Talk:Afrasianist phonetic notation

Is there really?

I placed a template on the article requesting better sources for the Afrasianist phonetic notation, but there is the underlying issue that we possibly should not have this article at all. I am somewhat sceptical that there really is something like a notation convention for linguists beyond the IPA. I know that this is frequently claimed, for example also for the field of Ethiopian linguistics - but the facts are usually that most linguists claiming to adhere to some such convention tend to use something very particular to their own work, and for which it is often very difficult to find any other linguist applying the same rules. This is one of the more frustrating consequences of working in a field that staunchly resists the acceptance of the IPA standard - it is next to impossible to ascertain pronunciations if no reference to the IPA is made in the publications. There is usually not any kind of agreed upon document that codifies the so-called convention. I am not aware of any such document for the Afrasiatic notation, and such a document, with sufficient documented endorsement by a good number of linguists, would have to be the basis for making the matter notable enough to warrant a page on Wikipedia. Practically speaking, if only Militarev and Stolbova use that set of symbols in this way (and they don't explain it on the linked webpage), then this is definitely not enough to save this article from eventual deletion. LandLing 13:23, 6 June 2021 (UTC)

Note that this is the convention used in our Proto-Afroasiatic language article.

I'd be happy with listing variants. The point is exactly that "it is next to impossible to ascertain pronunciations if no reference to the IPA is made in the publications," and as you pointed out, often no explanation of symbols is given. There's a lot of variation in Uralicist and Americanist notation as well, and of course even in IPA, so that's not much of a counter-argument. Anyway, it is extremely common for e.g. ⟨ʒ⟩ to be [dz] and for some diacritic to be used to convert central to lateral, e.g. ⟨ŝ⟩ or ⟨s̰⟩, and there should be some guide somewhere for people trying to figure out what all that means. — kwami (talk) 04:58, 14 June 2021 (UTC)

In theory I agree with you. There is no doubt that things like the Afrasianist notation exists as a loosely held concept, but any attempt to describe it, as the one you have put into that table, amounts to original research. Because the authors do not explain their usage, for the most part we really don't know what the symbols mean, except by going through a very tedious WP:OR synopsis. LandLing 11:18, 14 June 2021 (UTC)

Some of them do explain. They may not give IPA equivalents, but they do label phones by manner and place of articulation -- just as the IPA does -- and break down affricates into their plosive-fricative components. A bit tedious to go through, but fairly straightforward. — kwami (talk) 03:03, 15 June 2021 (UTC)

I'm also skeptical of this, tho for a slightly different reason. I don't think that there is a field standard: I'm not even sure there really is a field. There is significant overlap in the notations used by the handful of Afro-Asiaticists out there, thanks to the influence of long-standing Romanisation practices for the Semitic languages & the IPA, but the notation in the Etymological Database is not that in Militarev & Stolbova's earlier work is not that in Diakonoff 1988 is not that in Ehret's work is not that in Bomhard's. I don't think we have any citation of a work which describes what Afro-Asiaticists in general do. I think we probably shouldn't have this article. Pathawi (talk) 19:24, 19 January 2023 (UTC)

It would be fine to expand on the variation, but we should have something. I created this article because I was often unsure what AA symbols were supposed to mean. We could have multiple tables for different sources, or list "X or Y or Z" for each (proto)phoneme, but we should have some answer for other people who aren't familiar with the AA lit and want to look this up. — kwami (talk) 19:50, 19 January 2023 (UTC)

One thing we could do—& I could contribute to this—would be to make a table of IPA values posited to exist in the various reconstructions as rows & the graphemes used by different (notable?) scholars (or notable works, if it turns out that scholars aren't consistent over time) in columns. As the reconstructions aren't identical, not ever cell would be filled (& some rows would have only one cell filled), but it would be fairly straightforward, & I don't think would fall afoul of OR any more than "routine calculations" do. Pathawi (talk) 22:42, 19 January 2023 (UTC)

One problem I see with that is that they don't necessarily propose an IPA value, as with Egyptian ⟨h̠⟩. Even if a value is posited, the symbol for a protophoneme may be independent of it. Different researchers may have different ideas of what the protophonemes were but still use common conventional symbols for them. I think it might be better to start with the symbols, and give the various interpretations associated with them, than the other way around. — kwami (talk) 04:36, 20 January 2023 (UTC)

Egyptian ‹ẖ› isn't reconstructed for PAA as such with this grapheme by Diakonoff (‹x›), Orel & Stolbova (‹ḫ›), Ehret (‹x› or ‹xʷ›), or Bomhard (‹x›). Who's using this for PAA? Or are you proposing that the language-specific Romanisations for all of the daughter languages should be included as well? Pathawi (talk) 05:47, 20 January 2023 (UTC)

I'm proposing that common AA transcriptions be shown. Not necessarily reconstructions, but also comparative vocabulary. As for who uses it, the refs are examples. — kwami (talk) 12:22, 20 January 2023 (UTC)

That seems like way too much, probably also redundant for Wikipedia, & ultimately indeterminable. Each Afro-Asiaticist is drawing on a great number of languages. Romanizations are not fixed for all of these languages. Conventions vary a great deal from language to language. If you wanted to just focus on the graphemes employed in reconstructions, you'd be dealing with somewhere under fifty signs, among which there are some clear patterns, tho no two linguists use the same set of signs. The page as it is is almost there, but I think it misrepresents the nature of "convention". If you're trying to reach out to the comparative vocabulary that actually appears in these works, you have to account for the major rôle that Egyptian, Semitic, Somali, Bidhaawyeet ("Beja"), and Chadic play thruout (as well as the secondary rôles of Omotic & Berber), & what's here right now just begins to scratch the surface. I'm having a hard time imagining a useful way to do that. Pathawi (talk) 15:43, 20 January 2023 (UTC)

I don't mean transcriptions for individual languages, but common conventions used so that the languages can be directly compared. Basically what we have now: non-IPA symbols for AA as a whole and what people use them for. — kwami (talk) 21:34, 20 January 2023 (UTC)

Then ‹ẖ› isn't part of that. It's not reconstructed in PAA. It's used for Egyptian, in both major transliterations. The Afro-Asiaticists use evidence from Egyptian. Similarly, they use ‹ꜣ› & ‹ꜥ› & ‹ı͗› for Egyptian, ‹ʿ› & ‹ʾ› for Semitic languages, ‹ä› by some for Ethio-Semitic languages, Ehret uses ‹dl› for Cushitic… I don't think the category that you imagine actually reflects something concrete & identifiable in the core body of literature on Afro-Asiatic linguistics. If it's not possible to pin it down, then it shouldn't be the topic of an article. Pathawi (talk) 01:41, 21 January 2023 (UTC)

That's why ‹ẖ› isn't listed. — kwami (talk) 01:09, 22 January 2023 (UTC)

I'm having difficulty following this: You adduced ‹ẖ› as an example of a sign for which an IPA value isn't proposed. & it actually is listed.

It would, in fact, be relatively straightforward (tho it would be a bit of work) to chart out the overlap in reconstruction & the varying systems used, & then to note the relatively few reconstructed phonemes for which an author (or pair of authors) is unable to specify an articulatory value. But it's worth looking at the actual sources. Diakonoff 1988 makes preliminary hypotheses (often stated quite confidently) for all 44 consonantal phonemes he reconstructs. I'm not wedded to the idea that representations need to be through IPA, but the very idea that a "conventionalized set of symbols" is describable for Afro-Asiatic studies presumes that these systems have a degree of commensurability, which suggests that a comparative presentation should be feasible.

What I'm objecting to is two interrelated things:

1. I do not believe that a conventional or "semi-conventional" transcription exists. I believe that each author has their own transcription, but that due to shared influences these systems overlap. Because there is no general system, a reader cannot use a page like this to help them read material from actual Afro-Asiaticist sources: They simply have to know the specific notation employed by the specific author. A comparative page could allow that. I don't think a general description can. A couple examples:
   • You have ‹9› listed for /ɡ͡ɣ/. I think this isn't quite right: In Diakonoff 1988, it's /g͡x/ or /ɣ/. /ɣ/ is ‹ġ› for Orel & Stolbova 1995. By the starling.ring.ru database, they use the IPA /ɣ/, as do both Ehret & Bomhard. You've got ‹ḡ› for /ɣ/. That doesn't seem to appear in any of the sources listed for the page.
   • You list the convention of showing palatalised consonants with an acute accent, which I gather you're taking from Dolgopolsky's Nostratic work, which shouldn't be considered the same as the Afro-Asiatic work. I don't think that any of the writers I'm citing reconstruct this sort of palatalisation for Afro-Asiatic beyond the alveolars. When Bomhard talks about palatalisation in branches of Afro-Asiatic, he annoyingly uses ‹xʸ› (where ‹x› is any letter). I don't think that adding more symbols fixes this: It just creates a jumble, & shows that there is no conventional system.
2. If we take this article title to refer to a thing, that thing is not notable. There is no source that mentions it. As it stands, I think this qualifies as OR. A comparative table would not claim that a convention (or "semi-convention") exists, & would thus look a lot more like a routine calculation. Pathawi (talk) 02:56, 22 January 2023 (UTC)

‹ẖ› is not listed. There's just a comment about it. Could be removed w/o affecting much.

"Semi-conventionalized" may still be an overstatement. That's not really the point though. The point is to provide an explanation for symbols found in AA comparisons and similar works.

If this isn't a "thing", what "thing" would people look up to get this information? How are people supposed to know that a circumflex means lateral, if the paper just assumes it will be understood and doesn't explain their symbols? — kwami (talk) 04:36, 22 January 2023 (UTC)

They would find out what the source they were looking at was, & then look at a table which showed them that Diakonoff & Oret & Stolbova used a circumflex on three consonant signs that were laterals… & they'd also see there that Ehret & Bomhard marked laterals in completely different manners, so they'd be able to recognise when two different authors were doing the same thing with different notation. If you see a paper in the Journal of Afroasiatic Languages and Linguistics in which there is a circumflex over a letter, you actually cannot safely assume that it's a lateral or lateralised without a specific indication of the transcription scheme in use. Do you have a particular paper in mind? Pathawi (talk) 05:11, 22 January 2023 (UTC)

Yes, symbols may have multiple uses. That's why there may be more than one value listed per symbol.

What "table" would they look at? That's what this article is supposed to be. If we delete it, there's no table.

As an aside, you keep mentioning Ehret, but we shouldn't report him any more than necessary. His work is notoriously bad and not influential enough to be particularly notable.

I'm looking at Obst and Dempwolff from the 1910s and Berger from the 1940s; in all three, <s̰> is a lateral fricative, but only Berger explains that -- for the other two, you need to know the language to interpret the symbols. In Obst, a circumflex is an ejective, so <š̂> is [tʃʼ]. That can get quite confusing. Even if you don't have a unique ID for a symbol, a list of uses can often narrow things down enough that you can make sense of a paper. — kwami (talk) 06:16, 22 January 2023 (UTC)

My very first proposal & what I am consistently pitching is that there ought to be a comparative table, showing different scholars' notations in order to avoid confusion & because there is no general system. That's the table. I am proposing that a comparative table is the reasonable replacement for this, & the best alternative to deletion of this article. I said in my first comment that I'd be willing to contribute to such a table.

All of this work is notoriously bad and roughly equally notable. Most linguists think that Dolgopolsky's Nostratic is absolute bunkum. Ehret's h-index is 12. Bomhard's is 10. Diakonoff's is 6. Google Scholar's not all that reliable for citations counts, but it's the best thing we've got. Citations for Ehret's reconstruction way outstrips the others by far. I don't think it's good, but I don't think any of this stuff is good, but my opinion doesn't matter: It's clearly comparably notable. Pathawi (talk) 06:34, 22 January 2023 (UTC)

And I said I'd be fine with that, just that I thought the proposed structure was backwards. We shouldn't list different scholars' transcriptions for particular reconstructions because different scholars don't have the same reconstructions. — kwami (talk) 06:50, 22 January 2023 (UTC)