User talk:Mahagaja/Archive 13

Latest comment: 7 years ago by Angr in topic nighean
This is an archive page that has been kept for historical purposes. The conversations on this page are no longer live.

di- vs dí- prefix

We currently have Category:Old Irish words prefixed with di- and the entry di-, but in some instances the i is actually long. What is the real underlying vowel here, and where does the change in length come from? Should we use the long-vowel form as the lemma? —CodeCat 20:21, 1 October 2016 (UTC)Reply

Matasović reconstructs the preposition as *dī, connecting it with Latin (though if that's right, then our etymology of isn't), and says it was shortened in unstressed position, and Thurneysen says it was also shortened before a vowel, and both shortened and lowered to de- before ro-. So I guess the main entry for the prefix should be dí-. —Aɴɢʀ (talk) 20:41, 1 October 2016 (UTC)Reply
DIL seems to be a bit inconsistent with the length. For do·guid it has at the start of the entry, but for do·gní it has di instead. Is there any significance to this? —CodeCat 20:48, 1 October 2016 (UTC)Reply
I don't think so. —Aɴɢʀ (talk) 14:19, 2 October 2016 (UTC)Reply

ē?

What would Proto-Celtic ē become in Proto-Brythonic? UtherPendrogn (talk) 19:35, 5 October 2016 (UTC)Reply

It became . —Aɴɢʀ (talk) 19:37, 5 October 2016 (UTC)Reply

Old Irish noun inflection overhaul

I've redone the template for Old Irish noun inflections, and made it more like the verb template. Now the parameters are named instead of numbered, so that it's no longer necessary to add comments; it's self-explanatory. Forms can be given a special tag to indicate initial mutations that accompany it. There's also a parameter for the inflectional class, which categorises. See {{sga-decl-noun}}. —CodeCat 22:43, 11 October 2016 (UTC)Reply

OK, thanks. Is the idea that it should be used instead of more specific templates like {{sga-decl-o-masc}}? Or do the more specific templates call on the general one? —Aɴɢʀ (talk) 09:19, 12 October 2016 (UTC)Reply
No, they can both be used. This template is for when you'd rather specify each individual attested form, like for the verbs. It's useful for me because I can add declension templates that indicate the declension class and categorise accordingly, without having to interpret any of the attestations on DIL. I can just not specify any forms, and leave that to someone more experienced. —CodeCat 15:54, 12 October 2016 (UTC)Reply

*kumbā

Hey @Angr, I could use your help in the etymology of *kumbā. It seems there's a lot of confusion around it and its alleged cognates. Do you think *kumbaz could be an early borrowing or merger from PCelt? Also, could campus be a cognate? Thanks. --Victar (talk) 16:43, 12 October 2016 (UTC)Reply

*kumbaz would have to be a fairly late borrowing since it doesn't undergo Grimm's Law. If anything, the Celtic word is related to the PGmc ancestor of hump. I don't see how to connect it to campus since there's no way to get u-vocalism out of *kh₂em-. —Aɴɢʀ (talk) 16:59, 12 October 2016 (UTC)Reply
Ah, good call on hump. It could be that PGmc *kumbaz is unrelated (or at least not closely) and that OE cumb merged with a Brythonic borrowing to adopt the meaning of valley. --Victar (talk) 17:31, 12 October 2016 (UTC)Reply
@Victar: the jump from *ḱewh₂- to *kumbʰo- is pretty strange (the depalatalization of *ḱ, disappearance of *h₂, and extension with *bʰ). Is there more explanation available? —JohnC5 18:39, 12 October 2016 (UTC)Reply
You're right, it's very dubious. I took that from the reconstruction of *kuwos in Matasovic, and elsewhere I've seen it as *ḱówHwos, but it could be that *kumbā and *kuwos derive from different roots, despite all the alleged shared cognates. --Victar (talk) 19:06, 12 October 2016 (UTC)Reply
Oh, so in the entry for *kuwos, Matasovic mentions that *kumbā is also a cognate? —JohnC5 19:23, 12 October 2016 (UTC)Reply
No, he only claims कुम्भ (kumbha, a pot, jug) is a cognate which itself is cited as a cognate for some of the various other alleged cognates. He also goes on to say it might not be PIE in origin, so he wasn't very sure either. --Victar (talk) 19:32, 12 October 2016 (UTC)Reply
(e/c) No, he doesn't. He connects *kumbā with कुम्भ (kumbha, pot) and says Beekes further compares it with κύμβη (beaker) (although since Sanskrit has bh and Greek has b, it seems to me that it can only be cognate with one or the other, but not both, unless the Greek is a loanword from some branch where became b), and he connects *kuwos with शून्य (śūnya, empty, hollow), κύαρ (eye of the needle), cavus (hollow), and Tocharian B kor (throat). —Aɴɢʀ (talk) 19:35, 12 October 2016 (UTC)Reply
Right, exactly. I think the shared origin of PGmc *humpaz (hill, heap), from PIE *kumbʰ- (curved), and OE cumb being influenced by Brythonic makes the most sense. If so, what's the root of *kumbʰ-? --Victar (talk) 19:45, 12 October 2016 (UTC)Reply
But *humpaz has the same problem κύμβη does, namely it shows b instead of . And there is also a कुम्ब (kumba, thick end of a bone or club), which semantically could work with the Germanic, but not so well with the others. —Aɴɢʀ (talk) 19:54, 12 October 2016 (UTC)Reply
I'd missed this, but Beekes points out that the PIE form *kumbʰo- is impossible for a native PIE word since *k and *bʰ may not coöccur in a root. So Beekes prefers a Wanderwort. —JohnC5 02:57, 13 October 2016 (UTC)Reply
A reconstruction of the Proto-Indo-Iranian form I found is *kʰumbʰa-. We also have Albanian sumbull. --Victar (talk) 04:30, 13 October 2016 (UTC)Reply
Ok, so we could account for Celtic, PII, and Greek *kHumbHéh₂, but man, that would be a weird formation and definitely not native.
Orel give *ḱumbéh₂ for Albanian, Celtic, and Greek, but again this would be very weird given the *b. —JohnC5 05:00, 13 October 2016 (UTC)Reply
Here's de Vaan's take on it (page 11). I don't agree with the Latin cognate though. --Victar (talk) 05:07, 13 October 2016 (UTC)Reply
Given all this differing information (*ḱ/k/kʰumb/bʰéh₂), I am very uncomfortable reconstructing a PIE form at all. I would prefer to ascribe this to a Wanderwort. CodeCat, what do you think? —JohnC5 14:08, 13 October 2016 (UTC)Reply
I can't really say anything beyond what has already been said. —CodeCat 14:11, 13 October 2016 (UTC)Reply
I think most certainly there was an original PIE word that was a borrowing from a non-PIE language, as several published works can testify. The verdict is still out on whether the PCelt word is related. --Victar (talk) 16:48, 13 October 2016 (UTC)Reply

──────────────────────────────────────────────────────────────────────────────────────────────────── Tantalizing as all these options are, we mustn't forget that the only attested meaning within Celtic is "valley". Of course while it's possible for a word meaning "valley" to be cognate with words referring to pots and bowls and beakers, it isn't particularly obvious that it must be. The second meaning "barrel" seems to be assigned only for the alleged Middle Irish word comm, which Matasović however says does not exist, and which indeed I can find no evidence for in DIL. —Aɴɢʀ (talk) 14:27, 13 October 2016 (UTC)Reply

Matasović says the Old Irish form doesn't exist, not the Middle Irish. It could be simply we've yet to find an older example, or the Middle Irish form is a re-borrowering from Brythonic, retaining an older lost meaning. --Victar (talk) 15:59, 13 October 2016 (UTC)Reply
You're right, he says Old Irish. But if it were attested in Middle Irish, he would say so, and it would be in DIL, which gives Irish words from Old Irish through Middle Irish and up to Early Modern Irish. But the word isn't in the dictionary at all. Not only is there no entry for it, it doesn't occur in quotes in the entries for other words. I think the word isn't attested at all. —Aɴɢʀ (talk) 16:04, 13 October 2016 (UTC)Reply
Weird. I wonder where it came from than. Did someone just make it up? I know de Vaan cites it. --Victar (talk) 16:09, 13 October 2016 (UTC)Reply
I think I found it: there's a word, certainly Middle Irish and possibly Old Irish, coim, which sometimes means "vessel, container", but mostly seems to mean "waist, middle" or "protection, covering", so its semantics make a relationship to the "valley" word questionable at best, and its palatalized m means it can't come from *kumbā anyway. There is also a Modern Irish word com that means "coombe, cirque, mountain recess", but that's probably a loanword from Welsh, either directly or via English cwm, since it doesn't appear in DIL and so is probably a modern word. —Aɴɢʀ (talk) 16:22, 13 October 2016 (UTC)Reply
Hah, I was just about to post this:
Air. coim, coimm (f.), später auch com, comm (m-o) ‘body, breast, bosom, waist’; ‘protection, shelter’; auch ‘vessel, container’ (DIL, S. 129; vgl. LEIA, C-146). – Z.B. ITS V 158.16: a coim na cumhaile ‘(le bébé tombant) du sein de la servante’ (LEIA ebd.).
Here's the etyomology they give:
Air. coim(m) < proto-goidel. *kumbi-, air. comm < proto-goidel. *kumb~ = galloroman. *cumba ‘Tal, Trog’ ( > frz. combe ‘Talschlucht’)7. Vgl. akymr. cum, mkymr. cwm(m), bret. komb, alle der Bedeutung ‘Tal’, < proto-brit. *kumbos (VG I, S. 119; IEW, S. 592; LEIA ebd.).
--Victar (talk) 16:34, 13 October 2016 (UTC)Reply
I'm wondering if coimm, comm is the origin of University of Wales reconstruction *kombo- (cover). --Victar (talk) 16:41, 13 October 2016 (UTC)Reply
Very likely. I've always been suspicious of that wordlist precisely because it doesn't provide the attested reflexes of its reconstructions. —Aɴɢʀ (talk) 17:41, 13 October 2016 (UTC)Reply
Strange that Matasovic is so seeming adamant that Old Irish comm never existed. He could have cited coimm, comm and written his opinion on it. --Victar (talk) 18:28, 13 October 2016 (UTC)Reply
Indeed. He doesn't mention coim at all. —Aɴɢʀ (talk) 18:41, 13 October 2016 (UTC)Reply

Pronunciation of mathgamain

According to Stifter's Sengoídelc, all vowel letters other than u represented schwa when they stood in unstressed non-final position. Thus, I'd expect this to be /ˈmaθɣəṽənʲ/ instead. What is your take on this? —CodeCat 20:57, 17 October 2016 (UTC)Reply

It's true that unstressed nonfinal vowels don't contrast with each other; their quality is entirely predictable from the quality of the consonants surrounding them. But since we have to assume 5 separate short vowel phonemes in both stressed syllables and unstressed final open syllables anyway, it seems uneconomical to posit a sixth phoneme /ə/ that's found only in unstressed syllables when followed by a consonant. I'd rather keep /a e i o u/ as the short vowel phonemes and posit some sort of restriction on where they may occur. —Aɴɢʀ (talk) 21:06, 17 October 2016 (UTC)Reply
Stifter has no problem doing just that. And this is exactly what we do for Central Catalan, also. Unstressed /e/ and /a/ collapse into [ə], and we denote this with the phoneme /ə/. In fact, you could say that we already do it for English, too. Schwa in English represents an unstressed variant of certain vowels, the exact phoneme of which is unrecoverable, just as in Old Irish and Catalan. —CodeCat 21:09, 17 October 2016 (UTC)Reply
For Catalan and English, we know that the vowel in question actually surfaces as [ə] in pronunciation. We don't know that for Old Irish. We don't know whether /ˈmaθɣVṽVnʲ/ invalid IPA characters (VV) was pronounced [ˈmaθɣaṽɨnʲ] or [ˈmaθɣəṽənʲ]. If the former, then I am disinclined to give the phonemic form as /ˈmaθɣəṽənʲ/. And I think the former is a bit more likely, because the scribes are actually pretty consistent (by Old Irish standards, at any rate) in how these unstressed vowels are spelled: a between two broad consonants, e after a slender and before a broad, (a)i after a broad and before a slender, and i between two slenders. If they were really all pronounced the same, you'd expect more variation than is attested. —Aɴɢʀ (talk) 21:32, 17 October 2016 (UTC)Reply
Even so, I think you'd agree that there is complimentary distribution here, which is a hallmark of allophones of one phoneme. From what I gather, you are inclined to write full vowels because that's what the scribes wrote, and also because the phones happen to agree with similar phones in stressed position. However, it's quite possible for a language to have stressed /a/ and /e/ as full phonemes, but for unstressed [a] and [e] to be allophones of each other. To assign them to the equivalent stressed phonemes would sidestep this allophony, even though it's quite relevant to the phonemic system, which deals primarily with contrasts. —CodeCat 21:39, 17 October 2016 (UTC)Reply
Also note that Wikipedia's article Old Irish makes no mention of the actual pronunciation of the "schwa", but it treats it as a phoneme quite explicitly. —CodeCat 21:48, 17 October 2016 (UTC)Reply
But it isn't just stressed position, it's also unstressed position in absolute final position (where no consonant follows). There, all 5 vowels can follow any consonant. /ˈbaθa, ˈbaθe, ˈbaθi, ˈbaθo, ˈbaθu, ˈbaθʲa, ˈbaθʲe, ˈbaθʲi, ˈbaθʲo, ˈbaθʲu/ are all possible distinct words of Old Irish. —Aɴɢʀ (talk) 21:49, 17 October 2016 (UTC)Reply
Yes, I'm aware. It doesn't really change much though. —CodeCat 21:50, 17 October 2016 (UTC)Reply
It changes a lot for me. I'm prepared to believe in a merged phoneme in unstressed syllables, but one in unstressed syllables only when followed by a consonant exceeds my ability to suspend my disbelief. When I wrote the Old Irish phonology section at Wikipedia back in 2005, I didn't posit a phoneme /ə/; Benwing did that several years later, probably on the basis of Stifter. It is a defensible theory, but IMO the less economical one. Incidentally, the contrast between broad and slender consonants is lost in word-initial position in Old Irish too—it's always predictable from the following vowel—but because palatalization is contrastive in other positions, we treat the two sets as separate phonemes even in the position where they don't contrast. —Aɴɢʀ (talk) 22:06, 17 October 2016 (UTC)Reply

Adding a language

@Angr Since I've been reading a book about the language for a while, I thought I'd add a few Old Japanese lemmas. The coding somewhat eludes me, my tests so far (in the preview, I have not posted any lemmas yet, just Templates) won't let me add a romanisation of the word. The language is written in phonetic Chinese Hànzì. I've studied several other templates but can't find where and when it does all that. It seems to be basically automatic for the Japanese entries, where it immediately gives you several forms when you enter one. No need to try and fix it yourself, I'd be interested in figuring out the code myself, and there are a few technicalities I'll have to implement (some hànzì were used for two different sounds, which I assume will mess up the transliteration). UtherPendrogn (talk) 16:36, 25 October 2016 (UTC)Reply

I don't really know a lot about our coverage of Japanese, but I am quite sure that romanization is not automatic for it. It always has to be provided manually. For anything else, you're probably better off asking in the Grease pit. —Aɴɢʀ (talk) 16:39, 25 October 2016 (UTC)Reply
Cheers, I'll add it in manually then. I'll also ask in the Grease Pit. The etymology will always be in Latin anyway, since Proto-Japonic uses the Latin script. UtherPendrogn (talk) 16:58, 25 October 2016 (UTC)Reply

yRV'd??

Wiktionary:Main Page

itsUS,NOTamerican!81.11.206.39 08:57, 29 October 2016 (UTC)AMERICA=CONTINENT,SMARTppl.gethat!!81.11.206.39 17:47, 29 October 2016 (UTC)Reply

We're speaking English here, not Spanish. See sense 1 of American. —Aɴɢʀ (talk) 17:52, 29 October 2016 (UTC)Reply

Category:Pages using ISBN magic links

Why did the magic word ISBN begin to put all the pages in Category:Pages using ISBN magic links? Is there any technical reason? — TAKASUGI Shinji (talk) 07:56, 11 November 2016 (UTC)Reply

I have no idea. All I did was create the category so it would be blue (and could be hidden) rather than red. Whatever caused it to happen was already causing it to happen before I created the category. Maybe ask at the Beer Parlour. —Aɴɢʀ (talk) 07:59, 11 November 2016 (UTC)Reply

nighean

I'm happy to go along with the morphological process, but in the records that I've been looking at, you just don't see "nighean" before the 16C. At all, ever. I mean, if there's an example of 12C "nigęn" out there, I'll change my mind, but I really don't think you can say anything other than that "nighean" is derived from "inghean" in the 16C or so. (And I haven't seen it spelled as such before 1700 in Irish/Scottish language records, although IIRC there are hints of the pronunciation being something like it in Anglicised records.) --Catsidhe (verba, facta) 20:15, 16 November 2016 (UTC)Reply

I don't deny that for a moment. All I mean is that I really don't think it was /inʲɣʲen/ > /nʲiɣʲen/ by metathesis, but rather /inʲɣʲen/ > /inʲiɣʲen/ by epenthesis (whence also Irish iníon, pronounced /iˈnʲiːn/ with forward stress in many parts of Ireland even when it was still spelled inghean), and then /nʲiɣʲen/ by apheresis. And not just in Scotland (where the spelling nighean is found in the Book of the Dean of Lismore); Ulster and northern Connacht Irish as well as Manx also have n-initial versions. —Aɴɢʀ (talk) 20:32, 16 November 2016 (UTC)Reply
Oh yes, no doubt that "nighean" as spelled was a Scottish/Ulster form, but was also the base of Ní and Nic in women's names (as pronounced, that is). But still, I maintain that while, yes, it is also derived from OIr ingen, it did not spring therefrom fully formed, but via EMod Ir inghean. --Catsidhe (verba, facta) 20:41, 16 November 2016 (UTC)Reply
Or at least from a stage that could equally well be called Early Modern Scottish Gaelic, or what Ethnologue is pleased to call Hiberno-Scottish Gaelic. Writing {{der|gd|ga|inghean}}, however, gives the false impression that it's a loanword from Irish into Scottish Gaelic. We could say it comes from Middle Irish ingen, too. —Aɴɢʀ (talk) 21:11, 16 November 2016 (UTC)Reply
That's getting into a whole other ontological nest of vipers. I would have said that until the early 17C, that Scottish Gaelic and Irish were the same language, notwithstanding the differences evident since the Book of Deer. Even the Wiki page says that Classical Gaelic was common until the 17C, when the split into regional languages accelerated. If I had the chance to do it over from scratch, I would have had Primitive Irish > Old Irish (600-900) > Middle Irish (900-1200, as a transition language where OIr forms and orthography were fading and new orthography and grammar appearing) > Classical Gaelic (1200-1600) > Early Modern Irish (1600-18mumble) / Early Modern Scottish Gaelic (1600-18mumble) > Modern Irish/Modern Scottish Gaelic. That way we could unambiguously say gd nighean < EMgd nighean < CGae inghean without imposing an Irish primacy which the Scots might feel uncomfortable with. But as it is, the language as spoken in Scotland and Ireland at around 1600 is called, by us now, Early Modern Irish. --Catsidhe (verba, facta) 21:29, 16 November 2016 (UTC)Reply
Well, not really, because we don't have a separate code for Early Modern Irish, nor for Classical Gaelic (we don't use the code ghc). Instead, we call everything after 1200 or so "Irish" and "Scottish Gaelic", and can label it "Early Modern" or "early" in etymologies as we see fit. We could thus say modern SG nighean comes from "an early modern form inghean" (assuming that spelling is attested in some unambiguously Scottish text after 1200), which in turn comes from Middle Irish ingen, from Old Irish ingen. —Aɴɢʀ (talk) 21:37, 16 November 2016 (UTC)Reply
Return to the user page of "Mahagaja/Archive 13".