I recently sat down to read the wonderful article by the Naima Louali & Gérard Philippson (2004) "Berber expansion into and within north-west Africa: a linguistic distribution", Afrika und Übersee, pp. 105-130.
A wonderful article that discusses some of the chronological issues of the reconstruction of Proto-Berber. The internal coherence between the Berber languages strongly suggests a reconstruction of Proto-Berber no further back than the first millennium BC. At the same time, even if Proto-Berber forms a subbranch of Afro-Asiatic together with Proto-Semitic, their shared ancestor must be many thousands of years before that.
This would suggest that either Proto-Berber somehow 'koinized' at some point, to be a lot less differentiated than one would expect, or that many expected sister languages of Proto-Berber once existed but have now died out (or both!).
Louali & Philippson sensibly suggest that the Libyco-berber inscriptions of North Africa and the inscriptions of the Canary Islands (and perhaps the original indigenous Guanche language) are not forms of Berber, but rather "Para-Berber", that is, sister languages of Proto-Berber.
This would explain quite nicely why this material is so unreadable, but the few things recognisable still clearly look Berber in morphology.
Louali & Philippson also give in an appendix an overview of lexical items clearly reconstructible for Proto-Berber that relate to animal husbandry; However, no actual reconstructions are provided. I thought it would be nice to supply some reconstructions for these lexical items here.
*a-zgər'bull, ox'
L&P suggest a reconstruction *a-zgăr. And indeed the ă vowel in the Niger Tuareg azgăr generally require such a reconstruction. However, there are, to my knowledge no other *CCăC stems (only *CăCăC), and that vocalism would usually cause Mid Vowel Harmony that shifts the prefix vowel *a to e. The vocalism therefore is rather unusual.
The Zenaga form äzgər seems to point to *a-zgər, and this is in line with what we find in the other Berber languages.
*w/ʔulliʔ'sheep and/or goats'
The initial consonant gives some problems. Zenaga seems to point to an initial consonant *ʔ, while others sometimes have w and other starts with a vowel u.
*ta-βăl-e pl. *ti-βatt-ăn 'ewe'
Unproblematic reconstruction of a noun with an unusual supletive plural, and an unusual (but Pan-Berber) masculine plural-looking suffix *-ăn, rather than the usual -en, or for nouns with the feminine suffix -e: -iw-en.
Vycichl, who derives the feminine ending -en (without making the sound laws explicit) from feminine suffix -t + plural suffix -ăn suggests this word as evidence for this, and considers *ti-βatt-ăn is an assimilation of *ti-βăl-t-ăn. It is somewhat difficult to understand for this noun because 1. it does not have the feminine suffix -t in the singular, and 2. there's no direct evidence for a lt > tt assimilation. It works better for the noun*ta-ɣaḍ-t, for which see below.
*ta-aɣs-e pl. ti-aɣs-iw-en 'female sheep or goat'
Unproblematic reconstruction.
*ta-ɣaḍ-t pl. ti-ɣaṭṭ-ăn< *ti-ɣaḍ-t-ăn? 'female goat'
Unproblematic reconstruction, and a fairly convincing example of the assimilated *-t-ăn suffix, which blocked it from shifting to *-en.
*a-ɣăyd'(goat) kid'
Unproblematic reconstruction.
*a-akrăr/*a-kărr-əʔ pl. *i-akrar-ăn'male sheep'
The plural reconstructs fairly unproblematically. The singular, however presents some problems. Besides an etymon *a-akrăr, which is reflected, mostly, by the Tuareg dialects, the northern Berber varieties occasionally point to forms with some kind of -i suffix, which is not easily explained.
*a-kărăw pl. i-kərw-an'lamb'
The Tamazight and Tashelhiyt forms point to a reconstruction *a-kărăw for the singular. The Tuareg forms a-kərwa(-t) appear to be rather idiosyncratic back formation from the plural stem.
*a-zulăɣ pl. *i-zulaɣ'he goat'
The *a-zalaɣ plurals found in several Northern dialects are probably back-formations from the plural. The Siwi form zalaq pl. izaluqən is difficult to explain. Moreover the q in Siwi and Awjila azáləq pl. zulíq in this word are irregular
*a-aɣəβ'milk'
Somewhat unusual reflexes, here and there, but this reconstruction nevertheless seems fairly certain.
*a-dăɣăs/a-dɣəs 'colostrum'
The Tuareg varieties point to a formation with CăCăC pattern, while the Northern Berber varieties and Zenaga point to a reconstruction with a CCəC stem. One might consider that the Northern Berber varieties represent a back formation from the plural *i-dəɣs-an, but most northern dialects don't have a plural, and Tamazight has a different formation. Moreover, 'colostrum' is not very typical concept to have a plural form.
*ăksəʔ 'to herd; pasture'
Unproblematic reconstruction.
*a-mă-ksaʔ 'shepherd'
Unproblematic agentive formation of the previous verb.