Talk:Proto-Berber language

Reconstructions
Currently there are two competing vocalic reconstructions and a table which uses a third.

Consonants suffer from some of the same too, at least transcription-wise. The that may or may not merge with  is, presumably, the same as. Similarly probably equals the table's ? It looks like it might go together with as an emphatic.

I think I've seen a reconstructed as well elsewhere. Still much to do here... -- Trɔpʏliʊm • blah 14:53, 15 July 2009 (UTC)
 * I imagine */ɗ/ is intended to be the "glottalized" version of /d/, just like */t’/ is a "glottalized" /t/. The alternative would be that both were pharyngealized, which appearently isn't what Allati wants to convey. Mo-Al (talk) 04:13, 16 July 2009 (UTC)
 * Oops, I was confusing and  for a moment! Regardless, I'm pointing to note #14 which prefers pharyngealization. -- Trɔpʏliʊm  • blah 14:45, 19 July 2009 (UTC)


 * I've fixed the problem with the third reconstruction in the table; it should read less confusing now. However, I suspect that the table actually tries to say that Proto-Berber *i and *u are in free variation and do not actually contrast (the same has actually been assumed for early Arabic, if I remember the chapter on Arabic in Hetzron's The Semitic Languages correctly), and that Tuareg/Ghadamsi /ə/ corresponds to both /i/ and /u/ in Zenaga, so that one would actually have to assume a vowel phoneme */i ~ u/ for Proto-Berber. Which would result in yet another 5-vowel system effectively, with two short (high/closed and low/open, perhaps *a and *ɨ) and three long vowel phonemes. But if it actually depends on the lexeme in Zenaga and there is no free variation actually observable, it makes more sense to assume two separate phonemes. I don't really know. Anyway, in the third reconstruction Zenaga behaves a lot like Sardinian with respect to Latin, by simply collapsing the corresponding vowel quantities.
 * The profusion of competing reconstructions, also in the realm of the consonants, is certainly confusing. Clearly, parts of the article presuppose or take a different reconstructed system (or even several different systems?) as basis, as evidenced by the use of /β/, /ʔ/, palatalised velar stops, pharyngealised alveolar stops, perhaps /j/, and apparently, some also reconstruct /h/. Is there a consensus reconstruction that distinguishes all the contrasts reconstructed in the relevant (i. e., those which are in practical use for transcribing Proto-Berber forms), a kind of maximal system which can be losslessly converted into all the competing reconstructions? That's a simple practical issue which researchers need to keep in mind, as well. --Florian Blaschke (talk) 01:43, 25 September 2011 (UTC)
 * Also, it would be nice to know what König means with North, South, West and East Berber. --Florian Blaschke (talk) 01:45, 25 September 2011 (UTC)
 * I tried to clarify his use of names for subbranches. I don't think there is anything which could be called a consensus reconstruction. Personally, I would differentiate between "serious" reconstructions, in particular those by Prasse and Kossmann, and others, like those of Dolgopolsky and Militarev (Starling database), which in my personal opinion are to be treated with caution, since they appear to ignore basic insights of Prasse/Kossmann. In some of the earlier literature, "average" forms are used, but they don't show all the relevant distinctions, in particular for the vowels.--Schreiber91 (talk) 22:54, 1 March 2012 (UTC)
 * Thank you! So Prasse and Kossmann basically reconstruct an eight-vowel system, right? Short or ultra-short/reduced a, i, u, and long or unmarked a, i, u, e and o – where e and o are based on the evidence of Tuareg and Ghadames. However, in Tuareg languages, Prasse is said to claim that Tuareg e and o derive from Proto-Berber *i and *u (under unnamed conditions). Can you clarify this apparent contradiction? What exactly is the evidence of Tuareg for Proto-Berber *e and *o? (Personally, I wonder why the long vowels aren't marked as long vowels, which would enable to leave the short vowels unmarked: a, i, u, aa, ii, uu, ee, oo. Looks much clearer.) --Florian Blaschke (talk) 14:24, 31 March 2012 (UTC)
 * Exactly. Prasse wrote two papers on e/o, first arguing they are derived from i/u and also a (I think), and later arguing that they go back to Proto-Berber. I forgot what evidence Prasse mentions in his second paper, but I'll add some references to the article and will try to write more on the issue when I have more time. Yes, marking length would be clearer. One reason why it is not done seems to be that the early European study of Berber focussed on northern languages and considered them more representative of Berber. In Northern languages, the old length contrast does not exist any more, and Tuareg/Ghadames long vowels correspond to northern normal vowels. Second, Tuareg seems to have extra-long vowels, but I don't remember what their phonological status is.--Schreiber91 (talk) 11:34, 1 April 2012 (UTC)
 * Ah! So Prasse changed his mind. Thank you for fixing that!
 * Please do try to improve this article, too; it's so confusing now that even just a little work on the article will make it so much more tolerable. Yeah, I understood that the confusing notation is due to the tradition in the field, but sometimes you just have to dispense with legacies that hamper progress in a field.
 * (I was just going nuts thinking how Semitists and people working on other Afro-Asiatic branches and even Proto-AA reconstructions apparently think they can simply get away with ignoring vowels outright ... without even trying to exhaust every available shred of evidence – reconstructing from the bottom up if necessary, paying more attention to Modern South Arabian, Afro-Semitic etc. and de-emphasising Old Canaanite and Old South Arabian – to come up with consistent reconstructions that account for the vowels! At this primitive stage I'm not surprised that some more scrupulous historian linguists are very sceptical about AA and don't take it very seriously. If Indo-European philology had remained at that stage, it would never have made headway towards methodic stringency!) --Florian Blaschke (talk) 20:15, 1 April 2012 (UTC)
 * I definitely will work on the article, hopefully at the end of next week or so. I have to say that I'm really surprised by your impression of the attitude towards vowels. I think we have a fairly detailed idea of Proto-Semitic vowels and their role in morphology. I have to admit though that I don't know whether there is a consistent reconstruction of the development leading to Modern South Arabian and the southern Ethiosemitic languages (the development of the vowels in Ge'ez seems to be quite straightforward). Is that what you meant? (Old South Arabian, which you said should be de-emphasized, is virtually useless for reconstructing vowels anyway.) The reconstructions for other branches (with the exception of Jungraithmaryr et al.'s work on Chadic) include vowels. They are probably not very reliable and sometimes insufficiently rigorous at the current stage, but just the same applies to the consonants. It's clear that we are far away from having rigorous reconstructions for PAA and the three southern branches, and I do agree that we need more reconstruction from the bottom up (e.g. I think it's quite absurd that we have attempts to reconstruct Proto-Chadic and PAA, but no detailed reconstruction for the subbranches of Chadic), but I think there is no special problem with the way vowels are treated. But maybe I'm missing something?--Schreiber91 (talk) 21:47, 1 April 2012 (UTC)
 * I also added some remarks to what you said here.--Schreiber91 (talk) 21:58, 1 April 2012 (UTC)
 * Thank you; I have responded there.
 * Wikipedia might not be representative of the current state of the reconstructions, but the table here doesn't do much to inspire confidence. What I meant with my remark that while traditionally, Old Canaanite and OSA are accorded much weight for the reconstruction of Proto-Semitic because of their early and relatively extensive attestation, they are almost useless (apart perhaps from single words/glosses, loanwords and such, and Poenulus for Phoenician/Punic, as well as perhaps some of the Ugaritic evidence) for vocalic reconstruction (at least with respect to the short vowels), especially OSA, of course. As the vocalisation of Biblical Hebrew (at least with respect to the short vowels) seems to be based on some form of Aramaic, it can't be directly used for reconstruction, either, if my understanding is correct. That means that a reconstruction of the PS vocalism would be heavily based on Akkadian, Arabic, Aramaic, Ethio-Semitic and MSA. (A rather analogous situation is the reconstruction of the accent in PIE and subgroups such as Slavic, which must also apply different focuses than usual: OCS is central for Slavic reconstruction but provides no information on the Slavic accent.) --Florian Blaschke (talk) 12:41, 4 April 2012 (UTC)
 * Ah, no, the table is not representative, one could add vowels to the Semitic column, although they tend to be more controversial in the case of pronouns, where the evidence is a bit inconclusive. Maybe that's why the author chose to give consonsants only. The reconstruction of Semitic vowels is indeed based heavily on Akkadian and Arabic (and a bit of Ugaritic), but the developments leading to the other languages (with the possible exception of MSA and Sourthern Ethiosemitic) seem to be quite well understood (Semitic_languages has a nice overview, see also Semitische_Sprachen).--Schreiber91 (talk) 17:33, 4 April 2012 (UTC)
 * Fine, but how about the other columns? At least the Middle Egyptian vocalism has been reconstructed from what I've seen, but is not given, nor the Proto-Berber vocalism, let alone the other branches ... vowels are handwaved away at every twist and turn, recalling Voltaire's biting comment on the contemporary state of the art: "Etymology is a science where vowels signify nothing at all, and consonants very little." I find that table immensely frustrating as I am expected to believe in evidence which is cavalierly glossed over, without explaining why exactly that is permissible. --Florian Blaschke (talk) 16:20, 5 April 2012 (UTC)

Right, vowels could be added to some of the Eyptian words. Ehret, from whose book the table is extracted, sadly appears to have chosen not to inform himself about the reconstructed vowels of Egyptian (as Orel and Stolbova, who produced the other recent reconstruction) and preferred to use the vowels of Coptic. In the case of Chadic, the reconstruction of vowels just seems to be extremely hard and currently nobody seems to be interested in doing more work on Proto-Chadic (the vowels given in the table are just guesses). (In Berber *bn, the problem is simply that Berber verbs do not have stable vowels because of their root-and-pattern morphology, just as in Semitic) The problem is that work on Chadic, Omotic and PAA is still in a very experimental stage, and that reconstructing vowels is naturally far more difficult than reconstructing consonants. For instance, the time-depth of the several hundred Chadic languages, none of which was documented before the 19th century, might be comparable to that of Indo-European, and yet we only have four studies of Proto-Chadic phonology. Ehret is extremely confident about all the details including even tones, but in reality there is little in his PAA reconstructions that can be trusted.--Schreiber91 (talk) 17:26, 5 April 2012 (UTC)

Proto-Berber consonants
I couldn't find a full description of PB consonantal system in the sources available online, so I had to relie on indirect/partial descriptions. However, it's quite clear that some of the phonemes found in modern Berber languages were developped after the PB stage. There's also a tense/lax contrast (e.g. b ~ pp, d ~ tt, g ~ kk) not reflected in the table. Talskubilos (talk) 18:54, 16 October 2011 (UTC)
 * Great. Now the issues with the consonant table I saw above are solved. Thanks! --Florian Blaschke (talk) 16:47, 25 October 2011 (UTC)

Sisters or cousins?
In the second line of the article I read: "Proto-Berber was an Afroasiatic language, and its descendants (the Berber languages) are sisters to the Egyptian language, Cushitic languages, Semitic languages, ..." It puzzled me for a second. If Proto-Berber were a daughter of Afroasiatic, it would probably be a sister of any Proto-/Archaic Egyptian, Proto-Cushitic, etc. Then the daughters of Proto-Berber would maybe better be called cousins (in whatever degree) of the Cushitic languages, Semitic languages, etc. Wouldn't calling them cousins rather than sisters clarify the text? I think it would.Redav (talk) 23:34, 26 September 2013 (UTC)


 * I'm sorry you didn't get an answer for almost five years, but you are certainly right about that. It's a pretty basic convention for historical linguistics.--IfYouDoIfYouDon&#39;t (talk) 10:40, 7 May 2018 (UTC)

Proofreading needed
While reading this article with interest, I am noticing that there are parts that are in need of correction and stylistic improvement. There are some (minor but distracting) language issues and also some of the explanations are confusing and hard to follow as currently formulated. Unfortunately I lack both the time and background to attempt to revise it myself. Perhaps somebody could look at this? (For an example of what I am talking about, look at the sections on "History" and "Berber reconstructions". The clause "Proto-Afro-Asiatic is thus ten millennia B.C. prior" makes no sense, for example. Throughout these paragraphs there is furthermore a confusing switching back and forth between "B.C." and "BP" which is inelegant and taxes the mind of the casual reader. The section headed "Berber reconstructions" is almost impossible for me to make sense of as currently written, withough prior knowledge of the subject.) --A R King (talk) 09:54, 21 September 2015 (UTC)

My comment above referred to language, style and formulation, i.e. form not content. But there also seem to be some problems with the content. I cannot criticise the content in absolute terms because I lack the necessary knowledge, but the text appears to contradict itself: either some of the statements must be wrong, or need to be qualified, or else the way they are stated is too confusing. Again, here is an example I noticed: did Proto-Berber have case marking? I quote: --A R King (talk) 10:13, 21 September 2015 (UTC)
 * "Proto-Berber had no grammatical case." [OK]
 * "The Proto-Berber relics preserved at the lexico-semantic and syntactic levels show that the proto-Berber syntactic construction is of the ergative type." [That isn't grammatical case?]
 * "It [the apparent antecedent of 'it' is 'the proto-Berber statement core'] is not oriented in relation to its determinants (agentive subject, object…) whose syntactic functions are insured by casual elements including the casual affix (ergative) that indicates, as needed, the agent or the subject." [Huh...?]
 * That's because Allati, a researcher who has a view at variance with the better-founded one of Prasse and Kossmann, inserted his own POV into the article. See my discussion with above. --Florian Blaschke (talk) 20:39, 6 August 2019 (UTC)