User talk:Nicodene: difference between revisions

Archives: 2021–2

Hi Nicodene, and happy New Year—I wonder if you can shed some light at the long-lasting RFV for auctorico. It's not clear to me what the relationship between the forms auctorico, auctoriço, and auctorizo is, and though I initially thought it might be purely orthographic based on the ç form, the descendants at auctorico do look like they imply /k/ for the second c. Based on the existence of German sources attesting the spelling there might be a distinction between auctorico as an orthographic variant of -izo and auctorico as a form separately derived from -ico. —Al-Muqanna المقنع (talk) 19:38, 31 December 2022 (UTC)Reply

Hi @Al-Muqanna and apologies for the delay.

The RFV on auctorico can, of course, be closed thanks to the FEW citation. That's despite the entry being titled with an asterisk, which simply means, in Von Wartburg's style, 'unattested in Classical or Late Latin'.

Judging by what you have said regarding spellings of the type ⟨auctoriço⟩, I would expect them to have been used later than the early medieval attestations for auctoricare. I'd be surprised if the latter survived long enough to be confused with the former. Nicodene (talk) 02:19, 21 January 2023 (UTC)Reply

Can you try to translate the UDHR in Proto-Romance, Proto-Italo-Western Romance, Proto-Western Romance, Proto-Gallo-Romance, Proto-Ibero-Romance, Proto-Italo-Romance, and Proto-Balkan Romance, by using only words that are reconstructed from their respective descendants? Kwékwlos (talk) 20:30, 3 February 2023 (UTC)Reply

@Kwékwlos Seems a bit time-consuming to do all of them, but I'll try my hand at Proto-Gallo-Romance:

*/tót ɔ́ːmen nái̯sent líu̯ɾe ed eɡʷáːl en deɲtáːt ed en dɾéi̯ts sont dodáːt de raʣóːn e konoi̯sɛ́nʦa de ʤustétsa e déːβent tɾai̯táːɾ se kom fɾáːdɾe/

Nicodene (talk) 23:28, 3 February 2023 (UTC)Reply

I notice a mistake: Latin et > PGR */ed/ on one but */e/ on several others. Kwékwlos (talk) 12:24, 4 February 2023 (UTC)Reply

@Kwékwlos Intentional, one form being prevocalic, the other preconsonantal. I suppose one can still represent both phonemically as /ed/. Nicodene (talk) 12:36, 4 February 2023 (UTC)Reply

Alright. Next is Proto-Italo-Western Romance. Kwékwlos (talk) 19:08, 7 February 2023 (UTC)Reply

I'll humour you, as I find this interesting.

/tótᵗi ɔ́meni náscent líβri ed ekʷáli en deɲɲetáte ed en dᵉréktos sont dotáti de raʦʲóne ed konoscɛ́nʦʲa de justéʦʲa ed déβent traktáre se komo frátri/ Nicodene (talk) 19:31, 7 February 2023 (UTC)Reply

Why is t voiced in *ed, even though Italian does not show voicing of stops (except for some irregular cases)? Kwékwlos (talk) 20:58, 7 February 2023 (UTC)Reply

@Kwékwlos Because of Italian ed. No Romance language (modern or otherwise) reflects a /t/ outcome, as far as I know. Perhaps there occurred an early analogy with ad, which is likewise found nearly always in 'liaison' contexts. Nicodene (talk) 21:06, 7 February 2023 (UTC)Reply

And what about Proto-Ibero-Romance and Proto-Balkan Romance? Kwékwlos (talk) 19:49, 21 February 2023 (UTC)Reply

@Nicodene So I am trying to understand Proto-Balkan Romance. What does its UDHR look like (with IPA and Romanian orthography)? Kwékwlos (talk) 22:30, 21 March 2023 (UTC)Reply

I'm not good with Balkan Romance, sorry. Not yet anyway. Nicodene (talk) 23:01, 21 March 2023 (UTC)Reply

@Nicodene I've noticed you have done some work on Ibero-Romance. What would the PIbR UDHR look like (reconstructed from the Spanish and Portuguese UDHR)? Kwékwlos (talk) 15:53, 9 April 2023 (UTC)Reply

/ˈtodos los ˈɔmnes ˈnast͡sen ˈliβres e eˈgwales en diniˈdade e en deˈrei̯tos/

/esˈtan doˈdatos de raˈt͡son e konsˈt͡sjɛnt͡sja de ʒusˈtit͡sa e ˈdeβen trai̯ˈtarse komo jerˈmanos/ Xenos melophilos (talk) 19:35, 13 April 2023 (UTC)Reply

In a word such as vesâjo, final -o from -um is preserved. Is this really the case? Kwékwlos (talk) 11:27, 8 February 2023 (UTC)Reply

@Kwékwlos Yes, it really is. Specifically, the final vowel is preserved when needed to support a consonant cluster (at the stage */βizádɡo/). Some other examples are given here (under 'Arpitan'). Nicodene (talk) 12:39, 8 February 2023 (UTC)Reply

This template is designed to get all of its data from the entries linked to in its parameters, except for the name of the language and things like |bor= as applied to the first parameter. It used to be that putting a term without descendants in its numbered parameters would result in a big, fat module error. Given that the most common reason for such errors was people removing Descendants sections from entries that had no visible indication as to the presence of the entry where the actual error displayed, I got consensus to change that.

Now, instead of displaying a module error, it temporarily displays in the entry itself as if it were {{desc}}, and adds a maintenance category named after the language where the expected descendants section would be (I decided to do it that way because there's no point in alerting Proto-Indo-European editors about an issue relating to Franco-Provençal, even if the mechanics of Wikimedia categories dictate that the Proto-Indo-European entry is the one that shows up in the category).

It seems to me that this is the most civilized way to deal with such things. I've always said that a module error should only be used for emergencies- otherwise it's like triggering a fire alarm and evacuating the building because someone forgot to flush before leaving the restroom.

That said, the fact that the template doesn't halt everything and alert the armed forces when you do it doesn't make it right to abuse the template. Please don't put terms in the {{desctree}} positional parameters that don't have Descendants sections. Try to figure out other ways to make the display come out correct.

The obvious workaround would be to change the Alternative forms section so that only the alt forms you want to display in the ancestor's {{desctree}} use {{alter}} or {{alt}}, and to use {{l}}, etc., to link to the others ({{desctree}} only looks for the first 2 templates). If that seems too much like a kludge, we need to think about how the template should work, and make a proposal at the Beer parlour or the Grease pit so we can recruit a Lua editor to implement it (I know a lot about Lua in general, but not enough about the specifics to do more than elementary and extremely obvious emergency fixes).

Thanks! Chuck Entz (talk) 00:37, 12 February 2023 (UTC)Reply

When did "-e" disappear in Romance? Kwékwlos (talk) 14:15, 4 March 2023 (UTC)Reply

Well, never in Italian. But I suppose you're asking about Western Romance.

It would seem to have disappeared, at the PWR stage, after /n l r/. This is the state of affairs that remains in Venetian for example (per A Linguistic History of Venice, p. 62). Portuguese attests an early /e/ loss in that environment as well, considering that intervocalic single Latin /n/ and /l/ regularly delete in that language, without a trace, but not in words like cabedal or coraçon (hence these had early final /e/ deletion, making the previous consonant no longer intervocalic).

This isn't the case for final -e after other consonants, necessarily. A look at the descendants of Latin rete, for example, shows it surviving after /t/ in Portuguese (+Galician etc.), Venetian, and Istriot.

Nicodene (talk) 14:31, 4 March 2023 (UTC)Reply

The Glosas Emilianenses clearly shows preservation of -e after n and r. Kwékwlos (talk) 14:38, 4 March 2023 (UTC)Reply

I don't know if I would trust them as really reflecting phonetic writing. Final -e in infinitives is generally written for example but then you get <nafragarsan> = naufragare se habent. The preterite -avit alternates with -ot. Quomodo is written rather than como. Peccatos rather than *pecatos. -aria and -orio without indicating metathesis of /j/. The orthography is clearly heavily Latin-influenced, at a level well beyond what we see in later Old Spanish. Nicodene (talk) 14:52, 4 March 2023 (UTC)Reply

Could you look through their edits and finish the cleanup I started? They've obviously been doing some research in some good references, but I don't know Old French enough to tell if their cluelessness extends beyond just Wiktionary formatting. I hope they can learn how to do decent Old French entries- we could use the help. Chuck Entz (talk) 00:42, 13 May 2023 (UTC)Reply

We all have to start somewhere, I suppose. I've cleaned up the rest. Nicodene (talk) 01:27, 13 May 2023 (UTC)Reply

They have created four more entries since then, which look as bad as the previous. I've given them a warning. Benwing2 (talk) 02:08, 27 May 2023 (UTC)Reply

Based on the recent BP discussion, I wrote a script to clean up Latin etyma in Romance entries. It has some language-specific info, specifically on how to map language-specific suffixes like Spanish -dad to the corresponding Latin lemma and non-lemma etymon (in this case, -tās and -tātem). I can add support to the script fairly easily for the other languages I'm familiar with (Italian, Portuguese, Old Galician-Portuguese, French, Old French, Galician to some extent, Catalan to some extent, maybe some of the minority Iberian languages since they're pretty similar to other Iberian languages), but I'd probably need help for Romanian and maybe Venetian and Sicilian. It is running now for Spanish. It looks that it will change around 3,285 entries outputting about 375 warnings. It's partway through and I put the warnings generated so far here: User:Benwing2/clean-spanish-latin-etyma-warnings. Can you take a look? Some of them are false positives that can be ignored, but others represent real issues. BTW I notice you replaced {{inh+}} with From {{inh}} in at least one place, undoing a change by User:Sarilho1 (who is one of the primary Portuguese editors). Was there a reason for that? IMO the plus etym templates are especially important for Romance languages because they surface to the user whether a term is inherited or borrowed, which is often far from obvious for these languages. Benwing2 (talk) 05:07, 22 May 2023 (UTC)Reply

@Benwing2 I can help with the other languages when the time comes. As for the errors, I'll take a crack at sorting them out now.

If one of my edits happened to remove inh+, it was probably because I regard that as the default for Romance < Latin etymologies. That is, I'll mark borrowing with bor+ but leave inheritance unmarked. That said, I have no problem with switching to inh+. It wouldn't hurt to be more explicit. Nicodene (talk) 06:41, 22 May 2023 (UTC)Reply

Sounds good, thank you. I have pasted in the remaining warnings into the same location at the bottom. Benwing2 (talk) 07:48, 22 May 2023 (UTC)Reply

@Benwing2 I've finished running through the list. Nicodene (talk) 19:45, 22 May 2023 (UTC)Reply

Awesome, thanks! I am currently running the same script for French, and here are the French warnings: User:Benwing2/clean-french-latin-etyma-warnings. Note that the terms may appear to be in random order but it's just because they are sorted from right to left (which groups similar suffixes). Benwing2 (talk) 19:50, 22 May 2023 (UTC)Reply

@Benwing2 I've looked through those as well now. Is Italian in the works? Nicodene (talk) 03:25, 26 May 2023 (UTC)Reply

@Nicodene Thank you! I have already added support to my script for Portuguese and I'll be running it shortly; Italian will be next. Benwing2 (talk) 18:55, 26 May 2023 (UTC)Reply

See User:Benwing2/clean-italian-latin-etyma-warnings, User:Benwing2/clean-portuguese-latin-etyma-warnings and User:Benwing2/clean-catalan-latin-etyma-warnings. The script is making changes for all three languages now (and should be done within at most an hour). Benwing2 (talk) 01:58, 27 May 2023 (UTC)Reply

Hi. I notice you have been moving reconstructed Latin verbs from the first-person singular present indicative to the infinitive, and moving some nouns from the nominative to the accusative. I'm totally fine with the former practice and I think there was even a BP discussion about this, but I'm not so sure about the latter. IMO at least nouns in -us, -a and probably -is should stay at the nominative; not sure about nouns in -iō/-iōnem and such. I can see the argument that most Romance nouns are derived from the accusative, but some of course are not (e.g. uomo, moglie, sœur) and Old French and Old Occitan still had a nom/acc distinction. In general, having different forms lemmatized for reconstructed vs. non-reconstructed Latin complicates handling of Etymology sections, and there are of course gray areas (e.g. what about "Vulgar" terms that are attested e.g. in Appendix Probi)? It could also be argued that all these reconstructed terms are really Proto-Romance terms rather than Latin terms and belong under a Proto-Romance header. At the very least, I think we need a clear policy spelled out about which Latin terms are lemmatized where, including all the edge cases, so there won't be edit conflicts and it will be clear how to cite the terms in Etymology sections. Benwing2 (talk) 02:23, 29 May 2023 (UTC)Reply

@Benwing2 I do regard reconstructions based on Romance data as (Proto-)Romance themselves, whatever our current labelling is. Old French and Old Occitan nouns are themselves lemmatized by the accusative, or 'oblique', so adapting this practice seemed consistent to me. (It is also the practice of the Dictionnaire Etymologique Roman.) In cases like *fetonem, I can't see that a Classical-style nominative like *feto actually ever existed- Romance rather suggests *fetonis, as indicated there. For Proto-Ibero-/Italo/Balkan Romance, no active nominative case can be reconstructed for nouns, so lemmatizing by the etymological accusative seemed inevitable. I've yet to ever come across a case where Romance inherited a nominative singular that isn't attested in Latin, but if one comes up, we can add mention it in the reconstruction notes. Nicodene (talk) 03:13, 29 May 2023 (UTC)Reply

OK, well (a) I still think you should limit this practice to imparisyllabic third-declension nouns; (b) we still need a writeup on which forms are used to lemmatize nouns under which circumstances (you didn't address my concerns about edge cases). Are you willing to write this up and put it in WT:About Latin? Benwing2 (talk) 03:23, 29 May 2023 (UTC)Reply

@Benwing2 The issue remains that e.g. Proto-Ibero-Romance can't be reconstructed with a general noun-case system.

By edge cases do you mean attested 'Vulgar Latin', e.g. coleo? I don't see any reason to treat such entries differently from normal Latin ones with regards to lemmatization. They were subject to traditional Latin rules with regards to case inflection in the context of the texts in which they are attested. E.g. coleo is attested, per the provided quote, in the ablative plural, a form which certainly was extinct from popular speech in eighth-century France.

I can do a write-up of my current approach, but it would simply amount to 'use the accusative singular, except for feminine nouns in /-a/, where it's not worth bothering' (because /-am/ and /-a/ merged very early on, and the Romance descendants equally reflect either). Nicodene (talk) 03:38, 29 May 2023 (UTC)Reply

By attested I mean in things like Appendix Probi, Reichenau glosses, etc. I am fine with using nominative/first-person singular for these. As for using the accusative e.g. even for nouns in -us (where it unhappily erases the masc/neut distinction), this definitely needs discussion before you start moving a lot of terms. If you could, please create a BP discussion. As for there being no reconstructible nominative case in Ibero-Romance/Balkan Romance (I should add that categories like Category:Spanish terms inherited from Latin nominatives somewhat belie this), I don't see why this is an issue. Our reconstructions are somewhat artificial in any case in that e.g. we project the terms back to the 10-vowel Latin system distinguished by length when there clearly was no length in Proto-Romance; so I don't see why projecting a nominative case is problematic for nouns in -us and -is. Benwing2 (talk) 03:49, 29 May 2023 (UTC)Reply

@Benwing2 Yes I'll make a BP discussion with a write-up of the various points and counter-points.

I don't regard the reconstructions as projecting back to a Classical Latin pronunciation. Even if the head word is spelled as *fētōnem, the provided pronunciation is actually /feˈdoːn/. (The spelling, in effect, is simply there in honour of the etymology fētus + -ō/-ōnem.) This is briefly mentioned under About Vulgar Latin#Spelling, but a more detailed write-up may help. Nicodene (talk) 04:02, 29 May 2023 (UTC)Reply

Merci pour les quelques mises en page que tu as faites à mes quelques entrées! - Waelsch (talk) 15:22, 8 June 2023 (UTC)Reply