Wiktionary talk:About Proto-Semitic

, just reminding you about this unfinished page... —Μετάknowledge discuss/deeds 20:41, 4 March 2016 (UTC)

Proposals for standardisation of Proto-Semitic
The lack of any content on this page has made it difficult to standardise entries appropriately. I have the following suggestions and questions that we should come to a consensus on, add to this page, and then add to the entries themselves.
 * 1) The "Onomastics" (and similar) sections should be removed, as they do not support the PS entries and are not allowed by WT:EL.
 * 2) All entries should have a reliable reference supporting the reconstruction (not macrocomparativist literature).
 * 3) All entries should have inherited reflexes in both West and East Semitic, or else indisputable inherited cognates in AA. If they only have reflexes in West Semitic, they may be suitable for a Proto-West Semitic entry.

And some questions:
 * 1) Do we want to include Huehnergard's proposed *x̣ (which he writes as *xʼ)?
 * 2) We currently have *-u(m) in the nominal inflection table; is there any reason we don't simply separate the construct state and absolute state in the table?
 * 3) What should the standard for verb lemmata be?

—Μετάknowledge discuss/deeds 20:59, 2 August 2020 (UTC)


 * No to all.


 * 1) That you have recently found retarded irrelevant onomastics sections does not mean it is always bad. I like them (especially toponyms) sometimes on plant name entries, sometimes they do support (haven’t done this on Proto-Semitic as I remember, but on Proto-Slavic and Proto-Germanic variously, and there could be a standard for it).
 * 2) Bad reconstructions generally come from people adding things because they have found them in references rather than from assessing in how much they would themselves posit the reconstruction in question – isn’t that your preponderant observation, too? Naturally I disagree because I have many agreeable Proto-Semitic and other proto-entries without references, just stumbling upon the need to have an entry centralizing cognates, and assessing the probability of a Proto-Semitic form is rather easy if one has a sense of proportion. One just needs to know when one needs to stop. In reliable sources too there are starred things which should not be created. In other words the problem is less in people’s literature use but in lacking brains use. Referencing on this occasion
 * 3) As I have said already, I don’t believe in any larger division of the Semitic languages, or in the existence of a Proto-West Semitic any more than in a Proto-West Slavic. Some have argued that Old South Arabian was actually the first to split. I have sometimes been content with stated Proto-Semitic terms being present in Ethiopian, Old South Arabian, Arabic, Northwest Semitic, when it seems that Akkadian has innovated by replacing the Proto-Semitic term with a different word. The reader may himself see how much the likelihood of the Proto-Semitic term having existed is decreased by just the Akkadian lacking. What’s with ? Go back to cognates lists? It seemed to me it is still over the acceptable likelihood threshold. Sometimes the tolerances differ. I found  risky but it’s still okay on the other hand. BTW I see an analogy to Proto-Turkic pages still being created as “Common Turkic” when a Chuvash descendant lacks. One avoids to depend on one language.


 * Questions:


 * 1) Never seen that. And I am skeptical about all the attempts to get more than 29 consonants. When I say reconstructing Proto-Semitic is rather easy I refer to a notion that the phoneme set of Proto-Semitic makes the easy part.
 * 2) Probably just coding capabilities.
 * 3) I don’t see a current problem. Do you mean anything specific? The current formatting seems okay to me. Fay Freak (talk) 22:09, 2 August 2020 (UTC)
 * Suggestions: 1. Can you point out a PS example where the Onomastics section is useful? If a reflex is only found in proper nouns, those should be listed as descendants. 2. I don't think the reference needs to give the exact reconstruction, but if a comparison of the terms hasn't been considered by Kogan, Leslau, etc I would be suspicious, in that I doubt we can come up with anything they haven't written about. 3. I believe and  should be PWS. I don't see any case you have made to dispute the East-West division, which is consensus among workers and accords with an analysis on morphological grounds. Questions: 1. Then you should read about it. Huehnergard 2003 makes the case for it and gives a list of examples. —Μετάknowledge discuss/deeds 23:02, 2 August 2020 (UTC)
 * 1. I have said I remember there haven’t been Proto-Semitic examples yet. I can point to Proto-Slavic examples,  and ; I can’t exclude future use for Proto-Semitic, though this is less likely. But yes, I list them as descendants (in a simulated section, to cram the common noun descendants not). I thought the dispute point is not how to list them but whether to include onomastics. Which I don’t really see a need to discuss for Proto-Semitic specifically. You did not need to have special clearance to remove silly stuff like on.
 * 2. I don’t find it suspicious because one hasn’t yet tried to be complete or something, and it seems that the etc. behind Leslau and Kogan is rather short, it ends with just few, few have tackled the task of reconstructing Proto-Semitic; say comparing to Indo-European studies. Anyway in as much as references are desirable there shan’t be a dogm that referenceless reconstructions are to be deleted, but the principle of Wiktionary as a secondary source stays, assessing the merits of reconstructions as needed. You don’t point out a need that one did not know about – I also like books, innit –, it’s just signalling how great scholarly references are to maintain status in the academia class, for which it won’t be grateful.
 * 3. x̣ might be a nice fiction to create entries where Akkadian has ḫ and West-Semitic ḥ, although there might be other reasons for the variation: dialectal variation in Akkadian (there was also chronolectal one in the appearance of ḫ); or more cases of sporadischen Lautwandel in den semitischen Sprachen where one knows that there is a relation but does not know the original.
 * You have seen a consensus, the wording in this Huehnergard paper is surely strong. I have seen other things as said, and especially what I have not seen: I have perhaps never seen Proto-West-Semitic being claimed with such determination, which was however useful for the argument in that paper. It always looks like that when you read some papers which are similar in some points, it’s called . However the problem is: what would distinct grammar of Proto-West-Semitic be, as opposed to Proto-Semitic? It would not be a distinct language. Actually Proto-West Semitic, if we posit it, is Proto-Semitic. Fay Freak (talk) 02:17, 3 August 2020 (UTC)
 * It looks like despite all the excess verbiage, we're essentially in agreement on the first two. As for *x̣, it does not depend on the East-West split, but that split makes it maximally parsimonious. Naturally it is a "nice fiction"; all of a protolanguage is a fiction, but we try to make that fiction as neat and close to the truth as possible. I don't know what you're reading, but Kogan's Genealogical Classification of Semitic, which is one of the most critical of the East-West divide in mainstream scholarship, says that this split is "one of the least controversial and most broadly accepted subgrouping hypotheses" (in his chapter 2, which I recommend you read). Perhaps the selection bias is operating on you as well? —Μετάknowledge discuss/deeds 07:43, 3 August 2020 (UTC)
 * Not pinged nor currently active in editing Semitic languages, but I strongly support suggestion 3 (require inherited reflexes in both West and East Semitic or uncontroversial AA inherited cognates) and am somewhat supportive of suggestion 2; I think that it should not be necessary if there are many independent descendants (though in such cases it is likely that there are reliable references), but otherwise very desirable to require it. ←₰-→ Lingo Bingo Dingo (talk)  09:59, 7 August 2020 (UTC)


 * I can only touch on point 3 right now, although I’ll share my thoughts on the rest later. Comparative reconstruction requires two independent pieces of evidence for any item being postulated for a proto-language. East and West Semitic are the most widely accepted higher-level branches of the Semitic language family. If Wiktionary’s Proto-Semitic policies are to reflect the academic consensus, then it necessarily follows that any item postulated for Proto-Semitic must be confirmed by a witness in both East and West Semitic, unless a higher level (ie. Afroasiatic) connection can be conclusively demonstrated for an item attested in a single branch. Seeing as how regular sound correspondences between the branches of Afroasiatic have yet to be established and the relevant sound change laws therefore remain undescribed, this particular route is impossible at the moment. On that basis I support point 3. Rhemmiel (talk) 15:17, 8 August 2020 (UTC)


 * , I neglected to respond regarding my question 3 (presentation of verbal stems), so I will do so now. It is clear that there is a problem: nearly all of our entries have the form *CaCaC-, as if someone simply took the Arabic lemma form and lopped up the final -a. Firstly, Arabic uses a suffix conjugation that cannot be reconstructed to PS, so this form is a bit anachronous. More importantly, this completely hides the thematic vowels, which are lexical and not currently even presented in the entry, which is a glaring omission. For example, our current entry has a thematic *u in the short prefix conjugation, such that *-ʔkul would be a far better lemma (although the thematic vowel of the long prefix conjugation, in this case *a, would still need to be mentioned in the entry). —Μετάknowledge discuss/deeds 18:43, 8 August 2020 (UTC)
 * The suffix conjugation cannot be reconstructed? Very new view to me, I use to think it was there but not with the same purpose as in Arabic (not a “perfect”). I read however that the geminate stem ( in Arabic) was CaCCiC- in the suffix conjugation and Arabic and Ethiopic innovated them to CaCCaC- by analogy; base stems and geminate stems are the only ever verb forms I have created or linked. Inacquaintance with Akkadian usage makes the Proto-Semitic picture of verb morphology opaque to me, and such jargonized general linguistic papers help little on it: . So what forms do you reconstruct?
 * Well, you can just give the prefix form in the header? I think one reason for using the current forms is that the individual language dictionaries are based on them. Fay Freak (talk) 20:18, 8 August 2020 (UTC)
 * Indeed, I found the reconstruction of the PS verbal system very confusing, and to be the clearest explanation. Of course, Akkadian weighs heavily on this reconstruction (as it does for lexical reconstruction, given that I accept the primary East-West split), but in its broad strokes, it seems to accord with what I see elsewhere. Among those broad strokes is that there are two different prefix forms, each with their own thematic vowel. I think that the PS entries should function as a good dictionary for PS, treated as its own language, rather than following conventions that are alien to it, and you will note that Kogan cites verbs as I suggest, with the stem of the short prefix conjugation. —Μετάknowledge discuss/deeds 23:31, 8 August 2020 (UTC)
 * , any more thoughts on this? For the verbs in particular, I need to know what form we will use in the page title before I can make a conjugation template. —Μετάknowledge discuss/deeds 17:54, 15 August 2020 (UTC)
 * How come “per my own answer” ? We didn’t, but then already I declared it nice, and here and there words came to be mentioned with that character. And I reconstructed way more casually back then, and you gave two even older pages as counterexamples, which I of course had moved had I noticed them, having the impression that we agreed. Too bad only that at that point we have given up to know how the Proto-Semitic phonemes sounded. Fay Freak (talk) 18:52, 5 July 2021 (UTC)

both of you, Fay and Meta, decided the standard with your discussion above:
 * M: "Do we want to include Huehnergard's proposed *x̣ (which he writes as *xʼ)?"
 * F: "Never seen that. And I am skeptical about all the attempts to get more than 29 consonants. When I say reconstructing Proto-Semitic is rather easy I refer to a notion that the phoneme set of Proto-Semitic makes the easy part."
 * M: "Then you should read about it. Huehnergard 2003 makes the case for it and gives a list of examples."
 * F: "x̣ might be a nice fiction to create entries where Akkadian has ḫ and West-Semitic ḥ, although there might be other reasons for the variation: dialectal variation in Akkadian (there was also chronolectal one in the appearance of ḫ); or more cases of sporadischen Lautwandel in den semitischen Sprachen where one knows that there is a relation but does not know the original. You have seen a consensus, the wording in this Huehnergard paper is surely strong [...]"
 * M: "As for *x̣, it does not depend on the East-West split, but that split makes it maximally parsimonious. Naturally it is a "nice fiction"; all of a protolanguage is a fiction, but we try to make that fiction as neat and close to the truth as possible."

You, Fay Freak, are skeptical about the addition of *x̣, whereas Metaknowledge refers to a (reliable) book. Note that I had to search here in the talk page because I remembered this, which is not indicated in the table of About_Proto-Semitic. Yes I admit I blindly followed the standard when I moved (I realized after that it wasn't a new lemma but one you moved). So it's time to discuss a bit more, perhaps to finally add *x̣, since you say *ḥ develops another way in Akkadian. Malku H₂n̥rés (talk) 07:33, 6 July 2021 (UTC)
 * Well, if we use that then the worst thing that can be is that it is (or turns out) synonymous to ḥ.
 * And it is not in the table since, ehm, the table only includes sounds of known articulation. Fay Freak (talk) 18:18, 6 July 2021 (UTC)

TBH I'm not against *x̣! Though I believe it's more a nice fiction to account for a badly understood Akkadian (or East Semitic) sound shift, it's a simple and scholars-supported solution, which at least avoids confusion with a conservative development of *ḥ remaining so in East and West Semitic. It's just that our standard doesn't even mention it. As for it's value, you say it's unknown. Yet I doubt Huehnergard chose the letter  for unknown... Malku H₂n̥rés (talk) 21:09, 6 July 2021 (UTC)
 * If you know it and are so sure of it being a simple solution widely supported why don’t you put it into the table 🤗. Fay Freak (talk) 21:19, 6 July 2021 (UTC)

To be absolutely honest, Metaknowledge told me one year ago that I wasn't ready to create PS entries, and generally I consider I should just follow what more knowledgeable people decided on wiki (that is, the standard established by you and him), until further notice. The second reason is phonological: for /x/ we use *ḫ, Huehnergard uses *x, and *ḵ is the most coherent as it's the fricative counterpart of *k, therefore I think we should use *ḵ̣ with both diacritics, just like *ṯ̣. Malku H₂n̥rés (talk) 21:51, 6 July 2021 (UTC)

sorry I forgot to ping you. Malku H₂n̥rés (talk) 21:51, 6 July 2021 (UTC)
 * I didn’t like the sign itself from the beginning, however even Huehnergard, whose paper Metaknowledge linked above two years ago, uses *ḫ where we do, so his x̣ is not based on him using x, and I didn’t consider a visually distinguishable alternative, and now I think of h with ring below, h̥, cool? Almost like ḥ but with small difference for the initiated. Fay Freak (talk) 22:28, 6 July 2021 (UTC)


 * When in doubt there’s always the good old fashioned subscript route. ḫ₂/ḥ₂ might be something to consider for stylistic consistency and phonetic neutrality. Rhemmiel (talk) 12:54, 7 July 2021 (UTC)

I had been waiting for 's proposals for the sign to use, but neither him nor answered, unlike  who I didn't pingn Perhaps  would have interesting propositions to say. For a sign replacing *x̣, I'm not a fan of 's *h̥, too close to *ḥ. Underdot isn't used in PIE because ring below is used consistently (for syllabic sonorants), and conversely we shouldn't use ring below because we use underdots consistently (for emphatic consonants as well as *ḥ). As for Rhemmiel's subscript, the idea of Huehnergard is that a single phoneme fully merged with *ḫ in East Semitic and with *ḥ in West Semitic, whereas using ḫ₂ and ḥ₂ make two phonemes. Perhaps just *h₂ is good. I propounded *ḵ̣, which got ignored whereas it's the regular phoneme following our rules of romanization. Malku H₂n̥rés (talk) 10:01, 9 July 2021 (UTC)
 * We can’t use *ḵ̣, because a macron with a dot is terrible to the eye, and because we don’t even use *ḵ (k with a macron below). I have yet not succeeded to see its regularly following our rules of romanization.
 * I foresaw the argument of PIE using the ring for syllabic consonants. Everything is used for something somewhere else it would not have been encoded. There is a list of to choose from. Fay Freak (talk) 14:31, 9 July 2021 (UTC)
 * I wrote out a response earlier and completely forgot to save the page, so apologies about that. Malku, how can you claim that this is "scholars-supported" when it remains contentious in the community? We can commit to a Huehnergardian PS, as is neatly outlined in The Semitic Languages, or we can choose not to; either way, we should not deceive ourselves that we are following consensus when no such consensus exists. Fay, your claim about *x̣ and its phonetic interpretation is unfounded, and based on misinterpreting Huehnergard's changing notation. He does believe it to be an emphatic form of /x/, and therefore writing it with *x̣ is the only reasonable option. (Also, making up our own signs is a terrible idea in general.) —Μετάknowledge discuss/deeds 16:44, 9 July 2021 (UTC)
 * I am not informed though about the contentions, or what one could have contended against. Has one answered him at all? How has the story continued?
 * Well making up signs is a terrible idea in general, as it only creates divisions with the rest of the field. So I already went with x̣, but Malku H₂n̥rés wanted to play through the alternatives first 😜.
 * My note about the uncertainty of the phonetic value of *x̣ is not unfounded though. There are multiple options even if Huehnergard itself picked a favourite one. The more so if its very existence is contested. Fay Freak (talk) 18:17, 9 July 2021 (UTC)
 * : Read the Handbook p. 84–5 for a quick summary. Those who disagree seem to generally take the tack that there are very few unquestionably valid examples of this correspondence, and they would rather take those as unexplained shifts from *ḥ than posit a new, uncommon phoneme in PS. I do not claim to be able to adjudicate this; I merely want us to be consistent, document what we do, and be conscious of the logic behind our choices. —Μετάknowledge discuss/deeds 18:52, 9 July 2021 (UTC)
 * One of the two signs would have to be settled on, of course. But generally I agree with that making up our own transliteration symbols is not a good idea. Rhemmiel (talk) 09:36, 11 July 2021 (UTC)

it's supported by Huehnergard only, the -s for "scholars-supported" was unnecessary here. I've said that "I believe it's more a nice fiction to account for a badly understood Akkadian (or East Semitic) sound shift" and that *x̣ was an easy, but unreliable, solution. So either we keep *ḥ until one day something can account for this apparently-sporadic shift (which I support), or we use some sign standing for an emphatic velar fricative, assuming the nice fiction that it completely merged with *ḫ in East Semitic and with *ḥ in West Semitic. Malku H₂n̥rés (talk) 12:21, 11 July 2021 (UTC)

Noun-forming suffixes
Since noun-forming suffixes would be followed by inflection endings, shouldn’t and  be  and ? In other (proto) languages the lemma forms of noun-forming suffixes even include inflectional suffixes, such as : that is of course because the noun entries also include inflection endings, nominative singular masculine being the citation form in this example, and in analogy to it we would have to have the Proto-Semitic noun-forming suffixes hyphenated at the end like the noun entries are. What do I miss? That they unfortunately do not look like suffixes then? Fay Freak (talk) 17:10, 30 June 2021 (UTC)
 * You are definitely right. It looks ugly, but you are justified in adding a hyphen at the end. —Μετάknowledge discuss/deeds 17:29, 30 June 2021 (UTC)
 * I agree. Since noun entries end with a hyphen, including the ones with these suffixes, those same nominal suffixes should end with a hyphen accordingly. And its way easier than adding a parameter to the declension template, which wouldn't be very used. Malku H₂n̥rés (talk) 17:43, 30 June 2021 (UTC)