User talk:Mahagaja/Archive 21

You earlier removed HTML tags generated by Module:mh-pronunc to display small parentheses in IPA. I know these aren't specifically necessary and that you were concerned that they add articles to Category:IPA pronunciations with paired HTML tags. The problem is, given an on-going discussion in wikipedia:Talk:Marshallese language, we've been working not only to negotiate and reach consensus on IPA for Marshallese, but also to craft something that is not difficult on the eyes. The parentheses in question were selected to surround epenthetic vowels [(o)]—non-phonemic but phonetically-audible vowels that appear in unstable consonant clusters, but disappear when speakers enunciate words morpheme-by-morpheme. The problem with parentheses, though, is that they are visually disruptive in a way that unnecessarily slows reading at a quick glance, especially for vowels that by definition have less prosodic significance than normal vowels. In order to lessen the disruption, I decreased their size, experimenting with [(o)] or [(o)] instead. This is a language whose previous handling at Wikipedia and Wiktionary was fraught with serious problems in legibility of IPA transcriptions, confusing readers, so making the transcriptions both adequately accurate and smooth to read is important.

At the time, when you removed the HTML tags, I decided to adapt and see how well I could polish the appearance of the IPA with normal-sized parentheses. But the truth is, they're still visually-disruptive. I've searched to see if there are smaller encoded Latin script parentheses in Unicode, but I can't find any. And other proposed notation conventions (including breve [ŏ], inverted breve [o̯], superscript [ᵒ] and unmarked [o]) have failed to gain consensus, but parentheses have. I want to fully respect consensus, since it's important that the symbols chosen can better withstand scrutiny upon further review, but it is still important that the symbols involved display elegantly. - Gilgamesh~enwiki (talk) 19:23, 16 January 2020 (UTC)Reply

I guess it's all a matter of personal opinion. I don't find the full sized parentheses visually disruptive at all. Unicode does provide small parenthesis characters at U+FE59 and U+FE5A; they look like this: [﹙o﹚], but I find them more disruptive than normal parens. —Mahāgaja · talk 21:03, 16 January 2020 (UTC)Reply

Those could work for Marshallese. I'll experiment with them. Thank you for pointing them out to me. - Gilgamesh~enwiki (talk) 21:32, 16 January 2020 (UTC)Reply

I wouldn't recommend them: they're wide characters, like Han characters, and won't match the characters around them. — Eru·tuon 21:54, 16 January 2020 (UTC)Reply

Awwwww... So what's so problematic about having Marshallese IPA appear in Category:IPA pronunciations with paired HTML tags, then? - Gilgamesh~enwiki (talk) 23:42, 16 January 2020 (UTC)Reply

Nothing, I guess. There are already a bunch of entries in there that I can't clean up anyway. The category is mostly there to make sure people use characters like ʰ, ʲ, and ʷ correctly instead of formatting them as superscripts. —Mahāgaja · talk 07:33, 17 January 2020 (UTC)Reply

That's reasonable, and not a problem with those characters where the Marshallese IPA is concerned. I was just looking for less visually imposing parentheses for epenthetic vowels. Hmm... I may have to return to the on-going discussion and reraise this topic. - Gilgamesh~enwiki (talk) 11:13, 17 January 2020 (UTC)Reply

Since I pointed out gelach on RFV earlier, I'd also like to hand over another blatantly modern term I also found while cleaning up declensions: loigic. mellohi! (僕の乖離) 19:25, 19 January 2020 (UTC)Reply

@Mellohi!: Yeah, there a lot of these. A certain Wiktionarian used to assume that every entry in DIL was Old Irish. —Mahāgaja · talk 19:38, 19 January 2020 (UTC)Reply

I had long had this impression, too, having formerly followed that very same fallacious assumption in my earlier Old Irish edits until recently. Anyhow, what's the course of action? I propose the following: If Middle Irish, convert the entry from Old to Middle, if blatantly Modern, either delete the entry on sight or convert to an archaic spelling soft redirect under "Irish" depending on whether that spelling of the root is actually attested. Though you do have a point regarding gelach and existence of the Scottish Gaelic and/or Manx cognates of the modern term. mellohi! (僕の乖離) 19:51, 19 January 2020 (UTC)Reply

Your suggested course of action is pretty much what I do; I'll also simply move the entry (without leaving a redirect) to the Modern Irish form if we don't already have an entry for the latter. I just did that for Irish loighceoir, formerly Old Irish loigceóir. —Mahāgaja · talk 19:55, 19 January 2020 (UTC)Reply

Another thing to do is make a list of which of DIL's bibliography is which stage of the language. For SGA, we know Wb, Ardm, Sg, Tur, Bcr, and Thes. Fél, LL and LU "can usually safely be labeled "Old Irish". Don't know where to classify "Laws" though. mellohi! (僕の乖離) 20:20, 19 January 2020 (UTC)Reply

Yeah, I know; I'm the one who wrote that at WT:ASGA. I'm not sure about "Laws" either; but in general mga is safer than sga, because if a word exists in Old Irish it probably also exists in Middle Irish, but you can't assume the other way around. —Mahāgaja · talk 20:25, 19 January 2020 (UTC)Reply

Note that DIL's "Laws" citation label is always followed by a Roman numeral. DIL's bibliography notes that "Laws i", "Laws ii" and "Laws iii" are from the Senchas Már, which according to enwiki "was compiled into a single group sometime in the 8th century, though individual tracts vary in date" seemingly putting them safely in the Old Irish range. The other book also included in "Laws iii", the Book of Aicill, which apparently can be dated to the 7th century, also in Old Irish range. I'll be wary of "Laws iv" and "Laws v". mellohi! (僕の乖離) 20:47, 19 January 2020 (UTC)Reply

Sounds good! Feel free to add that info to WT:ASGA. —Mahāgaja · talk 20:49, 19 January 2020 (UTC)Reply

@Mellohi!: Another test I use is that if a word is mentioned in Thurneysen I consider it Old Irish, regardless of DIL's list of attestations. Maybe that's lazy, but I figure Thurneysen knew what he was doing. —Mahāgaja · talk 09:49, 20 January 2020 (UTC)Reply

Handing in another term to exterminate, cnáip. mellohi! (僕の乖離) 20:37, 24 January 2020 (UTC)Reply

This is an n-stem neuter in -mm, but its genitive singular nadma is rather bizarre and I don't recall another n-stem ever having a genitive singular ending in a. Also I can't declare it half-i, half-n since its dative singular is nadmaim (or whatever spelling). My main question is, is it a safe assumption that the genitive singular was ripped from the i-stem declension and thus assume that it aspirated vowel-initial words following it? mellohi! (僕の乖離) 17:05, 22 January 2020 (UTC)Reply

I don't know, and Thurneysen doesn't mention the genitive singular. My first thought is that it simply shows depalatalization of the -dm- cluster, but even then we would expect *nadmae in the oldest language. It may simply be a late form of that. —Mahāgaja · talk 17:16, 22 January 2020 (UTC)Reply

nadma appears repeatedly in the Senchas Már in addition to the Críth Gablach, so the decay of -ae must have occurred surprisingly early. mellohi! (僕の乖離) 17:37, 22 January 2020 (UTC)Reply

Thurneysen p. 210 does say that the genitive singular of neuter n-stems can be -a as well as -ae, and on p. 62 he says that -a for -ae (in general, not specifically this form) occurs as early as Ml. —Mahāgaja · talk 18:04, 22 January 2020 (UTC)Reply

Was doíni formerly the plural of this term? doíni is considered an i-stem but palatalization in the final consonant of doén is very poorly attested. mellohi! (僕の乖離) 17:02, 24 January 2020 (UTC)Reply

dóen is attested so much later than doíni that I think it's actually a backformation from it. —Mahāgaja · talk 17:08, 24 January 2020 (UTC)Reply

The depalatalization goes unexplained any way we slice it. mellohi! (僕の乖離) 17:40, 24 January 2020 (UTC)Reply

DIL notes that the primary semantics of the word are nigh-identical to those of labor; and thus it's a possibility the two nouns are in fact one. Is this a solid assumption? mellohi! (僕の乖離) 17:44, 24 January 2020 (UTC)Reply

If the inflection of the two senses is the same, we can merge them and say in the etymology section that the sense "illness" may be a {{semantic loan}} from Latin. The sense "grave" doesn't seem to exist in Latin, though. —Mahāgaja · talk 17:49, 24 January 2020 (UTC)Reply

For some reason, the EDPC reports cacht (and flagged as Middle Irish) as a masculine o-stem and a feminine a-stem. Looking at DIL's attestations, feminine a-stem is fair ("admider de .uii. cachta") but classifying it as an o-stem makes no sense since masculine o-stems have no business with having any form terminate in -a. Additionally, no palatal stem forms are attested (you'd expect those to appear if they were masculine o-stems or feminine a-stems...). Its genitive singular is cachta (e.g. comus cachta "power of impounding" listed in DIL), which smells like a u-stem. mellohi! (僕の乖離) 19:55, 24 January 2020 (UTC)Reply

Well, cht is never palatalized in Old Irish, so we expect the masculine o-stem to have the genitive singular cacht and the feminine ā-stem to have the genitive singular cachtae (or cachta per -ae > -a mentioned in a previous thread). In theory, genitive singular cachta could thus indicate an ā-, i-, or u-stem; but we really do expect the masculine meaning "slave" to be an o-stem and the feminine meaning "slave" to be an ā-stem, because it has the past participle ending -tos/-tā regardless of whether it's inherited from PIE or borrowed from Latin. But the abstract noun "captivity, duress" could also be a feminine i-stem *kaxtis < *kap-tis. And even the masculine o-stem could have been attracted to a different class (e.g. u-stem) because of the undesirability of having both the genitive singular and the nominative plural identical to the nominative singular. So in theory, we expect the following paradigms:

• "slave", masculine o-stem: cacht throughout the singular, as well as nom./gen.pl.; cachtu in the acc./voc.pl.; cachtaib in the dat.pl.
• "slave", masculine, attracted to the u-stem declension: cacht in nom./voc./acc./dat. sg.; cachto/cachta in gen.sg.; cachtae/cachta in nom./gen.pl.; cachtu and cachtaib as above
• "slave", feminine: cacht in nom./voc./acc./dat. sg. and gen.pl.; cachtae/cachta in gen.sg.; cachta in nom./voc./acc.pl.; cachtaib in dat.pl.
• "captivity", feminine i-stem: singular just like the masculine u-stem; probably essentially unused in the plural.

It's easy to see why these words would become conflated, especially in the later language (and it isn't attested in early Old Irish). —Mahāgaja · talk 20:38, 24 January 2020 (UTC)Reply

Thanks for clarification. mellohi! (僕の乖離) 20:55, 24 January 2020 (UTC)Reply

This verb has bizarre prefix behaviour. It has the consonantal behaviour of to- but the vowels do not check out. This is evident from the verbal noun and prototonic formations; they have te-. If it had the vowel behaviour of to-, they should have started with tu- or to-. I'm puzzled on where the /e/ came from. mellohi! (僕の乖離) 17:22, 31 January 2020 (UTC)Reply

@Mellohi!: The same thing happens when léicid is prefixed with ro-: 3rd sg. pres. subj. prototonic ·relic and passive sg. perfect prototonic ·reilced. Pedersen (VKG 2:562) tentatively suggests a phonological rule changing o to e before the ē in the following syllable that got syncopated away, i.e. *to-lēggīt > *te-lēggīt > ·teilci. Thurneysen (pp. 529, 532) says that ·relic changed its vowel under the influence of unrelated reilic (“graveyard”) (modern Irish reilig, from Latin reliquiae) and then ·teilci assimilated its vowel to that of ·relic, but that seems unlikely to me. Why would a verb form meaning "may throw" be influenced by a noun meaning "graveyard"? —Mahāgaja · talk 18:52, 31 January 2020 (UTC)Reply

Pedersen's rule doesn't account for do·tét and légaid which exhibit no similar infection behaviour whatsoever. Especially the latter, since it involves both a syncopated /eː/ and ro-prefixing. mellohi! (僕の乖離) 05:03, 1 February 2020 (UTC)Reply

No, it doesn't. I suppose that's why he only tentatively suggested it. Maybe someone somewhere has come up with a better explanation in the meantime, but I don't know what it is. —Mahāgaja · talk 07:38, 1 February 2020 (UTC)Reply

This verb has a suppletive past form, raink. The etymology for this suppleted form is rather straightforward as it's blatantly derived from the reduplicated preterite of ro·icc. However, I cannot track down the non-suppleted forms in rosh. Deriving them from Old Irish ro·saig and Middle Irish roichid is impossible since the outcome of slender /x/ in Manx should be zero, not sh (cf. cluiche and deich). ro·icc itself doesn't seem to check out either. Can't find any semantically similar Old Irish verbs that phonetically fit. mellohi! (僕の乖離) 04:44, 1 February 2020 (UTC)Reply

I have no idea. Maybe it's a loanword from Middle English, though what would explain the vowel? And I don't know how Manx treats /t͡ʃ/ in loanwords; in Irish it becomes /ʃtʲ/, but maybe it becomes plain /ʃ/ in Manx? —Mahāgaja · talk 07:38, 1 February 2020 (UTC)Reply

Having randomly noticed this in my watchlist, I poked around but all I could find was an article by Jennifer Kewley Draskau seeming to say some speakers don't [know how to?] inflect the verb suppletively anymore and inflect it regularly(?), which doesn't help with the etymology. (She elsewhere quotes an example of "written to be spoken" Manx using raink which she footnotes as "a preterite form possibly analogous to haink, meaning 'arrived'".) Building on Mahagaja's idea, perhaps Middle English "reach" influenced, or was borrowed and then influenced by, "ro·icc", thus explaining the vowel? But that's pure speculation. - -sche (discuss) 18:09, 1 February 2020 (UTC)Reply

@-sche: Kelly's 18th-century Practical Grammar seems to note the suppletion of raink in the paradigm of rosh (it precedes the note by "quere.", a term I do not understand however). If so, it doesn't seem to be a post-revival analogy. I agree that I'm inclined to posit that Middle English rēchen (“to reach”) was involved somewhere along the line. But I prefer to derive the analogical /o/ from ro·saig instead of ro·icc, since the former had /o/ all over its prototonic forms (and prototonic forms were pretty much always favored in the evolution of the modern Goidelic languages) and the latter never had /o/ prototonically. mellohi! (僕の乖離) 05:26, 2 February 2020 (UTC)Reply

I have trouble comprehending the etymology of con·túaisi and in·túaisi and ar·túaisi. (And the simplex túaisid.) Currently they are listed as deriving from *tawsos due to the simplex and the compounds also meaning "be silent", but this poses phonological issues due to the final-syllable -s-, which should have been destroyed on the way to Old Irish unless it was reduced from elsewhere. Macbain (with reference with the *tawsos etymology) and others derive in·túaisi from aith- + to- + the root of ar·sissedar, but the way they handled the prefixes makes no phonological sense whatsoever, probably due to them not recognizing étsi- as the prototonic stem of in·túaisi in the first place:

"éisd, listen, hear, Ir. éisdim, O. Ir. étsim. Ascoli analyses it into *étiss, *aith-do-iss, animum instare; the iss he doubtless means as from the reduplicated form of the root sta (cf. O. Ir. air-issim, I stand). an-tus-, great silence! Cf. Ir. éist do bhéal = hush! Root of tosd."

The verbal class (A II) of this set of verbs also makes Macbain's etymology untenable, since ·sissedar seems to come from a different verbal class (cf. at·tá and its relatives).

Matasović has never assigned either of the two verbs an etymology either. mellohi! (僕の乖離) 15:47, 3 February 2020 (UTC)Reply

I'll see if I can find anything about this when I get home this evening, but I suspect I won't be able to provide any more info than what you've already discovered. —Mahāgaja · talk 18:00, 3 February 2020 (UTC)Reply

@Mellohi!: As I feared, I can't track down a definitive answer either. Almost everyone agrees it's related to tóe (“silent, silence”), and it's clear it has to go back to something with a geminate -ss-, but I can't find any information on where that comes from. A *tawso-steyeti ("stands silent"?) seems unlikely, and *taws-teyeti and *taws-seyeti involve suffixes whose existence I'm unaware of. A *tawso-sagyetor would at least have a well-known denominative suffix, but it would give *·túaigethar, not ·túaissi. Even if the suffix were added athematically (or if there was a super-early syncope of the o), and even if it were exceptionally active instead of deponent, *taws-sagyeti would still give *·túaissig or something, and there's just no plausible way to get from *taws-sagyeti to *taws-seyeti/*taws-sīti. —Mahāgaja · talk 21:36, 3 February 2020 (UTC)Reply

Came across this one when trying to resolve our previous issues over con·túaisi, namely by me attempting to derive that verb as a verbing of some -tos or -tus nominal (which doesn't work due to vowel length mismatch anyway).

There are several bizarre phonological issues with this noun alone.

• This is one of the few, if not only, seemingly native Old Irish nouns ending in -st. Normal sound change should have turned Proto-Celtic -st- into /s/.
• DIL classifies this as a u-stem, yet its entire paradigm is consistently lowered (either to /o/ or /a/), not just the genitive stem. tus(t) is not found. Macbain not unreasonably treats this as an o-stem instead.
• EDPC does not mention this term at all.

mellohi! (僕の乖離) 18:07, 5 February 2020 (UTC)Reply

DIL doesn't cite a single genitive anyway. If it were from *tostos, we'd expect the genitive to be tuist; I suppose DIL calls it a u-stem because the genitive is tosta in modern Irish. Then there's the fluctuation between initial t- and s-. Neither Thurnyesen nor Pedersen mentions this word. I'm at a loss. —Mahāgaja · talk 19:24, 5 February 2020 (UTC)Reply

The bizarre initial consonant fluctuation seems to be due to conflation with its synonym socht, which does turn up in the EDPC. DIL also lists other such conflation-born forms like tocht and sosd. I don't know if they are of any help in explaining the final t of tost though since they both are dated long after Corm Y, LU and LL (where tost also turns up). mellohi! (僕の乖離) 20:01, 5 February 2020 (UTC)Reply

I wonder if tost itself could be a conflation. That could be the explanation for the -st. What if Proto-Indo-European *teh₂ws- formed a nominal like *th₂us-tus; that would give Proto-Celtic *tussus > Old Irish *tuss. If that got conflated with socht, the result could be tost. —Mahāgaja · talk 20:10, 5 February 2020 (UTC)Reply

The conflation solution could also deal with the paradigmatic root vowel being already universally lowered by the time of Corm Y and the lack of attestation of raised vowel forms. mellohi! (僕の乖離) 20:47, 5 February 2020 (UTC)Reply

Thank you very much for correcting βεβαιότερος for the modern section too. A thought: since my first days in wikt. i noticed that the table of ancient ajdectives is very awkward at the 'Derived' row. The degrees of the adj. and the adverb+its degrees are mixed up. I have shown to... maybe it was you?, or John?, the way i imagined it. I understood that any change would be difficult, also a separate degrees.table. A thought for the future. sarri.greek (talk) 10:33, 13 February 2020 (UTC)Reply

Thanks! Unfortunately I know far too little about editing modules to be able to help change the table display. —Mahāgaja · talk 10:37, 13 February 2020 (UTC)Reply

This is generally called a g-stem (e.g. DIL and Stifter) yet its genitive is ruirech, characteristic of a k-stem (cf. rí and brí, whose genitive singulars end in -g.), making it really hard for me to identify it as a g-stem. mellohi! (僕の乖離) 15:56, 25 February 2020 (UTC)Reply

There isn't really any difference between -k and -g after an unstressed vowel (or between -t and -d, either). They're only calling it a g-stem because it's an old compound of rí. —Mahāgaja · talk 16:06, 25 February 2020 (UTC)Reply

Do any other g-stem nouns other than those three exist? (On an unrelated note, the EDPC treats dét as an exclusively neuter o-stem, instead of the neuter nt-stem designation assigned by virtually everyone else.) mellohi! (僕の乖離) 16:42, 25 February 2020 (UTC)Reply

I can't think of any others offhand. —Mahāgaja · talk 16:44, 25 February 2020 (UTC)Reply

There are two nouns that arose as consonant stems with a stem consonant in w, namely crú (“blood”) and cnú (“nut”). Is it safe to thus create a noun class category containing these two nouns like this? mellohi! (僕の乖離) 17:35, 25 February 2020 (UTC)Reply

@Mellohi!: presumably bó (“cow”) belonged to this class too, but I'm reluctant to call it a class synchronically for Old Irish unless other authors (Thurneysen, Stifter, etc.) already have. Otherwise it's probably best to just call them irregular. —Mahāgaja · talk 17:38, 25 February 2020 (UTC)Reply

@Mahagaja: Those sources group the two w-stems I brought up under the u-stems, which makes absolutely zero sense whatsoever (especially cnú, which neither inflects like a u-stem nor is of the correct gender). mellohi! (僕の乖離) 17:58, 25 February 2020 (UTC)Reply

Are you sure it's reasonable to say that earráid is borrowed directly from Latin? I get that we should be defaulting to Middle Irish over Old, but there is a cited entry in the DIL for a pre-modern form. I saw your edit, and worried that I'd overreached, as I certainly have in the past, but then I saw that there was an eDIL link. embryomystic (talk) 21:36, 25 February 2020 (UTC)Reply

The only two cites in eDIL that look like they could conceivably be Middle Irish are the last two, cited from journals (ZCP and ITS). Without looking the articles up, there's no way to know how old the cites are. —Mahāgaja · talk 22:28, 25 February 2020 (UTC)Reply

This verb is very controversial. I'm 100% sure that the future form we list here *ɸiɸitsāti absolutely did not exist; everyone since Thurneysen derives the Old Irish future forms from *h₁ed-. EDPC does not think this verb existed at all, reconstructing *edeti instead of *ɸiteti, with the present stem ith- being secondarily analogical to ith (“grain”). Another view also sees the ith- in the present stem as secondarily analogical, but back-forming it from the suppleted verbal noun derived from the same root as ith instead. mellohi! (僕の乖離) 21:03, 26 February 2020 (UTC)Reply

I don't remember where I got *ɸiɸitsāti from, though I'm sure I didn't make it up myself. Go ahead and make any changes you deem necessary. —Mahāgaja · talk 21:12, 26 February 2020 (UTC)Reply