Reconstruction talk:Proto-Afroasiatic/maʕiʒ-

RFD discussion: June–October 2021
This is not a reliable PAA reconstruction (such reconstructions are so few that we might be better off not having PAA entries at all). Instead, this is just copied from, which is a good scholarly work, but which must be interpreted with care. The EDE lists every possible comparandum, without necessarily endorsing them, but Fay Freak has blindly copied putative descendants that Takács calls "phonologically problematic" and "out of [the] question" (for Dahalo and Mehri respectively, both on vol. 3 p. 158). In fact, the EDE presents the Chadic and Omotic comparisons as alternatives, not part of a single PAA origin. The whole effort is related to an Egyptian word whose exact meaning is not even known (!), so words with a wide range of meaning can be added as long as they fit the vague form — never mind the fact that, for example, the second consonant (Egyptian /j/ and Semitic /ʕ/) don't actually correspond in regular inheritance. Perhaps many of these words are related; they may have spread along with the diffusion of goats in Africa. However, we can't erect a PAA entry here without playing fast and loose with the data. —Μετάknowledge discuss/deeds 20:52, 25 June 2021 (UTC)
 * No, I haven’t copied EDE. I have not at all started from it, but from the SED, where they do claim it, and later discovered that the form was also in the EDE, with just some additional comparisons, including the Egyptian. The Dahalo form EDE claims borrowed from Nilotic is different from the one given here from the SED. I haven’t included any Mehri but Shehri with marked uncertainty. Omotic is always problematic, because perhaps it didn’t exist but is polyphyletic; but note that PAA was spoken before the domestication of the cow, therefore we can subtract the meaning difference without any loss. For the Chadic it could only be that some are related but not others/there have been mergers with similar or borrowed terms, but you can’t know which, I took this into account. The Egyptian is at least known to be a kind of horned bovid? How exact does one get with less common Egyptian animal names? The wide range of meaning I do not see, it seems moderately narrow. If the meaning correspondence fits closely then it is a wanderwort. So that’s how you delete everything. If it is looks too good it can’t be, if it is problematic it can’t be either. I guess there must be a pinch of problematicity.
 * I wish man would approach our Indo-European entries with the same insight that one lists every possible comparandum, without necessarily endorsing them. The bulk of PIE roots needs to be deleted, yet they aren’t because they use to be somewhat possible. Many of such entries however are even less and someone just took it over and made an entry as soon as he saw an asterisk, which I didn’t. Blindly copying is correct, but blindly deciding? First I decide it as plausible, then I blindly copied—before seeing the EDE. I know that one can’t start from the EDE, but when this is claimed and I do not likewise have weighty reasons to rebuke it we can list the claimed form cum grano salis. This is what’s done with, and, dude, who believes in ? Why don’t you RFD ? It is a convenient device to list comparanda apparently.
 * We could peruse to deploy reconstructions with three stars (meaning “possibly existed, but I don’t know”), but haven’t the technical framework for it so far, we only have this one reconstruction namespace for imagined forms.
 * What is with that other word for “goat” anyway, which this looks related to? 🇨🇬 and so on; it is also reconstructed for PAA,, there however the PAA is distinguished as “possible”, unlike this ma- form. Fay Freak (talk) 21:52, 25 June 2021 (UTC)
 * Currently, PAA reconstructions on Wiktionary are limited to four lemmas. We don't have basic vocabulary (such as Swadesh list): pronouns (but one), kinship terms, numerals, intangible concepts. However there are serious cognates of such lexical items. Why would a lemma for "goat antelope" have its entry whereas more relevant and reliable (ie. better preserved) don't? As a matter of fact, the descendants are problematic (treated alphabetically):
 * Berber: it is the sole family absent from the reconstruction. As far as I know, this family has a quite rich vocabulary about goat and this kind of domesticated animals.
 * Chadic: though the PC reconstruction seems plausible according to the claimed descendants, those are all redlinks, and that's why we don't have PC entries (or one for water, if it wasn't deleted). Why creating an entry for a stage beyond one having already its issues (scarcity of material and serious reconstructions)? The Hausa lemmata are likewise absent of the Wiktionary (I'm not saying they are wrong, but users can't rely on attested descendants if they don't have their entries).
 * Cushitic: Though Oromo, Somali and Afar are quite well attested, the sole proposed reflex is in Dahalo.
 * Egyptian: as Metaknowledge noticed, the precise meaning of mjzt is not known, and moreover the tentative one is not indicated on the entry. Furthermore, it's phonologically problematic for the laryngeal. I don't know much about evolution of Egyptian phonology from PAA, but the laryngeal would have been preserved (rather than reduced to a yod or lost).
 * Omotic: for once there's the Omotic family in a PAA reconstruction, however I remember that yourself intended to show me that Omotic is likely not Afroasiatic at all, which is contradictory. And there's nothing on the entry to indicate the uncertainty.
 * Semitic: The development from the PAA to the PS reconstruction seems regular. And only here. Maybe PS was the most conservative descendants of PAA, but the reconstruction seems Semitic-based. Its descendants are seldom (nothing in East Semitic, nor in South Semitic, nor in Canaanite), yet some poorly attested languages (e.g. Safaitic) have a reflex. What's more, most of the precedent tree disagreed with the current standardized one (just following "About Proto-Semitic"). So either we have to review our standard tree, or borrowings weren't indicated on the PS entry, and in this case it's rather a Wanderwort.
 * The phonology of the language is highly disputed as well as the classification of its descendants, but anyway this applies to all PAA entries. I agree with Fay Freak, the semantics are ok (given the context). When it's good, we keep it, and there's no "too good", unless it causes the entry to be suspect... May I propound that the entry is kept in an appendix or the like? I've noticed that there are around 30 and 50 lexical items having good Afroasiatic cognates, so that it would be cool to have some place to gather them, without necessarily meeting the standards (since they aren't entries there). Then it goes beyond the scope of the issue here. For instance having some "Proto-North-Afroasiatic" grouping Semitic, Egyptian and Berber so that the few material in Chadic and Cushitic wouldn't be a problem to make reconstructions (and also because those three families share great affinities).
 * Still beyond the question of the entry, the concept of deciding involves choice: you decide to blindly copy; you don't blindly decide to copy. As for the PIE entry you mention, I believe it would be better to delete all those displaying *a (unless there's a velar next to it or that inserting the second laryngeal wouldn't violate phonotactics, as in ) or having reflexes only in half of descendants families for the same reasons as, but here's not the right place to take these decisions (and I'm not expert of PIE). Malku H₂n̥rés (talk) 21:10, 27 June 2021 (UTC)
 * №37, as *maʕi(n)ʒ- ‘female bovid’. I'm also interested in the fact that you write that this is a noun, not a root. It is unlikely that bare root can have exact semantics. In a more complete article, the form is already given as follows *ma-ʕi(n)ʒ- ‘k. of bovid’; عنز < *ʔ/ʕi(n)ʒ- ‘k. of smaller bovid’. A note to the Egyptian form is written Note lack of -ʕ- vs. the Sem. forms. Gnosandes (talk) 10:19, 26 June 2021 (UTC)
 * In Afro-Asiatic unlike in Proto-Indo-European the trailing hyphen does not signify that the form is a root. It stands for omitted endings. (Which you find in declension tables in Semitic nouns here. The Proto-Semitic verb endings are more difficult to reconstruct due to later developments. We use the “suffix conjugation” for lemmata, while the prefix conjugation (the perfect; the “jussive” of Arabic verbs which is used with to negate the past) had no inflectional suffixes in the basic form. Which also is to say that we are very far from reconstructing verbs for Proto-Afroasiatic.)
 * There was little reason to give *maʕinʒ- instead of *maʕiʒ- since the nasal is virtually nowhere present, but apparently there to compare with the other goat word (*ʔ/ʕi(n)ʒ-). Fay Freak (talk) 13:16, 26 June 2021 (UTC)
 * Ok, although I just offered you a little more materials. Gnosandes (talk) 13:26, 26 June 2021 (UTC)
 * Sorry my ping failed. Please read (and answer) the long message above. Thanks to for WT:Beer parlour/2021/June. Malku H₂n̥rés (talk) 08:10, 28 June 2021 (UTC)
 * We seem to have a significant philosophical difference. You think that protolanguage lemmas can serve as simple storehouses for forms that look vaguely similar, even if the match in meaning is inexact or the sound changes don't match established ones. I think that protolanguage lemmas should be treated as serious entries in their own right, and only be created if there is a good basis to reconstruct them in the first place. Wiktionary practice, despite certain questionable entries, is on the whole toward the side of treating protolanguages seriously, as I advocate. —Μετάknowledge discuss/deeds 05:45, 29 June 2021 (UTC)


 * RFD-deleted. —Μετάknowledge discuss/deeds 22:11, 20 October 2021 (UTC)