɣ (-gg-<*ww), Auj a-zwâɣ, nə-zwaɣ, Sok zuggöγ, Ghdm a-
zəggaɣ, Ghat i-
zəǯǯaɣ-ən (pl.), Tai haggaγ-en, Ahg ihwaɣ (<*zwaɣ, v.), Ayr a-mi-zwăɣ,
Zng žobbä (<*zuwwa <*zuwwaɣ), Sml izwiɣ (v.), Izd a-zuggwaɣ, Rif a-
zuggʷaɣ, Mz, Shw a-zuggaɣ, Tim a-zəggāɣ, Nfs a-zəggwaɣ, Qbl a-zəggʷaɣ;
(2) Tai a-haggar.

67. road
1) Auj ta-baru-ṭ, te-bārû-t, Sok t-barut-t (otherwise <*-barud-t), Ayr a-
bər, ta-bəră-t; (2)
Ghdm a-brīd, Ghat a-barid, Izd, Mz a-brid, Rif, Qbl a-b̲riḏ, Shw a-briḏ,
Nfs brid; (3) Ghat a-brəqq-a, Tai a-bereqq-a, Ahg a-barəqq-a; (4) Sml a-
ɣaras, Zng t-urəs (<*tu-ʔrəs <*-u-ɣrəs); (5) Tim ti-zəmm-ət [] No Siwa.

68. root
1) Auj t-werž-ît (met. <*-z/ẓur unless from an Arab. dial. where warīz-at-
'vein' got the
meaning 'root' after the current meaning shift pattern), Ghdm, Sml, Izd,
Mz a-ẓur, Rif a-zwar, Shw a-zwər, pl. i-ẓur-an, Nfs a-zur, Qbl i-ẓur-an
(pl.); (2) Ghat, Tai i-dir; (3) Ghat i-kiw, Tai i-kiw-en, Ahg e-ke, Ayr e-
kăy [] No Siwa, Sok, Zng, Tim.

69. round
1) Siwa yu-brim-a (no etymology, perhaps a lw); (2) Ghdm kəwwər (v.), Tai
i-krirî, Qbl a-
kʷər-nənn-ay; (3) Ghat ǯiləllaw-ət (v.), Ahg ḡələll-ət (v.), Ayr gələll-
ət (v.); (4) Sml ta-zayyer-t; (5) Shw i-nnəḏ̣; Nfs i-me-gergeb [] No Auj,
Sok, Zng, Izd, Rif, Mz, Tim.

70. sand
1) Auj hə-mlal, Sok u-mlâl, Ghdm ta-mallil-t, Sml a-mlǟl (3) Ghat i-dehi,
Tai e-dehî, Ahg e-
dəhi; (4) Ghat i-ǯidi, Tai a-žedi (ž <*g), Rif i-ž'ḏi (<*-gdi-), Mz i-
žədi, Shw, Qbl i-ždi, Nfs židi; (5) Ayr ə-zezəl (e-dəhi 'colline', ə-gidi
'dune de sable', te-măsăs-t 'sable fin'); (6) Izd a-lḵiḵ; (7) Mz a-bərda,
(8) Tim šāl (<*kal, see 'earth') [] No Siwa, Zng.

71. say
1) Siwa umm; (2) Auj an, Sok, Ghat ən, Ghdm en, (past) inna, Tai in, Ahg
ənn, Ayr ănnu,
Zng yənnäh, Sml, Izd, Rif, Mz, Tim, Qbl ini, Shw inna; (3) Tai aməl, Nfs
əməl.

72. see
1) Siwa ẓər, Sok ẓorr, Zng ẓẓuʔr, Sml, Mz, Shw Tim, Nfs zər, Rif zr,
zəʔr, zā̄; (2) Auj a-gél;
(3) Ghdm əlləm (əll-əm, derived < wəl 'eye'), Qbl wali; (4) Ghat, Tai,
Ayr, Ahg əni (habitative hânney), Izd inni.

73. seed
1) Siwa a-ɣram; (2) Auj to-γwâu; (3) Sok a-kráz; (4) Ghdm a-mənd-i, Sml a-
mūd, Izd a-
mud, Tim i-məndi; (5) Ghat či-fəs-t, Tai ti-fes-t, Ahg tê-fes-t, Izd i-
fs, Mz a-i-fs, Nfs a-y-fs; (6) Ayr ă-masa; (7) Zng (ä-š)šunḍay-än (pl.);
(8) Mz t-izn-in [] No Rif, Shw, Qbl.

74. sit
1) Siwa, Ghdm, Rif, Mz, Shw, Nfs qim, Auj qīm, Ghat əqqim, Ahg ɣaym, Zng
imi (<*iʔmi
<*iɣmi), Qbl qqim; (2) Ayr găwăr, Sml ggawr; (3) Sml s-kkiws; (4) Izd
ɣužd-əm [] No Sok, Tai, Tim.

75. skin
1) Siwa, Ghdm, Sml īləm, Sok, Ghat, Tai iləm, Ahg eləm, Ayr elăm, Zng a-
yim, pl. ällamm-
ün, Rif irm (<*ilm); (2) Auj e-glim, Mz a-ǯlim (<*-glim), Shw a-glim, Nfs
u-glim, Qbl a-ḡʷlim; (3) Ayr e-găšek [] No Izd, Tim.

76. sleep (v.)
1) Siwa niddum, Ghdm nəddəm; (2) Auj iš; (3) Sok öṭṭös, Ghat, Tai, Ahg,
Mz, Shw, Nfs,
Qbl əṭṭəs, Sml ṭṭəṣ, Rif ʔṭṭəʔs; (4) Ayr ănsu; (5) Zng ṣummih; (6) Izd
gən, Shw gwən, Qbl ḡən [] No Tim.

77. small
1) Siwa a-kik, Auj a-kett-ût; (2) Auj me-ššék, Nfs mə-šək; (3) Sok mě-ẓî
(<*ẓVy<*ẓVg),
Zng mā-ẓẓūg, Sml im-ẓig (v.), Izd a-mə-ẓy-an, Rif a-mʔ-zzy-an, Mz a-m-zy-
an, Shw a-m-ẓy-an, Tim a-mə-zi-an; (4) Ghdm m-əṭṭid 'ê. petit, jeune';
(5) Ghdm əsdid 'ê. mince,menu, petit', Gjat, Ahg mə-ḍri, Tai an-ḍer-en,
Ayr a-mə-ḍray; (6) Tai en-tukk-en; (7) Qbl a-mə-žṭuḥ.

78. smoke
1) Auj a-b̲û, Ghdm u-b̲u, Ghat, Tai, Ahg a-hu, Ayr ə-hu, Zng ă-oʔbi, Smll,
Izd a-ggu
(<*a-wwu), Qbl a-bbu [] No Siwa, Sok, Rif, Mz, Shw, Tim, Nfs.

79. stand
1) Siwa əbdəd, bidda, Ghdm əb̲dəd, Ghat, Ahg, Ayr əbdəd, Tai ebeded, Zng
ăwḏəḏ
(<*abdəd), Sml, Izd, Shw bədd, Rif, Qbl, Tim b̲ədd 'se lever', Nfs udəd,
ewded; (2) Auj qīm (rather related with Berb. *ḳaym 'sit' than borrowed
from Arab ḳwm 'stand'; note also -y- vs. -w- in Arab.); (3) Sok še ti-škâ-
n '(be) on feet'.

80. star
1) Siwa irri, iri, Sok iri, pl. irây-en, Ghdm iri; (2) Ghat, Sml, Tim,
Nfs i-tri, Mz, Ahg, Ayr a
tri, Tai a-trî, Zng a-ḏari (-ḏ- <*t), Rif, Shw, Qbl i-ṯri; (3) Zng a-
wakkad-i [] No Auj, Izd.

81. stone
1) Siwa t-ɣoɣ-ət, Sok t-ɣaɣi-t, Nfs t-ɣaɣa-ṭ (ṭ <*t); (2) Siwa, Mz a-
dɣaɣ, Tim a-dɣa, Qbl a-
ḏɣaɣ; (3) Auj a-zérg (very likely < *-z-erg), Ghdm ī-rəg, Zng t-iʔrg-ət;
(4) Ghat, Ahg a-blal, Tai a-belal (Ayr a-blal 'pierre dur'); (5) Tai ta-
hun-t, Ayr tə-win-t (Ahg tə-hun-'large stone'); (6) Ahg é-kadé, Ayr e-
kăde; (7) Sml, Izd, Rif a-zru, Shw a-ẓru.

82. sun
1) Siwa t-fūk-t, Sok t-fuk-t, Ghat, Ahg ta-fuk, Tai, Sml ta-fuk-t, Ayr tə-
fuk, Zng tə-ʔffuk-t,
Izd ta-fuk-t, ta-fuš-t, Rif ṯ-ʔfuš-ṯ; (2) Auj tā-fû-t, Mz t-fuy-t, Shw ṯa-
fu-ṯ, Tim t-fuy-t, Nfs tu-fu-t (Ghdm tū-fə-t 'la lumière, l'éclat du
soleil'); (3) Ghdm īsi ('disque solaire'); (4) Qbl iṭiž.

83. swim
1) Siwa sîyəf, Izd ššəf; (2) Tai wot aman, Ahg əzzi âman ('to know
water') [] No Auj, Sok,
Ghdm, Ghat, Ayr, Zng, Sml, Rif, Mz, Shw, Tim, Nfs, Qbl.

84. tail
1) Auj a-zîf; (2) Sok ta-ma-qqal-t, Sml ti-m-ɣil-t (met. of Ahg e-mə-
llawɣ not to be ruled
out); (3) Ghdm ta-baḥṣuṣṣ (ḥ is expected to point to an Arabism, but may
be purely expressive; besides, I failed to find the source term in
Arabic); (4) Ghat a-rǯal, Ahg a-rḡal; (5) Tai ta-basus-t (met. of Ayr a-
sba not to be ruled out); (6) Ahg e-mə-llawɣ, Tim ta-mə-lqal-t (otherwise
rel. to 2); (7) Ayr ta-sba-t; (8) Ayr tə-lănkaw-t; (9) Zng tə-ššuḍḍ-ān;
(10) Izd a-bəṛdud; (11) Mz ta-zanḍ-in-t; (12) Nfs a-fettal (less likely <
Arab. fatīl- 'plaited, twisted; wick' or fatl-at- 'thread'; this
triconsonantal root is Proto-Afrasian); (13) Qbl a-ẓəʕkuk [] No Siwa,
Rif, Shw.

85. that
1) Siwa da-wok, wôk (m.), t-ôk (f.) [*-k]; (2) Auj íwan, ídīn [*-n], Ahg
wîn [*-n], Zng an, in
[*-n], Sml ɣwann [*-n], Izd aynna [*-n-], Rif ayənni [*-n-], Mz ənni [*-n-
], Shw win [*-n]; (3) Sok dâwa, udînak [*d-], Ghat dindəɣ [*d], Tai
wandin (m.), tandin (f.) [*-d-], Ayr wendaγ (m.), tendaγ (f.) [*-d-]; (4)
Ghdm wi (m.), ti (f.) [*-i/y]; (5) Nfs wīh, wîha [*-h], Qbl wihin *[-h-]
[] No Tim.

86. this
1) Siwa waya (m.), taya (f.) [*-ay], Auj wāya (m.), tâya (f.) [*-ay-],
Sok wā (m.). tā (f.) [-ā,
prob. <*aya], Izd, Rif aya, Shw ay, ayyi, Tim ay; (2) Ghdm wu (m.), tu
(f.) [*-u/w], Sml ɣwa [*-w], Mz u [*-u/w], Nfs wūh, wûha [*w-]; (3) Ghat
adəy [*-d], Zng ayḏ, aḏ [*-d]; (4) Tai wah (m.), tah (f.) [*-h], Ahg wah
[*-h]; (5) Ayr wina (m.), tina (f.) [*-n]; (6) Qbl wagi [*-g].

87. thou
1) Siwa šək (<*kək), Auj ku, kem, Sok šik, Ghdm šegg (<*keyy?), Ghat,
Tai, Ahg kay, Ayr
kăyy, Zng kuk, Sml kiy-ĩn, Izd ki, Rif šək, Mz šəč (<*kək), Shw, Tim
šəkk, Nfs šək, čək, Qbl ḵəčč.

88. tongue
1) Siwa, Ghdm, Ghat, Tai, Ahg, Ayr, Mz, Shw, Tim, Nfs, Qbl îləs, Auj,
Sok, Sml, Izd ils,
Zng əčsi (<*əlsi), Rif irʔs.

89. tooth
1) Siwa a-sayn, Auj a-sîn, Sok i-sîn, Ghdm a-sīn, Ghat, Tai, Ahg i-sin,
Ayr e-šen, Nfs sīn;
2) Zng ōkši, Sml a-ḵus; (3) Zng a-ʔräš (<*-ḳraš); (4) Izd, Mz ti-ɣməs-t,
Rif ṯi-ɣməʔs-ṯ, Shw
ṯi-ɣməs, Qbl tu-ɣməs-t; (5) Tim u-gəl.

90. tree
1) Sok a-sγâr, Tai tə-šɣar, Zng a-ššaʔr (<*-šaḳr); (2) Ghdm te-mdik-t;
(3) Ghat e-məɣ-t; (4)
Ghat, Tai, Ahg a-šek, Ayr e-šək; (5) Sml tä-ddä́g-et; (6) Izd a-s-klu; (7)
Rif a-zəmmur; (8) Shw ṯa-seṭa [] No Siwa, Auj, Mz, Tim, Nfs, Qbl.

91. two
1) Siwa, Sok, Ghdm, Mz, Shw, Tim, Nfs sən, Ghat, Izd, Sml, Qbl sin, Tai i-
ssin, Ahg ə-ssîn,
Ayr e-ššin, Zng šin-an, Sml sîn [] No Auj.

92. walk
1) Siwa ukəl, (2) Auj žî-gež, Sml zzi-gīz; (3) Sok šè tiškân ('go by
foot'? ti-škâ-n 'foot'); (4)
Sok mad, i-mmádda; (5) Ghdm z-bərdəz; (6) Ghdm rar 'passer, aller'; (7)
Ghat, Tai ek ; (8) Ghat ərǯəz (<*-rgəz), Ahg rəḡəh (-h <*z), Ayr răgăz
'marcher au pas', Zng iž̲-rigiž̲ 'aller à pied, marcher à pied'; (9) Ahg
egel; (10) Sml ɛ̌ddu, Izd, Qbl ddu; (11) Rif užuā̄ (<*ugur), Mz a-ǯur (ǯ
<*g), Shw uyyir, uggir, Tim igu (< igur) 'aller', Nfs ugur.

93. warm
1) Siwa əqqəd (v.); (2) Ghdm ze-qqāl, Rif a-z-ɣar (-r <*-l), Shw a-zɣal,
Nfs i-zeγel, Qbl
i-zɣil (v.); (3) Ghat, Tai ikkus-ən, Ahg, Ayr ukas (v.); (4) Zng a-rɣah
'chaud', Sml irɣ-an, Izd ərɣ, rəɣ (v.) [] No Auj, Sok, Mz, Tim.

94. water
1) Siwa, Ghdm, Ghat, Tai, Ayr, Zng, Sml, Izd, Rif, Mz, Shw, Qbl ama-n
(pl.), Auj imi-n,
Sok amâ-n, Ahg âma-n, Tim amā-n, Nfs am-ən.

95. we
1) Siwa nčini (-č <*k), Auj, Ayr nekkenî, Sok əččəni (əčč- <*ənč-), Ghdm
nekkenīn, Ghat,
Tai nəkkən-iḍ, Ahg nəkkən-iḍ, Zng nəkini, Sml nəkni, Izd nəkkin, nukni,
Rif nəʔššin, Mz, Tim nəšni, Shw nəčənin, Nfs nəčən, nešehen, Qbl nəkʷni.

96. what?
1) Ghdm, Rif mi, Ghat, Tai, Ahg, Ayr, Zng, Sml, Nfs ma, Izd may, Mz mama,
matta, Shw
matsa, Tim [] No Siwa, Auj, Sok, Qbl.

97. white
1) Siwa, Sml, Tim a-məllāl, Auj mill, Sok, Ghdm, Nfs melləl, Ghat məllul
(v.), Tai, Izd, Qbl
a-məllal, Shw, Ahg tə-məlle, Ayr imlal (v.), Zng molliǯ (-ǯ <*l), Rif a-š-
ʔmrar (<*-mlal), Mz d a-məllal.

98. who?
1) Auj mânī, Rif man, Mz manay-n, Shw mən; (2) Ghdm annu; (3) Ghat, Ahg,
Ayr mi, Zng
mə-ḏ (ḏ <*d <*t), Sml, Izd ma, Nfs mammu; (4) Tai wa, Qbl wi [] No Siwa,
Sokna, Tim.

99. woman
1) Siwa ta-lti, Ghdm ta-lta; (2) Auj tẹ-mí-gnī (ta-lta-n pl.); (3) Sok ta-
mětt-úṭ (Ghdm ta-meṭ
'femelle'), Ghat, Ahg ta-məṭ, Tai ṭa-məṭṭ, Ayr ta-mṭəṭ, Izd, Mz ta-məṭṭ-
ut, Shw ṯa-mṭṭ-uṯ, Qbl ṯa-məṭṭuṭ; (4) Sml ta-mɣar-t, Rif a-mɣar [] No
Zng, Tim, Nfs.

100. yellow
(1) Ghat a-ɣar (cf. Ahg a-ɣarirwaɣ); (2) Tai î-eraγ-en, Ahg, Ayr irwaɣ
(v.), Zng yərəʔ (<*yərəɣ), Sml, Izd, Mz, Shw, Nfs, Qbl a-wraɣ, Tim a-wra
[] No Siwa, Auj, Sok, Ghdm, Rif.

Appendix 2a
Differences in choice of words and scoring/etymological identification in
the North Tuareg 100-wordlist between V. Blažek (B)[89] and the author
(M)[90]:

4. belly: Ghat ta-ǯahu-t scored by B, excluded by M as a loan from a
Tamâhaq dialect
where *z > h (cf. Ayr ta-gəzu-t 'panse').
11. breast: Ghat, Tai a-dmar, Ahg i-dmar-en (scored by B for all three
languages); not scored
for Ahg by M because it is an animal's breast rather than a general term: i-
dmâr-en 'partie basse de la poitrine chez les quadrupèdes'.
14. cloud: for Ahg, B chose a-ḡənna also meaning 'ciel' and 'pluie' (<
Proto-Berber *a-
gVnna 'sky'), while M preferred a-ẓiar 'petit nuage isolé' related to Tai a-
ẓiar, the main term for 'cloud'.
23. eat: Tai and Ahg əkš scored by B with Ghat əč, while M scored
them differently on the following grounds: both Tai and Ahg əkš mean 'eat'
and 'bite' < Afrasian *(nV-)kič- 'to bite': Semitic (Aramaic and
Ethiosemitic) *nkṯ 'bite'; C. Chadic: Malgwa kúča 'bite off, tear off' (in
[Stolbova 2011 #146] *kʷVč- 'to pull (with force), tear off'; compared with
the Semitic, Ahaggar and Sidamo verbs); E. Cushitic: Sidamo kis- 'to bite'.

As for Ghat əč, it is < ət- š, with -š, a verbal suffix (cf.
different verbal forms of the verb 'to eat' in Ayr: ăttyu and ătt-šu,
Zenaga yətättä and yitY-šä, Izdeg t-š and the Common Berber "habitative"
form of 'eat': *tatt) < Afrasian *tVʔ- 'to eat': Semitic: Akkadian taʔu 'to
eat, graze', Soqotri teʔ 'to eat', etc.; Chadic *ti/aw/ʔ- 'to eat (soft
food)'; Cushitic C.: Beja tiyu 'to eat'; N. Omotic: Yemsa taʔ-, Gimirra
(Bench) tyaʔ-, S.: Dime ɛts-o, et(s), ʔič-in, Ari it-s-, ič-, Hamer it-s-a
'eat' (<*ʔit-s, with the same -s/š suffix as in Berber).
26. fat: Ahg ésîm is included by B (and scored with Tai isîm and
Ghat isim), but not listed by M because of its specific meaning 'graisse
fondu'.
28. fire: Ghat ə-fiw is scored by M, but not included by B.
29. leaf: Ahg éla (of excellent Afrasian etymology) is included by
B, but not scored by M because of its specific meaning 'feuille minuscule
d'alser'.
51. man: B scores Ghat ai and Tai eî with Ahg éi meaning 'mâle, de
sexe masc.' not scored by M who gives only ales for 'man'; B also includes
'men, people' for all three languages (<*midd-an) while M does not as
inclusion of suppletive plural forms in addition to corresponding nouns in
singular does not, from his point of view, comply with the requirements of
lexicostatistics.
52. many: besides the forms derived < *a-yg(g)-in in all the tree
languages, M scores a synonymous form hull-an 'many' recorded in Tai and
Ahg (not in B).
64. person: besides the forms derived from *iman scored by both B
and M for all three languages, B scores the following synonyms in Ahg:
âləs, middan, iyän and, in Ahg ǎw-âdəm and Ghat aw adem (marking the second
element as an Arabism). M scores neither of these synonyms on the following
grounds: aləs is 'man' rather than 'person'; midd-an 'people' is plural;
iyän derives from iyä-n 'one' in masc. (with fem. iyä-t) and conveys the
idea of a male individuum (glossed 'individualité' by Foucauld) rather than
man as a human being implied by 'person' vs. 'man' in the Swadesh's list;
in ǎw-âdəm and aw adem the semantic stress is on the second element, an
Arabic loan, to be ruled out of the list. Instead, M scores Ahg a-raḡah
(<*a-ragaz, cf. Zng ä-räġäž̲ 'homme (quelconque), personne, individu').
66. red: while both B and M score Tai ihaggaγen with Ghat zəǯǯaɣ
(<*zəwwaɣ) and Ahg ihwaɣ (<*-zwaɣ), for Tai M also scores a synonym a-
haggar < *a-zawwar (equal to Kel-Ui and Wargla a-zawwar, etc.) unrelated to
*zəwwaɣ.
75. skin: In addition to the general term *ilm common to all three
languages, B scorres Ahg aqqər which has a specialized meaning 'peau non-
tanné' and as such is not scored by M.
81. stone: B scores Tai i-kedi 'rocher' with Ahg ékadé; M rules out
the Tai word for a slightly modified meaning (the common Tai term for
'stone' is a-belal scored for all three languages by both authors). B also
scores Ahg tə-hun-t with Tai ta-hun-t both glossed as 'big stone' which M
considers a specified meaning and does not scare as the third synonym.
84. tail: For Tai, M quotes ta-basus-t; no term in B.
90. tree: Besides the terms derived from *a-hVšk- in all three
languages, M quotes (1)
Ghat e-məɣt and (2) Tai tə-šɣar, absent in B.
92. walk: B scores the terms derived from *ikk for all three
languages (in M, Ghat ekk is not scored as its meaning is rather 'go to'
than 'walk, go') and, for Ahg, the following synonymous verbs: demer,
eḵlek, eḡ and rəḡəh. While rəḡəh, indeed, conveys the meaning of 'walk'
('marcher au pas' and 'promener') and is scored by M with Ghat ərǯəz (not
in B), ekk, demer and eḡ all mean 'aller à, chez' and are not scored by M
(as well as eḵlek/eḫlek which is an Arabism) scoring, instead, Ahg egel
'marcher' as a synonym.
99. woman: B scores, besides the singular, the plural forms for all
three languages,
whereasM scores only the singular ones.
100. yellow: besides the forms derived from a common Berber *wVraɣ
~ *rVwaḳ (< North Afrasian *w/yVraḳ-), there is Ghat a-ɣar (considered by B
a metathesis and scored with Tai î-eraγ-en and Ahg i-rwaɣ). M prefers to
regard Ghat a-ɣar as unrelated and scores it differently.

As a result, according to B, Tai and Ahg have 97.4% of common scores
corresponding, after Starostin's formula, to split date 1,440; and Ghat vs.
Tai-Ahg has 92.6%, corresponding to split date 1,050.
According to M, Tai and Ahg have 92% of common scores corresponding,
by the same formula, to split date 1,010; and Ghat vs. Tai-Ahg, 86%,
corresponding to proto-Tamâhaq split date 680.[91]

Appendix 3

Genealogical Tree of Berber Languages (based on lexicostatistics of 17
languages)

Note that (1) the Proto-Berber split in the 11th cent. BCE and the
separation of Proto-Tuareg identifiable wth Garamantes coincides with the
proposed time of the latter's appearance on the historical scene;[92] (2)
the split of Proto-North Berber in the 5th cent. BCE into Proto-Atlas
group[93] and the rest of North Berber corresponds to the spread of
Libyan tribes along the South Mediterranean area ("the Libyan Coast")
recorded in the same century by Herodotus who referred, inter alia, to
the Atlantes living near Mt. Atlas. The Atlantes lived near Mt. Atlas
The Atlantes lived near Mt. Atlas
Appendix 4

The Berber-Phoenician case

Let us analyze the examples adduced in [Malášková-Blažek], one by one:

A. Berber forms of Phoenician/Punic origin with the epigraphic or literary
documentation of sources:

1. Berber *ā-b[i]ẓālīm[94] (> aẓalim) ~ Neo-Punic bṣl 'onions' (also
Hebrew bāṣāl, pl./coll.
bəṣālīm).
Though the borrowing is obvious, in order to keep to the facts, the
borrowed forms should be
reconstructed in Berber as *ā-ẓālīm without -*b- appearing only in later
loanwords from Arabic which leaves the problem of loss of -b- in Berber
unsettled (even more so as this loss is unique[95]). The only explanation I
can think of is a source Punic form hā-b̲ṣəllīm (if such form is possible)
sounding to the Berber/Libyan's ear as [awṣalīm] > āẓālīm. However, judging
by Greek and Latin rendering of Punic plosives in a postvocal position,
their spirantization started not earlier than the 2nd century BCE
[Lyavdansky, 283], which corresponds to the Neo-Punic attestation of bṣl.
If this explanation holds water, the words of the aẓalim type could not
have been borrowed in Proto-Berber, but were independently borrowed from
later Punic or Hebrew in individual – only North, at that[96] – Berber
languages (or into one language then spreading to the others). All other
forms are Arabisms except the South Tuareg ones of the *ā-mẓəlli type
'onion' which should likely be explained as a metathesis of *ā-ẓālīm.

2. Berber (all branches) *ā-gādīr, pl. *ī-gūdār 'wall' ~ Neo-Punic gdr
'wall' (also Hebrew
gādēr 'dry-stone wall' < Sem. *gadīr-/ *gidar- 'stone wall/structure').
It is hard to believe that the word mentioned in one Neo-Punic
inscription corresponding to Latin murum[97] became source of a common
Berber term with such diverse meanings beside 'wall' in individual
languages as 'cloison de terre' (Siwa ǯadir), 'banquette' (Ghadames u-
ǯadir), 'ciment' and 'cimetière' (Ayr ǎ-gédir, E.Tawllemmet ǎ-gadir),
'falaise, rocher à pic; ravin, précipice; talus' (Rif a-ždir), 'bord, rive'
(Mzab a-yder), 'falaise, rocher à pic' (Iznassen a-ždir), 'talus; terrain
en forte déclivité; escarpement; monticule' (Qabyle a-gadir).[98]
In spite of similarity in the vocalic pattern, I'd rather bet on
common Semitic-Berber root with a reconstructed meaning 'assemblage or
stones, stone structure', presumably integral, with the *-Vr extension, to
the common Afrasian "root family": Afrasian *g(ʷ)Vd-: Semitic *gʷVdd- ~
*gVdgad- 'wall, house, room' (Syrian Aramaic gūd-, gudd- 'wall', Arabic
ǯudd- 'earth embankment used as boundary', Geez gädgäd 'wall, enclosure',
Amharic gʷada 'additional house in which household objects are held',
etc.); Chadic W.: Hausa gídā́ 'compound', Guruntum gìdi 'hut', C.: Glavda
gúd(à) 'room', Musgu gu(d) 'house', Zime-Batna gìḍ 'enclosure'; E.
Cushitic: Oromo godoo 'hut', Burji godoo 'cattle-pen', Harso, Dihina koto
(k regularly < *g) 'house'; perhaps, also isolated W. Berber: Zenaga ti-gid-
ith and Kwarandzyey a-gada 'wall' (more likely than from *a-gadir as
follows from Malášková-Blažek, yielding Zenaga agǝdri 'sol, parquet');
Afrasian *gadil- (with the *-Vl extension): Semitic: Arabic ǯadl-at-
'pilon'; Berber: Ghadames a-gadil 'wall', amagdūl 'forteress ruinée…au
plateau des idoles…', Ayr, E.Tawllemmet e-gadel 'screen', etc.

3. Berber (all branches) *agliH (imperative), *yuglaH (perf.), in
individual languages 'go',
'depart', 'arrive', 'move', 'accompany', 'leave', 'move camp', etc. ~
Phoenician g-l-y 'to move, remove', Hebrew gālā 'to emigrate, uncover',
Ugaritic gly 'to present oneself, make one's way, go', Aramaic-Syriac gǝlā
'to depart, uncover', Mandaic gla 'to exile, uncover', Arabic ğalā 'to
emigrate, make clear, unveil, uncover', Geez galaya 'to cut away, separate,
divide'; Aramaic > Akkadian galû 'to be deported'.
To start with, I prefer to distinguish three Semitic verbs, two of
which are very likely related while the third (West Sem. *gly/w 'uncover,
lay bare': Hebrew gly 'uncover, lay bare', Aramaic-Syriac gǝlā 'uncover',
Mandaic gla 'uncover', Arabic ǯlw 'uncover, undress oneself', Geez galgala
'lay bare, empty') is rather not .[99]
One of the two other roots – *gly has a very specific cultural meaning
'to exile, go into captivity' that could not have evolved in every language
independently and should be reconstructed on the Proto-Central Semitic
level:[100] Hebrew gālā 'go into exile', gālūt 'exile; exiles', Aramaic:
Syrian gālūtā 'captivity'. gālīlā 'captive, exile', Mandaic gla 'to exile'
(Akkadian galû 'to exile, be deported' is considered a loanword from
Aramaic), Arabic ǯly 'emigrate, leave one's country', ǯalā 'exile,
emigration', ǯāliyat- 'banishment, captivity far from one's motherland;
contribution' (Daṯina Arabic ǯalā 'emigrate'), Geez tagalgala (< ta-
galgala, reduplicated stem) 'go into captivity, be taken into exile' .
This Proto-Central Semiitic verb is likely to have stemmed from Proto-
Semitic "biconsonantal" root *gl 'to go, move': Ugaritic gly 'to present
oneself, make one's way, go', Phoenician gly 'to move, remove', Hebrew gālā
'(to have) to leave, dissapear', Syrian gǝlā 'to depart', Arabic ǯwl 'go,
run, turn', Geez galgala 'evacuate, separate', Tigrai gälälä 'move away
from a place, make way', Amharic gäläll alä 'make way, part', gäläggälä
'separate two people who are coming to blows', Jibbali ególi 'to shift
(animals) from one place to another', eglé 'to shift from one place to
another' (trans.), gútli 'to shift (from a place), change (a religion)'
(intrans.)
I see no reason to consider the above biconsonantal Proto-Berber verb
*gl with a general meaning of movement without any cultural overtones (such
verbs are usually not borrowed in Afrasian languages) a loan from Punic
instead of its natural derivation from Proto-Afrasian *gVl- 'go, run,
follow, enter': Chadic W.: Sha gal 'to run', gulay, C.: Buduma gā́là, E.:
Tobanga gə̀le 'follow', Mobu gale 'go out'; E.Cushitic: Saho gil 'run', Afar
gal 'go', Oromo gala, Somali, Arbore, Burji gal- 'enter'; N.Omotic: Ometo,
Anfillo gal- id.

4. Berber (all branches) *ē-niHir, pl. *ī-niHir-an '(oil) lamp' ~
Phoenician nr 'lamp,
candelbrum' (also Hebrew nēr 'light, small clay lamp filled with oil' <
Proto-Semitic *nū/īr 'light').
Obviously borrowed in Berber, but it is hard to determine whether it
happened in the Proto-Berber period or later in individual Berber languages
or in one language with subsequent spread to the others ("chain
borrowing"); in the two latter cases, the source word could be Punic or
Hebrew.

5. Berber: *ā-niHās or *ā-naHās 'copper' ~ Phoenician nḥšt 'bronze', Hebrew
nəḥōšät,
nəḥūšā 'copper, brass, bronze'.
The Berber terms in question, to exclude undoubted Arabisms among them,
are represented by Shilh (anas) and two East Berber languages:
Ghadames[101] unas and Sokna nās. Unless there was a related form without
-t not preserved in Phoenician sources, the unusual trait of the presumed
borrowing is lack of this suffix in the borrowed Berber words where one
would expect the form like *ta-nas-t (could then Hebrew nəḥūšā have been
the source word?)[102].

6. Berber *ā-γānīm 'reed' (North, East and West branches) ~ Punic ḳn
'reeds' (pl. in -m not
attested but quite imaginable), Hebrew ḳānä̲, pl. ḳānīm.
Clearly a loan in Berber either from Punic or Hebrew. Again, the
choice between borrowing in Proto-Berber and "chain borrowing" depends on
whose glottochronological dating – by Blažek or the present author – is
closer to the historical reality.

7. Berber *aγriH (imperative), *yuγraH (perf.) 'to call, tell, cry' and
'read, study' ~ Neo
Punic ḳrʔ 'to read', Hebrew ḳārāʔ 'to call, read, recite, invite', Arabic
ḳrʔ 'to call, read, recite, invite'.
I prefer to derive the Berber verb 'to call, tell, cry' with a very
general meaning (see a similar case in point 3) from a common Afrasian
biconsonantal root *ḳr 'to cry, speak loudly': Proto-Semitic *ḳrʔ 'call,
shout' (from which the secondary meaning 'to read' evolved in Hebrew and
Arabic,[103] whereas in Modern South Arabian it's an Arabism); Chadic W.:
Hausa ḳā́rā́ 'exclaiming, crying out, a noise, a sound', Karekare kàarùwàa
'cries of praise, ritual speech', C.: Zime-Dari kárárá 'bruit, cri',
Dghwede kwàrayà 'speak', E.: W.Dangla kòrìyà 'cri percant', etc.;
E.Cushitic: Oromo ḳorḳoro 'ask', Darasa ḳor- 'ask (in investigation)' .
The meanings 'to read' and 'to study' are due to influence rather than
borrowing – and from Arabic at least no less likely than from Punic or
Hebrew.

8. Berber: Ahaggar γärät 'être comme il faut, être poli, bien élévé et
distingué de manières'
~ Phoenician ḳrt ʽcity', Hebrew ḳärät (and Arabic ḳaryat-).
A good example of borrowing (from any of these three Semitic
languages) with an attractive further semantic development, however,
attested in one Berber language only and therefore irrelevant for the
relative dating of Proto-Berber split.

9. Berber (all branches) *ā-γ[i]ssīm'cucumber, small melon' ~ Punic *ḳš'
'cucumber'
(reconstructed from κισσου μεζρα and cissu mezra), Post-Biblical Hebrew
ḳiššūʔā, pl. ḳiššūʔīm 'musk melon', Arabic ḳu/iṯṯāʔ- 'cucumber; melon'.
The term ḳiššūʔīm in the meaning 'musk melon' occurs (only in pl.) in
Biblical Hebrew, while in Post-Biblical Hebrew it means 'cucumber' like in
Punic. Either a borrowing in Proto-Berber or "chain borrowing" from either
Punic or Hebrew (with some preference for the latter because of the meaning
'melon').

10. Berber: Ghadames a-šašid, pl. šašiden 'almond'; the palatalization of
the expected *γ also
appears in tašarḍemt 'scorpion', (?) Zenaga täyššaḍ 'arbre Balanites
aegyptiaca ... dont les fruits tūgga'n s'appellent dattes du désert …' ~
Phoenician šḳdm pl., Hebrew šāḳēd 'almond'.
Only the Ghadames parallel fits phonetically and semantically.
Anyway, irrelevant for our topic.

11. Berber: *ā-zātīm, pl. *ī-zūtām: Ghat azāčim, azatim 'olive oil',
Ahaggar ahâtim 'olive'
~ Phoenician zt 'olive', Hebrew zayit, pl. zētīm 'olive, olive tree',
Arabic zayt- 'olive oil'.
Undoubtedly loanwords from either Punic or Hebrew into North Tuareg
only, and as such irrelevant.

B. Berber forms of probable Phoenician/Punic origin without the direct
epigraphic documentation of sources, but with evidence in other Canaanite
languages

12. Berber: Qabyle agusim 'nut, nut tree' ~ Hebrew ʔägōz, pl. ʔägōzīm
'walnut, nut tree'.
An obvious loanword with an unexpected -s- in Qabyle. Irrelevant.

13. Berber (all branches) *ā-galzīm 'axe, pick-axe, etc.' ~ Hebrew garzän
'axe, pick-axe'.
A loanword in Berber, though the questions about -l- versus -r- and m vs.
n remain unanswered; the plural Punic form, if ever existed, would have
been *grzn-m which does not make the answers easier.

14. Berber (North, East and South branches) *almid (imperative), *yulmad
(perf.) 'to learn,
study' ~ Hebrew lāmad 'to learn, become tamed, teachable' (Arabic lamada
'to submit to, be obsequious' cannot be a source of the Berber forms).
Obviously, a loanword either from the non-preserved Punic word or Hebrew.

15. Berber *ā-māHdīr, pl. *ī-mūHdār 'kind of hoe': North and East
(Ghadames) branches ~
Hebrew maʕdēr 'hoe, mattock', Arabic maʕdūr 'pickaxe'.
Evidently, a loanword either from the non-preserved Punic word or
Hebrew or, less likely, Arabic (because of -ū- vs. *-ī- in Berber).
Irrelevant for limited attestation.

16. Berber: *ā-mindām (common Tuareg) 'something like that, so and so' ~
Post-Biblical
Hebrew mindaʕam & middaʕam 'something', derivable from maddaʕ 'knowledge,
science', plus -mā 'any'.
Convincing as a loan from Hebrew as it is hard to imagine the
unpreserved Punic source word of the same structure and meaning; anyway,
Tuareg only and as such irrelevant for the relative dating of Proto-Berber
split.

17. Berber (North, South and West branches) *ā-γurūm, pl. *ī-γurūm-an
'bread, bread crust'
~ Hebrew ḳərūm 'crust, membrane, skin', from ḳāram 'to overlay, cover with
crust, incrust', Syrian Aramaic ḳərāmā 'covering, membrane', Arabic (coll.)
ḳuram- 'Durra-bread', ḳurāmat- 'rest of bread in the oven'.
While the very fact of borrowing is indisputable, the source is most
likely Hebrew with its too specific meaning 'bread crust' to be of exactly
the same meaning in Phoenician, even if the related term in it ever
existed; note also -ū- in Hebrew vs. *-ū- in Berber.

18. Berber (North, South and West branches) *argim (imperative), *yurgam
(perf.) 'to
vilify', 'mock', 'curse', 'blame', etc.' ~ Hebrew rāgam 'to stone,
imprecate, cover with a heap of stones', Arabic raǯama 'to heap up stones;
make nasty remarks about, curse, abuse'.
The meaning 'imprecate' for Hebrew rāgam is adduced in [Leslau]
arguably erroneously: it is absent from [HALOT] and I failed to find it in
the available translations of the Bible. If there is no such meaning in
ancient Hebrew, Arabic remains the only source of this obvious borrowing in
Berber. The only problem seems to be Ahaggar erġem, in Foucauld's notation,
erḡem,[104] in the notation system I use. Malášková and Blažek quote
Vycichl who "explained that the Ahaggar word could not be of Arabic origin;
in this case it should look ǝrhǝm, cf. ǝlhin Gespenst" < North African
Arabic el-žinn; tǝlhamust "Buffelkuh" < North African Arabic el-žāmūsa."
However, this is true of borrowing from those North African Arabic idioms
where *g > ǯ/ž, but not in case of Egyptian Arabic where *g remains g –
cf., for instance, the verb 'to come': ǯa in Libyan, Algerian and Moroccan
Arabic vs. ga in the Cairo Egyptian Arabic; another source of borrowing
might have been a Yemenite Arabic dialect with its occlusive g (<*g)
considering an important role of its speakers in the Arabic conquest of
North Africa.[105] Therefore, an Arabism borrowed in Ahaggar from a ǯ-
language would have undergone the change ǯ > h, but borrowed from a g-
language should have rendered ġ/ḡ (< *g) which in the absence of the
meaning 'imprecate' in Hebrew (and with a shallow chance to have been
present in Phoenician even if the verb rgm expectedly 'to stone' did exist
in it) leaves the only opportunity to account for this Berber loanword:
from a North African Arabic g-idiom.[106]

19. Berber *ā-rumūn, pl. *ī-rumūn-an 'grenade (fruit)' (with a question
mark) ~ Hebrew rimmōn
'pomegranate (Punica granatum)', Arabic rummān-.
In most of the Berber languages, it's an Arabism (like in Ahaggar
errummân) while in a few, it might have been borrowed from Hebrew or
unpreserved Punic word, but not into Proto-Berber. Irrelevant.

20. Berber *ā-ẓārīf, pl. ī-ẓūrāf 'alum': North, South and East (Ghadames)
branches ~ Post-
Biblical Hebrew ṣārīp̲ 'alum'.
A borrowing either from a (Post-Biblical?) Hebrew or unattested Punic
term.

21. Berber: *ā-taggār (with a question mark in Malášková-Blažek's paper) ~
Post-Biblical
Hebrew taggār 'merchant, dealer, trader', Arabic tāǯir 'merchant'.
All Berber forms are Arabisms except Shilha of Sus tateggärt 'sale
certificate' which must have been a loanword from Post-Biblical Hebrew
taggār (or a non-preserved Punic source). Irrelevant.

22. Berber *(t)ā-daffūH(t) 'apple' ~ Hebrew tappū aḥ, Arabic tuffāħ-
'apple'.
Forms which might be loans from Hebrew or a non-preserved Punic source
are attested only in three North Berber languages (Shawiya, Djerba and
Nefusa plus 'white figs' in Qabyle), the rest are recognized as Arabisms.
Irrelevant.

The main conclusions in Malášková-Blažek's paper relevant for discussing
the dating of Proto-Berber split are:
- The Berber populations were in direct contact with speakers of two
Semitic languages,
namely Phoenician/Punic from c. 800 BCE till the 5th century CE and Arabic
from the end of the 7th cent; the hypothetical third one, Hebrew, is
questionable.
- In all Berber branches there are early Phoenician/Punic loans - at
least 6(7?) in the West,
and as many as 19 in the Northern branches, indicating that the first
contact preceded the disintegration of the Berber dialect continuum. This
agrees with the absolute chronology: the first contact could be
extrapolated on the basis of adoption of the Phoenician script (took place
in the area of Northern Morocco) and after the foundation of Carthage to c.
800 BCE, while the disintegration of the historically attested Berber
languages is dated between 680 and 460 BCE, if so-called 'recalibrated'
glottochronology is applied (Blažek 2010). The estimates by G. Starostin
(p.c.) and A. Militarev (1991, 154), dating the disintegration of Berber to
1480 BCE and c. 1300 BCE[107] respectively, are too early with regard to
dispersion of the Phoenician/Punic loans.

My conclusions and comments:
1) The Berber populations were in direct contact with speakers of three
rather than two
Semitic languages, the third one being Hebrew (even not to completely rule
out the forth – Aramaic spoken by the Jewish enclaves), cf., e.g., "Jews
lived in multiple communities in North Africa for >2,000 y[ears]" [Campbell
et al.]
2) Taking into consideration my comments on the examples adduced in
[Malášková-
Blažek], those of them regarded as loanwords from Punic can be classified
in the following way:
(2.1.) if the later split date by Blažek is accepted, they are to be
reckoned Proto-Berber;
(2.2.) if the earlier split date by the author is accepted, they are to
be reckoned loans in
individual Berber languages borrowed independently or through "chain"
borrowing either from attested and not attested Punic or attested Hebrew
terms; these terms are ## 1 (onions), 4 (lamp), 5 (copper; with some
preference for Hebrew as the borrowing source), 6 (reed), 9 (cucumber,
melon), 13 (pick-axe), 14 (learn), 17 (more likely from Hebrew), and 20
(alum);
(2.3.) a pseudo-Proto-Berber term not to be accounted for other than
a loan from Arabic: # 18;
(2.4.) terms which are attested within one branch only, cannot be
classified as common Berber and, therefore, are irrelevant for the dating
of Proto-Berber split debate: ## 1, 8, 10, 11, 12, 15, 16 (likely from
Hebrew), 19, 21, and 22;
(2.5.) terms inherited from Afrasian rather than borrowed: ## 2, 3
and 7.
(3.) Even loanwords unquestionably attested in more than one or in all
Berber branches can
be explained by a spread of every borrowed term from one Berber language to
another or from the source language independently in various Berber
languages rather than by borrowing in Proto-Berber. If we had been in the
dark about the much later times of borrowings from Arabic in all Berber
languages, we could easily take them for loans in Proto-Berber. The same is
true of the Roman/Latin influence, much more limited in number and more
comparable to the Punic one (see the classic study by H. Schuchardt
[Schuchardt]). With this scenario in view, it is difficult or impossible in
most cases to decide between Punic and Hebrew as the source of borrowing.
4. As for extrapolation of "the first contact…on the basis of adoption
of the Phoenician
script [by the Libyans]", the very fact of this adoption needs convincing
and very concrete – not just based on general considerations – proofs.
Until such proofs are presented I am inclined to see the Libyan script's
origin in South Semitic rather than Phoenician alphabets:[108] out of
eighteen Old Libyan and tifinaɣ symbols similar to Semitic ones both in
pattern and sounding, I could discern only three similar to Neo-Punic and
six similar to several varieties of South Semitic symbols (see Appendix 1).

5. Proto-Berber split took place ca. 11th century BCE. This is my most
recent dating
obtained by several procedures (all based on "recalibrated" Sergei
Starostin's methods applied by both Blažek, George Starostin and myself) –
"standard" recommended for 100-word lists and "experimental" for 50-word
lists: processing (a) a 100-word list of 40 Berber languages[109], (b) a
100-word list database of 170 Afrasian languages fully representing all the
branches, groups and subgroups including 15 representative Berber
languages, (c) a 50-word list database of the same Afrasian and Berber
languages as in (b).
The etymological identifications in all the lists in the Berber
lexicostatistical database, linked to the Afrasian etymological database
rely on complete etymologies wherever established. All the results showed
the late 2nd millennium BCE for the Proto-Berber split. The "disintegration
of the historically attested Berber languages… between 680 and 460 BCE"
obtained by Blažek [Blažek Migrations] using the same glottochronological
method is too late. The disagreement is accounted for by partial difference
in our choice of items (words) in the diagnostic list, including synonyms
when there are more than one of them in the same language; and certain
different etymological identifications (see Appendix 2a). With all due
respect to Vaclav Blažek, one of the few Afrasianists whose etymologies I
treat seriously (though mine are sometimes different), I cannot agree with
this dating. Quoted in the same paper, the date for disintegration of
Northern, Eastern and South Berber sub-branches, i.e. 1,080 BCE, obtained
by George Starostin in 2010 is close to mine though his date of separation
of East Berber branch (not the Western, represented by Zenaga as mentioned
by Blažek) from Proto-Berber – 1480 BCE – is too early (in a personal
communication, however, G.S. admitted it: for this Berber branch he relied
on reconstructed forms only instead of terms attested in individual
languages.)
Anyway, the key argument by Blažek which I cannot agree with is that
an earlier date is not compatible with the distribution of Phoenician-Punic
loans in all Berber sub-branches as "their spread is thinkable only in the
1st mill. BCE."

Appendix 5

Berber loanwords in Nubian.[110]

1. Berber (all branches) *Ham-an 'water' > Nile Nubian *aman 'water,
the Nile'
(Behrens and Militarev). This loanword apparently implies that when the
Nubian speaking population came to settle in the Nile valley, they
encountered there a Berber speaking one from whom they learned the name of
the river – water par excellence.
2. Berber (all branches) *a-drar 'mountain' ~ Nubian (Dongola) darr
(Militarev; cf. also North Omotic *dar 'mountain' which, in principle,
could be the alternative source of borrowing in Dongola).
3. Berber *a-ksum (all branches except Tuareg while there is Ahaggar
ta-kasa 'viande cuite sous la cendre', probably related) ~ Nubian (Kenuzi,
Dongola) kúsu meat' (Militarev).
4. Berber (all branches) *ird-Vn 'wheat' ~ Nubian *erde 'millet'
(Militarev).
5. Berber *ill-Vn 'millet, sorghum' (except Tuareg)[111] ~ Nubian
*ille 'wheat' (Militarev).
6. Berber *ug(g)id 'man' (except Tuareg) ~ Nubian *ogiǯ (<*ogid?) id.
(Militarev).
7. Berber: Ghadames tu-kessi 'kleines Holzteller', Ahaggar a-kus
'Holztopf', Tuareg-Niger a-kūss 'Schüssel' ~ Nobiin (Mahas and Fadidja
dialects) gōs, Dongola kōs 'Holzschüssel' (Behrens); the comparison seems
valid, cf., however, late Egyptian kꜣs 'vessel' and Arabic kaʔs- 'bowl'
(and Hebrew kōs), any of which can be an alternative source of borrowing in
Nubian.
8. Berber: Siwa a-fil, a-filû, Ghadames a-flilI, Sokna i-falil,
Ahaggar é-felêli 'Zwiebel' ~ Nobiin fille, felle, Dongola billɛ id.
(Behrens; comparison immaculate).
9. Berber: Ahaggar aġûleh 'Wolf' ~ Nobiin ǯellek id. (Behrens;
comparison doubtful for phonetic and chronological reasons; the Ahaggar
word is isolated in Berber).
10. Berber: Ghat ta-ġiger-t, Ahaggar ġiġer 'Maus' ~ Nobiin ǯigir id.
(Behrens; plausible).
11. Berber: Siwa ižedi, ǯedi 'Staub', Zwawa iǯḏi 'Sand' ~ Old Nubian
əskit 'Erde', Nobiin iskid 'Sand, Staub, Erde' (Behrens; comparison
unlikely: either of the derived forms of Proto-Berber *i-gid 'sand, earth'
can hardly be borrowed in Nubian as iskid).
12. Berber: kul 'alle, jeder' in various North Berber languages ~ Old
Nubian ⲕⲉⲗⲗω 'alle', Nobiin kúlli 'jeder' (Behrens); this is the most
enigmatic case: while the Tuareg forms like Ayr kəllu 'tous' and E.
Tawllemmet kul 'tout', less likely Arabism, can be added, this Berber
pronoun is usually regarded as such which makes its borrowing in Old Nubian
impossible. However, there is a reliable common Afrasian *ʔa-kʷall- 'all,
many' with reflexes, besides Semitic, in Egyptian (Pyr.) ṯnw 'each'
(<*kVlw) and South Cushitic Dahalo ʔákkale 'all' to which Berber *kul can
otherwise belong as an inherited lexeme.

Abbreviations:

Ahg – Ahaggar, Auj – Aujila, C – Central, E – East/Eastern, Ghdm –
Ghadames, Izd – Izdeg, Mz – Mzab, N – North/Northern, Nfs – (Jebel) Nefusa,
Qbl – Qabyle, S – South/Southern, Shw – Shawiya, Sml – Semlal, Sok – Sokna,
Tai – Taitoq, Tim – Timimun, W – West/Western, Zng – Zenaga

Conventional symbols:

H in a reconstructed proto-form stands for an unspecified laryngeal or
pharyngeal consonant.
V in a reconstructed proto-form stands for any or an unspecified vowel -
*a, *i or *u
S conveys a non-specified sibilant or sibilant affricate.
K conveys a non-specified voiceless velar or post-velar consonant.
/ between vowels in the reconstructed forms point means 'either…or' or
'both', e.g.: *a/isi means that either "a" or "i" is to be reconstructed or
that both vowels are equally possible.

Bibliography

AADB – Militarev, A. and O. Stolbova. Afrasian etymological database.
starling.rinet.ru

Alojaly, Gh. Lexique Touareg-Français, Cobenhague, 1980.

Behrens, P., Wanderungsbewegungen und Sprache der frühen Saharanischen
Viehzüchter. Sprache und Geschichte in Afrika 6, 1984, 153-154.

Bender, M.L., Omotic Lexicon and Phonology. Carbondale, 2003.

Biedermann, H., Altkreta und die Kanarischen Inseln'. Almogaren I, 1970.

Blažek 2000 – Blažek, V., Toward the discussion of the Berber-Nubian
lexical parallels. Études berbères et chamito-sémitiques. Mélanges offerts
à Karl-G. Prasse, ed. Salem Chaker. Paris - Louvain, 31-42.

Blažek Migrations – Blažek, V., Afroasiatic migrations: Linguistic
Evidence. [email protected]

Blažek Tuareg – Blažek, V., On application of glottochronology for South
Berber (Tuareg)
languages. [email protected]

Blažek V., Levant and North Africa: Afroasiatic linguistic history. The
Encyclopedia of Global Human Migration.

Blench, R., Berber plant names and the archaeobotany of the Maghreb.
International Workshop on African Archaeobotany, University of Las Palmas
de Gran Canaria 26-29th June 2018.

Brogan, 0., Inscriptions in the Libyan Alphabet from Tripolitania and some
notes on the Tribes of the Region, Coloquium on Hamito-Semitic Comparative
Linguistics. London, 1970.

Campbell, Ch. L. et al., North African Jewish and non-Jewish populations
form distinctive, orthogonal clusters. Proceedings of the National Academy
of Sciences of the United States of America. Vol. 109, No. 34, 2012.

Dallet, J.-M., Dictionnaire Kabyle-Français, Paris, 1982.

Foucauld, le père Charles de, Dictionnaire touareg-français, Paris 1952.

Gelb, I., A Study of Writing. Chicago, 1963.

HALOT – The Hebrew and Aramaic Lexicon of the Old Testamentum, by L.
Koehler & W. Baumgartner. Leiden-New York-Köln, 1996.

Hints-Badawi – Hinds, M and E. Badawi, A Dictionary of Egyptian Arabic,
Librerie du Liban,1986.

Kerr, R.M. , Some thoughts on the origins of the Libyco-Berber Alphabet.
Études berbères V, Actes du "Bayreuth - Frankfurt - Leidener Kolloquim zur
Berberologie". Köln, 2010, 41 - 68.

Krahmalkov, Ch., Phoenician-Punic Dictionary. Leuven, 2000.

Lazaridis, I. et al., Genomic insights into the origin of farming in the
ancient Near East. Nature. http://dx.doi.org/10.1038/nature19310 (2016).

Lanfry, J., Ghadamès, II, Glossaire. Alger, 1973.

Leslau, W. Comparative Dictionary of Geʕez (Classical Ethiopic). Wiesbaden,
1987.

Lhote, H., Les touaregs du Hoggar. Paris, 1984.

Littmann, E., L'origine de l'alphabet libyen, Journal asiatique, série 10,
tome 4, 1904, 423-440.

Lyavdansky, A., The Phoenician Language. The World Languages. Semitic
Languages. Moscow, 2009 (А.Лявданский. Финикийский язык. Языки мира.
Семитские языки. Москва, 2009).

Malášková-Blažek – Malášková, Z. & V. Blažek, Phoenician/Punic loans in
Berber languages and their role in chronology of Berber. Folia Orientalia,
Krakow, 2014.

Masqueray, É., Dictionnaire français-touareg (dialecte des Taitoq). Paris,
1893.

Militarev-Kogan – Militarev, A. and Kogan, L. Semitic Etymological
Dictionary. Vol. I. Anatomy of Man and Animals, Münster, 2000.

Militarev-Starostin – Militarev, A.Yu. and S.A.Starostin. Common
Afrasian–North Caucasian cultural terms. Preprints of the Conference
"Language Reconstruction and Prehistory of the Orient", Moscow, part 3,
Moscow, 1984 (А.Ю.Милитарев, С.А.Старостин. Общая афразийско-
севернокавказская культурная лексика. Лингвистическая реконструкция и
древнейшая история Востока. Тезисы и доклады конференции Часть 3).

Militarev 1984 – Militarev, A.Yu., Comparative-historical Afrasian
studies today: what can it contribute to the historical science? Preprints
of the Conference "Language Reconstruction and Prehistory of the Orient",
Moscow, part 3, 22-23 (А.Ю.Милитарев, Современное сравнительно-историческое
афразийское языкознание: что оно может дать исторической науке?
Лингвистическая реконструкция и древнейшая история Востока. Тезисы и
доклады конференции. Часть 3. Москва, 22-23).

Militarev 1984a – Militarev, A.Yu., Language of Meroitic Epigraphy in the
Light of its Genesis. Journal of Ancient History, 1984/2, 153-70
(Милитарев, А.Ю., Язык мероитской эпиграфики как исторический источник в
свете его генезиса. Вестник древней истории, 1984/2, 153-70).

Militarev 1991 – Militarev, A.: I. General Introduction, pp. 148-162; II.
Guanche Languages, § 1. Phonetics, pp. 163-173; V. Historical Phonetics and
Lexicon of Libyan-Guanche Languages, pp. 238-265. In: Aikhenvald., A. & A.
Militarev. Libyan-Guanche languages. Afrasian Languages. Languages of Asia
and Africa. Series IV, Vol. 2. Мoscow (А.Ю.Айхенвальд, А.Ю.Милитарев,
Ливийско-гуанчские языки. Афразийские языки. Языки Азии и Африки, IV, кн.
2. Москва).

Militarev 1994 – Militarev, A. The Garamantis in the context of African
history by the eyes of a linguist. Garamantis (the African Atlantis).
Moscow, 1994, 230-282 (Милитарев, А., Глазами лингвиста: Гарамантида в
контексте североафриканской истории. Гарамантида (африканская Атлантида).
Москва, 1994, с. 230-282); and its abbreviated German version in Militarev,
A., Eine Garamantiade im Kontext der nordafrikanischen Geschichte - Zum
Schicksal eines Volkes aus linguistischer Sicht. Hellenismus. Beiträge zur
Erforschung von Akkulturation und politischer Ordnung in den Staaten des
hellenistischen Zeitalters. Akten des Internationalen Hellenismus-
Kolloquiums 9.-14 März 1994 in Berlin. Tübingen, 1996, 725-748.

Naït-Zerrad, K. Dictionnaire des racines berbères (formes attestées). Paris
and Leuven, 1998.

Nehlil, Étude sur le dialecte de Ghat. Paris, 1909.

Nowak, H., Neue Felsbilderfunde auf der Insel Hierro'. lnstitutum Canarium
Nachrichten 37/38,
1981.

Peña-Rodríguez - Peña, P. A.and M. Á. R. Rodríguez. Manifestaciones de la
religiosidad
mediterránea en la Protohistoria canaria (circa s. X a.n.e. al s. XV
d.n.e.).
https://www.researchgate.net/publication/277801293.

Peña, P. A. Excavaciones arqueológicas en el sitio de Buenavista
(Lanzarote): nuevos datos para el estudio de la colonización protohistórica
del archipiélago canario. Gerión 63 2011, vol. 29, núm. 1, 59-82.

Prasse, K.-G. A propos de l'origine de H touareg (tahaggart), Copenhague,
1969.

Schuchardt, H., Die romanischen Lehnwörter im Berberischen. Kais. Akademie
der Wissenschaften in Wien. Philosophisch-historische Klasse.
Sitzungberichte, 188. Band, 4. Abhanglung. Wien, 1918.

Starostin, S.A., Comparative-historical linguistics and lexicostatistics.
Time Depth in Historical Linguistics, McDonald Institute for Archaeological
Research, Oxford, 2000, 223-259.

Stolbova 2011 – Stolbova, O.V., Chadic Lexical Database, Issue IV.
Velars. Moscow.

Stolbova CED – Stolbova, O.V., Chadic Etymological Dictionary. Moscow,
2016.

Takács, G., Etymological Dictionary of Egyptian. Vol. Three. Leiden-Boston,
2008.

Tomback, R.S. A Comparative Semitic Lexicon of the Phoenician and Punic
Languages. Missoula, 1978.

Valério-Yakubovich – Valério, M. and Yakubovich, I., Semitic word for
'iron' as an Anatolian loanword. Studies in Linguistics and Semiotics.
Collection of Papers for the Anniversary of Vyach. Vs. Ivanov (Исследования
по лингвистике и семиотике: сборник статей к юбилею Вяч. Вс. Иванова), ed.
T.M. Nikolaeva. Moscow, 2010, 108-116.

Vigne et al. The Early Process of Mammal Domestication in the Near East.
New Evidence from the Pre-Neolithic and Pre-Pottery Neolithic in Cyprus.
Current Anthropology, vol. 52, No. S4, 2011 (volume supplement).

Voskarides et al. Y-chromosome phylogeographic analysis of the Greek-
Cypriot population reveals elements consistent with Neolithic and Bronze
Age settlements. Investigative Genetics, 2016.

Vycichl, W., Dictionnaire étymologique de la langue copte. Leuven, 1983.

Wölfel, D. J. Monumenta linguae canariae. Graz, 1965.

-----------------------
[1] The present paper is a substantially revised, updated and enlarged
version of my paper "Tamahaq Tuaregs in the Canary Islands (Linguistic
Evidence)", Aula Orientalis, No.6, Barcelona, 1988, pp. 195-209, published
thanks to my friend Prof. Gregorio del Olmo Lete. It was my first
publication abroad and in English which, due to the then difficult
communication between Moscow and Barcelona, a few errors slipped in, the
grossest being the date of the 7th -8th centuries B.C., instead of A.D., of
the postulated Tuareg migration to the Canary Islands. This left the whole
idea up in the air (to say nothing of Tamâhaq in my original manuscript
replaced by Tamaraq – as perceived by French speakers). Anyway, this paper
passed nearly unnoticed and practically played no part in the subsequent
studies on the Canarian populations' ethno-lingustic history and
prehistory. Most publications I've come across since then either address
the genetic and archaelogical aspects touching the linguistic subject only
in passing or, if addressing a broader interdisciplinary perspective, are
stereotyped, repetitive or highly superficial (even compared with the
remarkable work by D. J. Wölfel [Wölfel] of the mid-20th century). This is
clearly demonstrated by unusually scanty articles in Wikipedia like Guanche
Language and Guanches, etc. and in Glottolog on Guanche with its
predominant references to sources old or immaterial to the subject.
As for the recent studies in Libyo-Berber ethnic, cultural and
linguistic prehistory, they either do not recur to such a mighty tool of
investigating the past as comparative-historical linguistic analysis and
etymology or use it on an extremely outdated level; on the rare exceptions
– studies by V.Blažek and a glottochronological sketch by G. Starostin –
see below. The indispensable work of compiling new dictionaries of
individual languages or lists of related words embracing the whole family
cannot supersede reconstruction of protoforms necessary for comparison with
related terms in other Afrasian branches, i.e. for "deeper" etymology.
Extralinguistic implications passing over this painstaking labor of immense
complexity are false or at best far from complete as, for example, those by
R. Blench [Blench] asserting late farming in the Libyo-Berber speaking area
and borrowing a bulk of agricultural terms from Phoenician and Latin:
without tracing the corresponding terms (except those having proved and
exact source of borrowing) to their inherited Afrasian roots implying
common Afrasian origin of these terms and, hence, of corresponding
domesticated plants and agricultural activities, any such conclusions are
flimsy.
[2] The widely used (including in the author's earlier publications)
term 'Guanche' is appropriate only for the aborigines and language of the
island of Tenerife. According to the 17th century author Nuñez de la Peña,
'Los naturales de Tenerife se llamaban guanchinet, que los Españoles
corrompieron en guanche, que quiere decir… natural de Tenerife… guan quiere
decir persona y chinet Tenerife' [Wölfel, 609]. In fact, guan should be
analyzed as *wa-n 'he/this of…'
[3] Out of over thousand words, mainly cultural terms and toponyms
known from all the insular idioms (considered extinct since 17th century),
most items were recorded in the 14th-17th centuries by Spanish and some by
Italian, French, English, Portuguese and other visitors to the Canary
Islands as well as 'Europeanized' natives of the islands. The transcription
variants of the same Canarian word – some similar and easy to bring into
accordance, some diverse enough to require exigeant analysis – followed by
a translation in the recorders' mother tongues often help to reconstruct
the original Canarian phonetic word.
[4] Cf., however, Tenerife guijon 'navio(s)', guihon 'ships', guyon
'navire', etc. [Wölfel, 542], likely to read as [wiHon] < *wiʔon, an
arguable metathetic match of Egyptian (Pyramid texts) ỉnw.t < *ʔinVw (V in
a reconstructed protoform stands for any or an unspecified vowel - *a, *i
or *u, according to the conception of Proto-Afrasian vocalism postulated by
Igor M. Diakonoff and accepted in most advanced today's comparative
Afrasian studies) 'ferry boat' corresponding to or borrowed from Egyptian
in Hebrew ʔǎnī 'ships, fleet' and Ugaritic ʔany(t) 'ship' (these examples
as all that follow, except otherwise referred to, are quoted after the
Afrasian etymological database [AADB]).
[5] probably consisting of several languages/dialects: for scarcity of
data preserved, inner classification and genetic status in the family is
not clear.
[6] Militarev 1991.
[7] This and other dates of protolanguages disintegration to follow
are the author's glottochronological calculations based on Sergei
Starostin's modification of M. Swadesh's method [Starostin], which yields
much more reasonable dates wherever they can be independently checked than
those produced by the traditional Swadesh's glottochronology. For
lexicostatistics and glottochronology of Berber see Appendixes 2, 2a (the
data) and 3 (the genealogical tree) of the present paper.
[8] Less likely, some of the Canarian words quoted below may have been
borrowed rather than inherited.
[9] It is tempting to unite the two into one implying an arguable
evolution from an epileptic fit/ possession by spirits (a shaman's trance?)
( the one who deals with spirits ( (evil) spirit ( god. Cf. a somewhat
similar semantic development in Afrasian *kVrw- (cf. IE *kʷer- 'magical
sign, omen'): Semitic: Arabic kry 'sommeiller', Tigre kərri belä 'talk
nonsense', säb kəyarät 'soothsayers' (säb is 'man'); Egyptian (22nd Dyn.)
kꜣ (<*kVr?) 'soul'; Berber: Ahaggar tă-karaww-at, Ayr i-kerker-ǎn
'épilepsie', E.Tawllemmet a-t-kər 'possessed (by an evil spirit)'; Chadic
W.: Hausa kùrwā 'soul; ghost', 'the soul or personality which is supposed
to leave a sleeper, returning when he awakes', C.: Bura kir 'the self; the
will'; E. Cushitic: Oromo ekeera (<*ʔVkVr-) 'ghost'.
[10] Cf. Palma garahagua 'ruín como un perro', garehagua 'méchant
comme un chien', etc. placed by Wölfel into a separate entry [Wölfel, 484-
5, entry "Hund"] which probably represents *garVhawa, a metathetic variant
of *hirgʷan-.
[11] -n in the Auslaut instead of -r in the two latter forms is likely
due to assimilation to the previous -n- (nota genitivi).
[12] S- conveys a non-specified sibilant or sibilant affricate and -K,
a non-specified voiceless velar or, in this case, probably a pharyngeal.
[13] As for al- in the two latter forms, it is another – besides a
regular a- with the masculine nouns in the singular – article-like prefix,
occasionally encountered in Berber (cf. Siwa a-gurzəni, a-gurzinî ~
əlgurazən, lugrāzén 'dog') likely from the Afrasian deictic pronoun *ʔVl-:
Semitic: Akkadian (Babylonian) ullû 'that', Tigre ʔəll-i/a (masc./ fem.)
'this', Arabic definite article ʔal-; E. Cushitic: Dasenech ello, Tsamai
ʔúlo 'this', etc.
[14] There is no chronological contradiction to the borrowing from
Hausa in Proto-Canarian: Hausa branched off the main West Chadicic stock as
early as in the 4th mill. BCE, according to my glottochronological
calculations.
[15] mocan is a fruit of Visnea Canariensis [Wölfel, 575].
[16] Our joint paper with Sergei Nikolayev on Afrasian animal terms
(in preparation) will contain a discussion, in which I adduce my arguments
for this hypothesis, while my co-author who, on the contrary, is inclined
to see the Proto-Afrasian Urheimat in Ethiopia adduces his, based mainly on
the set of Proto-Afrasian zoonyms reconstructed by the author. One of my
main arguments is based on the fact that a set of common Afrasian terms
likely pointing to insipient agriculture and surpassing by far similar
terms reconstructed if any for other proposed macrofamilies (Nostratic,
Sino-Caucasian, Austro-Tai, Austronesian, etc.) corresponds to the earliest
agriculture so far attested in the Levant. Сf.: "…farmers related to those
of the Levant spread southward into East Africa" [Lazaridis et al, 1]. And
also: "The first farmers of the southern Levant (Israel and Jordan) and
Zagros Mountains (Iran) were strongly genetically differentiated, and each
descended from local hunter–gatherers." [ibid.]
[17] the process likely reflected in the Saharan rock paintings and
engravings of the "Pastoral period" whose beginning is roughly dated to the
6th mill. BCE.
[18] Voskarides et al, 7:1.
[19] Vigne et al. Volume 52, Number S4 " October 1, 2011 (Volume
Supplement)
[20] glossed by Vycichl as 'partie de la personalité humaine; esprit',
etc. [Vycichl, 69].
[21] Also in [Militarev 1984, 22]. The alternative explanation would
be the Egyptian-Canarian contact in the Canary Islands implying an Egyptian
expedition there but there seems to be no other evidence of this.
[22] Takács, 733.
[23] Militarev 1984, 22.
[24] That only the Ghadames term's order of root consonants matches
the Egyptian one can be accounted for by the former (and, perhaps, similar
unregistered forms in other East Berber languages) retaining the original –
Egyptian – order, while in non-Eastern Berber, metathesis apparently took
place due to the incompatibility tendency between the two lateral
consonants in immediate contact: l and s which, according to the regular
correspondence established by the author (Berber *s ~ Egyptian š < Afrasian
voiceless lateral sibilant *ŝ), must have still pronounced as lateral
sibilant in early Berber. If this reasoning is correct, the "secondary"
order of radicals in Canarian salema probably features it as a borrowing,
feasibly from Tamâhaq/Ahaggar, rather than going back to the Proto-Berber-
Canarian level. The Berber term was borrowed in Hausa as lámsà (likely from
Ghat) and in Central Chadic language Bura as šalmwi (from some other Berber
language) 'a k. of fish'.
[25] but not *maŝiḫ as in [Takács, 625] as Afrasian *ŝ yields *s/š,
not *z continuing Afrasian lateral affricate *ĉ in Proto-Berber. See also
the discussion there of my (in Милитарев 1991, 150-151) and some earlier
authors' identification of mšwš with a-maziɣ, self-designation of Berbers.
I agree with Takács that there is no convincing etymology for *a-maziɣ: one
of the etymologies – from Berb. *i-zwaɣ 'be red' which I suggested
following Sarnelli's idea – excludes, as Takács justly observed, any
comparison with Egyptian mšwš as Berber *i-zwaɣ is < *zwḳ cognate of N.
Omotic *zoḳ 'red' (< Afrasian *ʒ/ǯawḳ-). Other etymologies relying on W.
Tawllemmet əžžəɣ 'marcher d'un pas altier, marcher comme un noble' or žǎɣǎɣ
'ê. brave, intrépide' fitting well phonetically and semantically are of a
too narrow representation to claim the origin of the common Berber *a-
maziɣ.
[26] But not, to my knowledge, 'wild dog' as quoted in [Blažek
Migrations].
[27] There are two stunning things about the Egyptian ꜣbyḳwr vs.
Tuareg *abaykur case. One of them is that it seems to be the longest ever
known historically preserved word (for over four millennia – think of it!)
in its full and precise integrity. The other is what appears to be an
effort – requiring a sophisticated linguistic appreciation – to convey
velar Berber k not by the corresponding Egyptian k but by the postvelar
"emphatic" ḳ the reason, in my view, being that in Egyptian, k before w/y
could change into ṯ [č] and thus would not have rendered the exact
phonation of the Berber-Libyan term. A somewhat strange fact is that the
latter is preserved in the Tuareg group only (dropped in the other
groups?), while the Proto-Tuareg branched off the rest of Berber not before
the late 2nd century BCE – a millennium later than the term in question was
recorded in Egyptian.
[28] which would have been but one of the "pre-Berber-Canarian"
speaking groups spreading along the Mediterranian coast and in the Sahara
since 6,000 BCE (see 1.2.2.) – the one inhabiting Libya, whose idiom was
the protolanguage of both Canarian and all extant (and ancient extinct)
Libyan-Berber languages. As for other "pre-Berber" groups, they might have
been absorbed by the new waves of Libyan-Berber speaking migrants
colonizing the same areas.
[29] Militarev-Starostin, 39.
[30] No wonder: cf. "The Mitanni were closely associated with horses.
The name of the country of Ishuwa, which might have had a substantial
Hurrian population, meant "horse-land"" [Hurrians, Wikipedia].
[31] V. Blažek [Blažek Migrations] arrived at a later date (on our
differences see p. 47-8).
[32] See, however, examples in Appendix 5 rather implying a long-
standing – if temporary – sojourn of Berber speaking population in the Nile
area in southern Egypt or northern Sudan which could be either a group of
Garamantes, who, for some reason, stayed there for some time or a "stray"
group of earlier Proto-Berber speakers, cf. "In the 3rd and 2nd mill. BCE
the linguistic traces of Berber related idioms appear in the Nile Valley…
This is in agreement with the information of Herkhuef, a commercial
emissary who visited Upper Nubia c. 2230 BCE, about the ruler of the
district J3m fighting against the tribes Tmw by the fourth cataract. The
Tmw are usually connected with antique Libyans and modern Berbers. From the
area between the 2nd and 4th cataracts the Tmw are mentioned also in the
time of Ramesses II (1290-1224 BCE) on the stele of his official Ramose who
sought workers among Tmw (Behrens 1984, 137-39)" [Blažek Migrations].
Anyway, this set of Berber loanwords in Nubian implies long-time rather
than brief contacts.
[33] Cf. a mind-blowing case — which could even be taken for a
whimsical scholarly mystification if not for the much-respected author — of
the word for 'god' in the North Omotic language Gimirra (Bench) which is
garamantči [Bender], the implications probably overlooked by Bender's
leaving no comments.
[34] See [Peña].
[35] [Ibid., 59]. What I could not deduce from this publication is
whether all the finds are considered by the archaeologists of Punic
provenance testifying to the Phoenician presence in the islands or they
admit that part of them may bear evidence of the indigenous population in
the early 1st mill. BCE. Anyway, the undoubtful identification of Punic
artefacts does not necessarily speak for the Phoenicians' presence in
Lanzarote: these could be items and even construction technique picked up
by proto-Canarian speakers from the Phoenicians on their westward route
along the Mediterranean coast.
[36] See [Foucauld].
[37] Certain discrepancy in conveying long/short vowels and e/ə in the
Berber, especially Tuareg, examples adduced in this paper is accounted for
by different notation systems used by various authors.
[38] With the exception of rare cases when Afrasian *ǯ (Semitic *ḏ)
yields Berber *d.
[39] The symbol V in a reconstructed form stands for a vowel of an
unidentified quality.
[40] rendered by h and, in Spanish transliteration, also by j,
sometimes x or g – inter alia depending on the period when the word was
recorded.
[41] Ahaggar h is of a mixed origin, reflecting three Proto-Berber
consonants, namely *z, *h and *b̲ (otherwise *hʷ).
[42] The disturbing factor here, like in 3.2.2.5, questioning the
presumption of borrowing is lack of the tV-…-t confix in the Canarian
terms: words are normally borrowed in a morphological form of the donor
language. However, if the donor language is understandable to a degree by
the recipient language's speakers, the latter may, in principle, catch on
the stem of the borrowed word and leave out the affixes.
[43] Prof. Prasse suggested in a personal communication that Canarian
-harenemen might be understood as *-ɣenemen (cf. Qabyle iniɣəm, pI. iniɣm-
an 'dry fig') with metathesis; this would imply analyzing the Canarian form
rather as t-ahar enemen < *iniɣm-ən with the fallen uvular (or the one
unheard by the recorders) which is less likely, in my view.
[44] This Common Tuareg *a-zayh is likely related to the Common Berber
verb *azub̲, perfective *yuzab̲ 'to skin': Ghadames úzəb̲, Ghat az, perf.
yuza, Ahaggar ah, perf. yuhǎ, Ayr azu, Semlal azu, perf. yūza, etc.
[Prasse, #114].
[45] Possibly < Common Tuareg *tā-m-azāy-t (> *ta-məzay-t >*ta-mzī-t)
< *ta-m-azayh-Vt, related to *a-zayh (see fn. 44).
[46] Direct identification of this toponym with the Arabic tribe Beni
Hawara discussed in [Wölfel, 476-477, 611] is, of course, anachronistic in
this form – with the Arabic "Beni" part of it.
[47] I expect an important test of this hypothesis to come from
geneticists with their progressing chronological methods.
[48] The date obtained by V.Blažek is different – the 11th century CE
(on our differences see Appendix 2a).
[49] The comments made in Spanish: '...esta devocion se entendia por
los juramentos, ruegos y peticiones que hacian; no le sacrificaban mas de
rogarles por los temporales para herbaje a sus ganados' [Abreu de Galindo,
Padre Fray, Historia de la Conquista de las siete islas de Gran Canaria.
Año de 1632, quoted after Wölfel].
[50] E.g. Ahaggar ere iẓẓâl-ən 'he who is right', lit. 'the being
right', iẓẓâl-ən being a participle of the verb əẓẓəl 'to be right'.
[51] For example, yorəh-awən Yǎlla 'que Allah vous donne en retour!'
[Foucauld, 1609].
[52] añepa with this meaning was also attested in Gran Canaria.
[53] The form anzpa was suspected by Wölfel to be a mistake or
misprint but it is certainly not.
[54] The definite dimensions of the ǎ-ñhəf pole (from 0,3 to 0,5 cm in
diameter and from 1,5 to 1,8 m long) mentioned in [Foucauld, 1348] point to
a specific function of this implement.
[55] Prof. Prasse (in a personal communication) cast doubt on this
comparison, namely, on Tenerife (and Gran Canaria) p conveying Berber f.
Though Berber f (the Proto-Afrasian *p merged with *p in Proto-Berber)
normally corresponds to Canarian f, the plosive p is not rare in recorded
Canarian words, while there seem to be several cases of tenable lexical
comparison between Berber terms with f and Canarian ones with p (see, for
example, 3.2.2.2.). Anyway, the similar entangled phonetic developments of
the Canarian and Ahaggar forms in question make the proposed comparison
practically infallible.
[56] Also known as a proper name of a 'caballero canario' [ibid.]; for
the meaning of mencey see 3.2.1.4., reste is another high title [ibid.]
[57] The comparison suggested by Wölfel is with Ahaggar ager 'être
plus grand que'. Fitting quite well semantically, it does not explain the
evident labialization of the Canarian term.
[58] K.-G. Prasse employs the terms "perfect" and "imperfect".
[59] Cp. īb̲er a-meqqar 'rigole principal de distribution à l'intérieur
du jardin' [J. Lanfry, Ghadamès, II, Glossaire, Alger, 1973, p. 26] and the
comments on Herro hero: 'Y para assi hazerlo [proveer la isla de agua] se
recoge el agua en una alberca al pie del arbo1..., Se la reparten con buen
orden...' (Antonio de Viana) [Wölfel, 511].
[60] Милитарев 1984, 22-23. Whether Afade ohárbotú was borrowed via
Proto-Berber or directly from Egyptian (chronologically possible: by
glottochronological calculation, Afade separated from Proto-Kotoko as early
as 2nd mill. BCE), it is a phenomenal evidence of a word's phonetic
preservation. The Bura and Buduma verbs are rather loans from Hausa, while
the Ngizim one looks like a later loan from Tuareg. Anyway, Wölfel, though
he did not arrive at the idea of comparing the Libyo-Berber terms with the
Egyptian one, insightfully compared them with Kanuri revo-nin 'write'
[Wölfel, 461] (likely from Hausa), Old Nubian uran 'Schreiber' [ibid., 462]
(another Libyo-Berber loanword in Nubian – see Appendix 5). Note another
word, compared by Wölfel, the ancient Greek téras 'Götterzeichen' ('sign,
emblem, wonder, monster'), according to Frisk [see Frisk, H., Griechisches
etymologisches Wörterbuch. Heidelberg, 1960, 878], an archaic term without
secure connections (I am thankful to the Indo-Europeanist Dr. Corinna
Leschber for consulting me on this word); as for the comparison with Basque
ira-atsi 'añadir, escribir', ira-korri 'leer', it is rejected by R.L. Trask
since -ra- appears a causative of a simpler verb (I am obliged to John D.
Bengtson for this information).
[61] The problem which could be formulated as 'Arabs in the Canary
Islands?' (just like 'Phoenicians in the Canary Islands?', 'Cretans in the
Canary Islands?' and so on) needs to be specified. The question is not
whether some Arabic seamen ever reached one of the Canary Islands in
accordance with this or that tradition, but: was there any cultural impact
on the indigenous populations traced archaeologically, genetically or
linguistically? As for the Phoenicians, their role in the prehistoric
insular cultures including the religious aspect is recently claimed in
various studies [see, for example, Peña-Rodríguez], though as a comparative
linguist, I am more convinced by lexical, etymological and inscriptional
evidence in which I have not so far come across any Phoenician traces. As
for the Arabs, such a 'marked' culture as Islamic Arabic leaves perceptible
traces everywhere, a taboo on eating pork being one of the most
conspicuous, while pig-breeding was among the main occupations of the
Canarians.
[62] The only Berber parallel Wölfel adduces is Taitoq sabor 'enceinte
(qui entoure une tente)' which hardly fits semantically.
[63] And xerco, jerco which is a 'reverse' formation of singular from
xercos perceived as Spanish plural, though in this case it is certainly
not, judging by the comparative Berber data.
[64] See discussion on these subjects incuding the origin of famous
Tuareg crosses in [Lhote]
[65] There are several Latin borrowings referring to a religious
sphere, in the Tuareg lexicons, e.g. Ahaggar aneḡlus [Foucauld, 1332], Ghat
aǧəlus, Ayr ǎngǎlos [Alojali, 142] 'ange' < Latin angelus; Ahaggar abekkâḍ
[Foucauld, 52], Ghat abəkkaḍ, Ayr abǎkkaḍ [Alojali, 6] 'péché' < Latin
peccatum, etc.; interestingly, in some non-Tuareg Berber languages, the
same borrowed terms lost their original meaning (cf. Ghadames angalūs, mot
peu compris, équivalent à: inspiration, esprit [Lanfry, #1125] or Qabyle
āḅəkkaḍu, mal ou maladie? Sens mal défini [Dallet, 20]), which probably
points to a weaker impact of Christianity on these languages' speakers than
on the Tuareg-speaking tribes. There is also a term of Berber origin:
Ahaggar Mess-îneɣ 'Dieu' (lit. 'maître/propriétaire-notre') [Foucauld, 1245-
6], Ayr, East Tawllemmet Məšš-ina 'Dieu, Le Seigneur' [Alojali, 132]
(alongside with Yalla 'Allah' borrowed from Arabic), apparently reflecting
a pre-Islamic monotheistic tradition.
[66] Cp. Tenerife chaxiraxi 'Nra Donna', chaciraxi 'Santa Maria',
chijoraji: 'Adoraban por cosa celestial y suprema deidad a la Virgen de
Candelaria y al niño en su mano derecha...' [Wölfel, 441]. With all that
probable retrospective 'christianization' of the image, the surprizing
thing about the term is its Tamâhaq origin. The reconstructed form of it is
*ča-HiraH-i < *ta-HiraH-i (Tenerife č < *t), lit. 'she-of-HiraH-my', cp,
Ahaggar ta 'celui qui' which is a feminine counterpart of wa (see point
3.2.2.6.) Now, *HiraH is well-attested in the forms achguayaxerax 'el que
todo 1o sustenta', 'the maintainer of all' (< *at wa yaHeraH), guayaxerax
'sustentador y autor de 1o criado' (< *wa ya HeraH), etc. (in [Wölfel, 362-
365], the Ahaggar forms are compared). The verbal form *ya-HeraH '(he)
sustains, preserves' is the Ahaggar ɣereh 'serrer; conserver, placer en
lieu sûr' [Foucauld, 1773-4]; h < *z (proving its Tamâhaq origin in
Canarian, cp, point 3.2.2.) assured by the related Ghat verb əɣrəž, əɣrəǯ
'serrer, enfermer' [Prasse, #582] and the Canarian variant form guayageraz
(on z/h see point 3.2.4.). So, the exact reconstruction of chaxiraxi would
be *ta ɣirah-i 'She of my Sustainer'.
[67] Or continuity rather than heritage. See arguments for the theory
asserting that modern Tuaregs, the Tamâhaq Tuaregs in particular, are
direct descendants of the Garamantes in [Militarev 1994].
[68] See the presumably full inventory of symbols of different
varieties of Old Libyan and tifinaɣ scripts in [Militarev 1994, 270-276].
[69] The idea of the Libyan alphabet's origin from the South Semitic
script seems to be first put forward by Enno Littmann in his L'origine de
l'alphabet libyen, 1904 [Littmann] and nearly unanimously rejected. From
recent reviews of the subject see, for example, Kerr (2010).
[70] With possible exceptions in Hausa a iiloo 'iron band on spear
shaft', most likely borrowed from Tuareg; and late Egyptian zn (if < *zl)
'ploughshare': if connected, this term can be a loan from Berber, but not
otherwise for both semantic and chronological reasons – Egyptian of the
Greek period is much younger than the Proto-Berber one.
[71] On that complicated common Semitic term, presumably an Anatolian
loanword, see [Valério- Yakubovich].
[72] The one and only partially analogous development I know is a
complete loss of the Phoenician (or Hebrew) source-term's initial b- by the
borrowed North Berber forms (see Appendix 4 #1).
[73] Otherwise interpreted as wədān.
[74] ʕāšōt occurs only once in the Hebrew Bible and has no etymology.
If so, it is not clear on what grounds its translation as 'worked,
processed' in the Septuagint and Vulgata is based. I suspect a mistake in
the Hebrew text: instead of ש (shin) it should have been שׂ )sin) rendering
ʕāŝōt from ʕāŝā 'make, manufacture'.
[75] Graphic variants of the same character are separated by a comma.
[76] There are two styles of the East Numidian script, a 'horizontal'
one used in the so-called monumental inscriptions, and a 'vertical' one
used mostly in epitaphs. The two seemingly identical graphemes for w and l
are, in fact, distributed between them: w is rendered by , l by
in a horizontal, and w by , l by in a vertical style.
[77] This sign (four dots) unattested in East Numidian, seems to have
developed from sign #8 (three dots) to be later borrowed in the Ahaggar
tifînaɣ.
[78] The development of this symbol from three lines to three dots is
an important argument for the transmission from the Tripolitanian, most
likely Garamantes', script to the Ahaggar tifinaɣ.
[79] Three series of symbols for a voiceless sibilant (corresponding
to Ahaggar s phonetically - see below) are hard to explain.
[80] Unattested in what we call Tripolitanian inscriptions, this
sign's variants (and ) occur in the epitaphs of the cemetery
near Djerma, which I tend to unite, under the name of Fezzan inscriptions,
with Tripolitanian ones into a separate subgroup of the Libyan script.
[81] This sign renders ty = č.

[82] The grapheme transliterated as gg renders in the Ahaggar
tifinav the voiced stop g (see "g" in Appendix 1 Table "Special affinity of
Libyan and North Semitic (Phoenician) scripts") originated from *-ggʷ (<*-
ww) and thus opposed to the voiced velar spirant ḡ (or, perhaps, its
palatalized match gy) <*g. In our inscription, ggʷ is rendered by the
succession of two signs: and (for w).
[83] The combined character (read as lt) made of (for l) and
(for t) is current in the Ahaggar tifînaɣ.
[84] The interesting fact is that the triconsonantal root of this well-
known Libyan tribal name written in the 'native' alphabet appears to
coincide with that of the Semitic name for Egypt (Hebrew Miṣrayim, Arabic
ʔal-Miṣr-, etc.)
[85] The dot in East Numidian is a sign of word-border and is surely
the same in Canarian inscriptions; applied rather inconsistently.
[86] Consulting [Foucauld], and using his transcription of Tuareg
words.

[87] The symbol [] points to the absence of the word in a
corresponding lexical source(s) or its borrowing (usually from Arabic).

[88] (1) and (2) are regarded as two different roots by both Blažek and
Prasse; they are also
likely to continue two different Afrasian roots, perhaps, eventually
related on the pre-Proto-Afrasian level).
[89] After [Blažek Tuareg].
[90] The Ghat terms are quoted after Nehlil; the Taitoq, after
Masqueray; the Ahaggar, after Foucauld.
[91] There is some discrepancy in 85% of common scores between Ghat and Tai
vs. 87% between Ghat and Ahg likely accounted for by one-two undetected
(and hardly detectable) loanwords from Ahg in Ghat: that borrowing from a
Tamâhaq idiom in Ghat even in the nuclear lexicon did take place can be
seen in an evident loanword: Ghat ta-ǯahu-t 'belly'; the inherited Ghat
form should have been ta-ǯazu-t (or ta-ǯažu-t or ta-ǯaǯu-t) <*ta-gazu-t,
cf. Ayr ta-gəzu-t 'belly'. Consequently, if the date of Ghat and Tai split
is chosen for a Ghat-Tamâhaq split, it would be the last third of the 6th
century.

[92] Cf. "The Garamantes were probably present as tribal people in the
Fezzan by 1000 BC" (Garamantes – Wikipedia)
[93] or Proto-Shilḥ-Beraber/Tamazight, the Shilḥ subgroup represented
in the tree by Semlal and the Beraber/Tamazight subgroup, by Izdeg.
[94] All the reconstructed Berber protoforms quoted in [Malášková-
Blažek] are after Prasse; in some of the Berber and Semitic forms adduced
in their paper, the notation is slightly changed in accordance with the one
I am using in the present study. The data by these authors are given in the
paragraph following the example's number while the paragraph(s) that
follows contains my comments.
[95] The only case where a partly similar development could be
hypothesized, Berb. *ʔuzzāl allegedly from Phoeniciam brzl, appears to have
quite a different explanation (see 6.1.3.2.)
[96] The loanwords of Punic or Hebrew origin with b- are represented
only in Sened ti-běslim 'oignon' and Augila (coll.) bẓālîm 'cipolla' which
is certainly not enough to claim the Proto-Berber status; there are also
isolated forms considered Arabisms: Jebel Nefusa bṣål (coll.) 'cipolla' and
Ghadames ta-baẓẓul-t 'ail sauvage'.
[97] Adduced in [Krahmalkov], but absent in [Tomback].
[98] See [Naït-Zerrad].
[99] Whatever full is formal coincidence between the derived forms of
this root and the other two in individual languages, the meaning shift from
the verbs conveying various kinds of movement to 'uncover, lay bare' is far
from evident; another argument in favor of these verbs' casual homonymity
is high frequency of the *gl combination of root consonants.
[100] dated glottochronologically prior to 3,000 BCE: an unexpected
meaning for such an early date.
[101] I was surprised to read in Wikipedia Berber Languages that
Ethnologue, "mostly following Aikhenvald and Militarev (1991)", classifies
Ghadames as Northern Berber: "including Nafusi and Ghadames within Zenati".
In the work mentioned, as in all my subsequent studies, Ghadames is
unequivocably part of the eastern branch; Jebel Nefusa included by some
authors into the Eastern branch is, lexicostatistically, definitely a North
Berber language of the Zenet group.
[102] Note also the meaning 'copper' in Hebrew and Berber vs. 'bronze'
in Phoenician.
[103] Probably in Hebrew only, whereas in Arabic it was under Hebrew
influence – otherwise it's hard to explain why the meaning 'to read' seems
to be absent from Aramaic and Ugaritic, the languages closer genetically to
Hebrew than Arabic.
[104] Meaning in Ahaggar 'fair une epigramme contre', the same meaning
in Ayr and E. Tawlemmet ərgəm – no way to be borrowed from other language
than Arabic, most likely Egyptian Arabic ragam 'revile, curse' (Hints-
Badawi).
[105] See [Corriente].
[106] In this case, a really great scholar Vycichl was certainly
wrong.
[107] Not exactly. In [Militarev 1991, 154] there is reference to my
glottochronological calculations presenting the last third of the 2nd
mill., not c. 1300 BCE, as the time of Proto-Libyo-Berber split. My most
recent results are just closer to the end of that period.
[108] My guess is that the Libyans borrowed or, rather, adopted their
earliest alphabet from a variety of Southern Semitic script and later added
three more symbols from the Punic one.
[109] A representative list of 17 languages see in Appendix 2.
[110] The lists of examples discussed below were composed
independently and, curiously, published simultaneously by Peter Behrence
[Behrens] and the author [Militarev 1984] which the two authors learned, to
their mutual surprise and collegiate joy, from each other at a conference
in Vienna in 1987. Some of these and a few less convincing cases were
suggested by Zyhlarz (1934-35), Wölfel (1955), Vycichl (1961), Bechhaus-
Gerst (1984) and more recently by Blažek [Blažek 2000], with references to
these authors' studies.
[111] A common Afrasian term *w/yVl- ~ *lay/w- 'k. of corn, bean,
seed', including Cushitic like Saho-Afar ilaw 'grain, wheat' and the Omotic
forms, any of which can be an alternative source of borrowing in Nubian.

Berber-Tuaregs-Canarians.doc

Recommend Documents