Talk:dunc

About dunque and unque

We see that many of the western Romance cognates show the infamous adverbial /-s/ (another example is French onques). It's actually possible that the /-s/ was also added in Tuscany as well, early enough to participate in the general sound change /-as/ > */-ai̯/ > /-e/, as in Latin amicas > Italian amiche 'girlfriends' (not *amice) or Latin portas > Old Tuscan porte 'you carry'. Nicodene (talk) 19:16, 21 May 2022 (UTC)

Yes, that did cross my mind. I brushed the idea off, thinking that it could bave been an influence by the numerous Italian adverbs ending in -e (bene, male, sempre, -mente), but the -s is a more solid explaination.

On the subject, is there any publication actually accounting for this -s that we've been talking so much about lately? Google isn't really helpful in looking for things that consist in a single letter.

To justify this *adunqua reconstruction I put here, it's a way to organize my thoughts before writing a more elaborate post on the scriptorium. Catonif (talk) 21:34, 21 May 2022 (UTC)

Here are some useful sources that comment on that /-s/:

• Jensen's Old French and Gallo-Romance syntax, page 398: https://i.imgur.com/wVym8cQ.png
• Lausberg's Lingüística románica: Morfología (Spanish translation), page 145–6: https://imgur.com/a/R1O8xMG
• Penny's A history of the Spanish language, page 131: https://i.imgur.com/UYfcMRN.png
• Wolfe and Maiden's Variation and change in Gallo-Romance grammar, page 349: https://i.imgur.com/3lzeRzP.png

Incidentally, if you're right to reconstruct an /a-/ in *adunqua, that suggests another possible etymology: ad + unquam. Nicodene (talk) 07:42, 25 May 2022 (UTC)

@Nicodene Thank you very much for these sources and sorry for the late reply. About /a-/, the more I look into it, the more I get doubtful about its presence in the Proto-Romance form, and start to think more that it emerged later, although still lean towards the theory of it being there. ad + unquam does very well qualify phonologically, but it seems hard to justify its semantical shift and its syntactical strangeness (*to ever / *everly ?). Catonif (talk) 07:58, 31 May 2022 (UTC)

Yes, it's not very plausible semantically.

I took a look at the FEW entry and found some interesting information. Here is a translation in English, with my comments in square brackets:

'Dunc [meaning 'then'] is attested eight times in Latin inscriptions, four of which come from Central Italy, one from Dalmatia, and one from Provence (Grober ALL 2, 103; Zimmermann ALL 5, 570; Pirson 252). It is more likely to have originated by analogy with dum from the pair tum, tunc (Schuchardt Lbl 12, 413) than from donique (Z 16, 243). Dunc was affected by tunc not only phonetically but also semantically. Outside of Gallo-Romance it survived in Italian dunque, Piemontese dunch, Old Italian donqua, Old Genoese/Old Lombard/Old Venetian donca (AGI 10, 145, 249; SFR 7, 50), Sicilian nunca (RLomb 40, 1117), Arezzo donqua (R 18, 613), Viareggio donca (Z28, 163), Arcevia donga, Monferrato dunca (AtrP 23, 353), Lombard/Venetian donca, Trieste donchia (A GI 10, 460); Gorizia dunza, Catalan doncs, Old Spanish doncas. With the exception of Catalan [???], all of these forms mean 'therefore'. In Catalan as in Gallo-Romance both meanings [???] coexist. Accordingly, the meaning 'therefore' must have developed early on. It is therefore quite likely that ['Vulgar'] Latin already possessed the meaning 'therefore'. It is also true that Loftstedt was able to prove the meaning of 'therefore' for the particle dum in Late Latin texts (see Frstschrift zum 65. geburtstage von P. Person; Ipsala 1922.) -- Phonetically, dunc was in many areas influenced by unquam, hence Old French donque, Old Italian donqua, etc. (MLR Gr 2, 625). In Gallo-Romance there is often an adverbial -s, cf. Old French donques, etc. In Modern French, this donques lives on in the pronunciation dõk, which serves to distinguish the word [donc] from dont.'

I don't quite understand the parts that I've marked with question marks. I'll think about it more when I've the time. Nicodene (talk) 12:40, 31 May 2022 (UTC)

Ok I've incorporated this information into the relevant entries, more or less. May have missed some things. Nicodene (talk) 08:01, 1 June 2022 (UTC)

Thank you for incorporating the information, though I see you've descended all of them directly from dunc, which seems pretty crude, considering that while we might consider the */a-/ and the */-s/ as variations which emerged later, the final */-a/ was unlikely missing: the descendants which don't exhibit it have undergone predictable apocope. As for the /-k-/, it is safe to presume a de-labialization of an older /-kʷ-/, still preserved in Tuscan. All descendants would thus be under the *dunqua page.

By the way, from the pages I've removed the descendants Central Marchese (which is a variety of Central Italian and not considered a language by the WT:LL) and Tuscan, since it contained the terms that lacked Tuscan anaphonesis (w:it:Anafonesi), and were therefore found in texts outside of Tuscany, mostly Umbria. Moreover, Tuscan is just another way to refer to Standard Italian in contexts of phonological evolution, rather than the standardized form. Catonif (talk) 18:05, 2 June 2022 (UTC)

┌┘

Hi @Catonif. I have moved the conversation here from your talk page.

The forms in question were taken from the FEW entry translated above. Specifically donqua (Arezzo), donca (Viareggio), and donga (Arcevia). The first two are Tuscan (not Umbrian), and the last is indeed Central Italian, which Wiktionary may not recognize as a separate entity yet, but actual Romance linguists do. Cf. the Oxford Guide to the Romance Languages, which treats Tuscan (together with Standard Italian) separately from Central Italian. Likewise Wikipedia. And removing 'Tuscan' as a category seems counterproductive to me: none of these forms are part of Standard Italian, so they should be distinguished somehow.

You're right to question the lack of anaphonesis in the above Tuscan forms, but as the AIS maps for 'lingua' and 'giuntura' demonstrate, non-anaphonetic forms (lengua, giontura, le congionture) are in fact found around Arezzo and Viareggio, where peripheral Tuscan dialects are spoken (per, for instance, Pellegrini's linguistic map of Italy).

The FEW is perhaps the best Romance etymological dictionary that exists, in any case, and is reliable in such matters.

Speaking of which, I wasn't the one who decided to place all those derivatives under the entry for dunc; that is what the FEW itself does (the entire entry is titled, and dedicated to, dunc). It mentions some of the derivatives as being (subsequently) influenced by Latin unquam, so I have placed those under the reconstructed *dunquam.

However, deciding whether any given form derives from *dunquam(-as) or *dunc-as requires a detailed examination. So let's begin. (Pinging @Oigolue, who is good at Western Romance dialectology.)

---

For Sardinian, *dunquam would have yielded *dumba in Logudorese, and *dunqua in Campidanese, neither of which exist. So the source of duncas has to be *dunc-as, the latter morpheme being an adverbial /-as/. No external borrowing can explain the Sardinian duncas, considering its /u/, its final /-s/, and the fact that it is attested from early medieval times (per Wagner's Dizionario Etimologico Sardo).

For Castilian ('Spanish' in the narrow sense), @Ser be etre shi and I were unable to find any medieval form along the lines of *donquas; there was only doncas. For comparison, there are plenty of medieval attestations of nunqua < Latin nŭmquam 'never'. So doncas, if native, appears to derive from *dunc-as rather than *dunquam-as. On the other hand, it could simply have been borrowed from Old Aragonese doncas.

For Old Aragonese, we have both the aforementioned doncas and dunquas/donquas. The latter appears to be a native derivation from *dunquam-as; no other plausible explanation comes to mind. The form doncas, if native, can derive from donquas via a one-off sound change (which I find quite unlikely), or it could derive from *dunc-as. The third possibility is that it's borrowed from Old Spanish or Occitan doncas, or from Catalan donques.

Catalan donques, if native, can only derive from *dunc-as, considering the absence of /w/ in its pronunciation ([dɔ́ŋkəs, dɔ́ŋkes] per the DCVB). Or it could be borrowed from Aragonese or Occitan doncas.

Occitan doncas, if native, can equally reflect *dunc-as or *dunquam-as. There is no obvious way to determine which one it derives from. It could also be borrowed from Catalan donques, Aragonese doncas, or Old French dunkes. There also exists an Occitan donca, either from *dunc-as (with secondary loss of /-s/; more on that later) or borrowed from Old French dunque, Franco-Provençal donca, or Piemontese donca.

Old French dunkes/dunque (etc.), if native, can only derive from *dunquam(-as). (The expected outcome of *dunc-as would have been *donches, with palatalization of /k/ before /a/.) Or it could have been borrowed from Occitan donca(s) or Franco-Provençal donca.

Franco-Provençal donca, if native, can only derive from *dunquam. (As with Old French, the /ka/ of *dunc-as would have shown palatalization.) Or it could be borrowed from Old French donque, Occitan donca, or Piemontese donca.

Friulian doncje can only be a native Rhaeto-Romance form derived from *dunc-a(s). There is simply no other explanation for the palatalized cj. (More on the /s/-loss later.)

Ladin donca, if native, can only derive from *dunquam, considering its non-palatalized /k/. Or it could be borrowed from Venetian donca. I lean towards it being a borrowing, considering the above Friulian form, which cannot derive from *dunquam.

Tuscan donqua and dunque (also dunqua) can only be native and derived from *dunquam-(as). No borrowing can explain it, since the only other Romance outcomes with /w/ are found in Aragonese, which is, to put it mildly, not a likely source of borrowings in Italian.

Finally, we have 'Pan-Italian' donca (Piemontese, Lombard, Venetian, Tuscan, Central Italian, Neapolitan), dunca/nunca (Sicilian), dunche (Tuscan again), and dunchi (Sicilian again), which we will treat as one group. These forms, if native, cannot derive from *dunquam, because all of these languages retain native /kwa/, without exception. It is exceedingly unlikely that these forms can all be borrowed from Franco-Provençal or Occitan donca, considering that they extend down the entire length of Italy, all the way into Sicily. Furthermore, the parallelism between /o/ in Neapolitan donca and /u/ in Sicilian dunca corroborates the autochtony of the term in southern Italy (a recent borrowing, having swept through the area, would have probably ended up having the same stressed vowel in both languages). The general picture is of a native survival of *dunc-as across the peninsula, with widespread loss of /-s/.

---

Now to address the /s/-loss. In Standard Italian, at least, we know that the regular outcome of Latin final /-a(ː)s/ was either /aj/ if stressed or /-e/ if unstressed (certainly via an */aj/ stage). Hence *dunc-as would be expected to yield dunche, which is in fact an attested Tuscan form. Sicilian dunchi, if native, also shows a regular outcome of atonic /-a(ː)s/.

However, we find the more common donca/dunca all across Italy, with no such trace of a final /-s/. Does that make *dunc-as an implausible etymon?

It doesn't for at least two reasons. First, total loss of unstressed /-s/, without any remaining trace, is attested for both noun and verb endings in parts of Gallo-Italic and even Tuscan (per Gerhard Rohls's Grammatica storica della lingua italiana, §308).

Second, *dunc-as was an adverb, meaning that the final /s/ was not actually needed to mark any grammatical distinctions (person or number) nor, incidentally, to distinguish it from any other word (there was no such noun, for example, as *dunca).

And, in fact, we can compare the reflexes of another adverb with a similar ending: Latin forās (“outside”). Looking at AIS map 356, we find derivatives of the fora type all across Italy, from the extreme north to the extreme south, showing loss of /-s/ without any traces. Even the Rhaeto-Romance derivatives reflect /-s/ loss, despite the fact that these languages regularly conserve Latin /s/ in word-final position. We also find fora (without /-s/) in parts of Central Sardinia, the most isolated and linguistically conservative part of the island.

In fact, the /-s/ loss isn't even limited to Italy. In Occitan we find that the derivatives of *dē-forās 'from outside' show a universal loss of /-s/ (search 'dehors' here to consult the ALF). Cf. also Catalan fora (compare medieval fores) and Spanish fuera (compare obsolete fueras).

On the whole, then, *dunc-as is satisfactory on phonological grounds and adequately explains all of the 'Pan-Italian' outcomes.

---

Concluding thoughts:

• Sardinian duncas and Friulian doncje derive, without a doubt, from *dunc-as.

• 'Pan-Italian' donca/dunca/dunche/dunchi most probably derives from *dunc-as.

• Tuscan donqua/dunqua/dunque and Aragonese donquas/dunquas derive, without a doubt, from *dunquam(-as)

• Gallo-Romance, on the whole, shows a mixed situation. Perhaps Occitan donca(s), whatever its origin, exerted a strong influence on all the surrounding languages.

Nicodene (talk) 06:22, 3 June 2022 (UTC) Nicodene (talk) 06:22, 3 June 2022 (UTC)

Thanks for the movement of the page.

I see your point about Tuscan, even though it still seemed weird to me that the only term listed under Tuscan was one with peripheral characteristic uncommon to the rest of Tuscany.

It is interesting how most languages don't necesserily reflect the /-kw-/, though its indisputable presence in two languages as far away as Old Aragonese and Tuscan makes me wonder how could have it been absent in the space between them.

About Gallo-Romance, it seems more plausible for it to have developed from the /-kw-/ version, rather than everything being borrowed from Occitan.

About the 'Pan-Italian', retention of /kwa/ does have exceptions: Neapolitan addovonca and Tuscan dovunque (both from (ad) de ubi unquam) show parallelism with adonca and dunque. More parallelism can be seen with the Gallo-Romance terms of donk(w)a(s) by looking at the descendants of umquam.

About the -s loss, since *dunc-as is unrelated to unquam, what explanation would its /a/ have? All terms exhibit -a and some of them an additional -s, and hypothesizing and original -as with later loss of -s (which we mustn't forget the rarity of) instead of an original -a with an adverbial -s (attested in grater scale afaik) still doesn't justify the /a/. Catonif (talk) 10:58, 3 June 2022 (UTC)

@Catonif The adverbial /-as/ is mentioned by Lausberg (§699; cited in the screenshots from earlier), who provides the following examples:

• *primas > Old French primes, Old Spanish primas
• *certas > Old French certes, Old Occitan certas, Old Spanish ciertas

To the former I would add Italian prima (“before”), keeping in mind the complete /s/-loss in Latin forās > fora (found in multiple parts of Tuscany, per the AIS map from earlier). Spanish mientras < Latin dum + interim also bears mention.

Perhaps this /-as/ came into being as a result of the earlier adverbial /-s/ being attached to some common adverbs ending in /-a/, such as nunqua and ora, both mentioned by Lausberg. He also suggests that the Latin aliās could have played a role, though I struggle to see how, considering that it doesn't survive anywhere at all in Romance. On the other hand, the forās that we've been discussing shows a pan-Romance distribution and has the right ending.

------------------------

It now occurs to me that, in theory, the Old French donque(s) and Franco-Provençal donca could natively derive from *dunc-as if we suppose that the morpheme boundary prevented /k/ from palatalizing. The existence of donc as an independent word in both languages, with its final /-k/, would have helped in that regard. The general Gallo-Romance situation remains something of an enigma for our purposes.

As for the Italian retention of /kwa/, I was really just referring to the retention of /w/ in that sequence. Tuscan dovunque is an expected outcome, like other words suffixed with -unque < Old Italian unque/unqua (adverb) < Latin umquam. Adverbial /-(a)s/ may also have played a role here too, considering the /-e/.

Neapolitan addovonca, on the other hand, is genuinely surprising to me. So too are the medieval Gallo-Italic derivatives of unquam, which show lack of /w/ (and are indeed mentioned in the FEW entry). I'm a bit baffled, actually, considering that I couldn't find any trace of /kwa/ > /ka/, nor of /ɡwa/ > /ɡa/, anywhere in Italo-Romance or Gallo-Italic on the AIS maps from earlier, nor the maps for acqua and quattro. I'll have to look into this further. Nicodene (talk) 12:53, 3 June 2022 (UTC)

I didn't know about NavigAIS before this by the way, very cool tool. About Sardinian, the terms quando (9, 1032) and quanti (50) show delabialization of stressed /ˈkwa/ to /ˈka/ in both Logudorese and Campidanese, and so does unstressed /kwa/ in Pasqua (777). The delabalization from donqua to donca could also be influenced/borrowed from Gallurese/Sassarese (cf. Gallurese linga, where Sardinian yields the predicted lingua and limba in map 106). I also noticed Italian anche (< anco) as another adverb that's shifted it's word ending vowel to -e, unrelated to the adverbial -s, since -os > -i. Catonif (talk) 20:19, 4 June 2022 (UTC)

@Catonif Sardinian does have some cases of word-initial /kwa/ > /ka/, but in internal position the outcome of /kwa, ɡwa/ is consistently /ba, pa/ or /kwa, ɡwa/ (depending on dialect). Cf. the derivatives of Latin aqua, equa, quinquaginta, lingua, arcuatus, siliqua.

Considering that (a)duncas is already attested in early medieval Sardinian, I don't see how it could be the result of influence from Gallurese and Sassarese, which developed later as a result of colonization from the mainland. The complete absence of any such form as *dumbas or *dunquas, anywhere on the island, is also suspicious.

Vulgar Latin Pasqua can't be used here because Latin had Pascha (borrowed from Greek), without any /w/, and its direct survival in Sardinia needs no special explanation. The heavy Byzantine influence on early medieval Sardinia may be relevant.

Anche is rather doubtful; cf. Treccani's 'etimo incerto'. Why would Old Occitan ancui, meaning 'today', yield an Italian word that has an entirely different meaning? And why would the final -ui have been eliminated, when Italian has the same ending in several common words? (Lui, costui, cui...) Nicodene (talk) 22:40, 4 June 2022 (UTC)

Whoops, didn't know about Pascha. About anche, I was referencing to development from the older anco, which is attested and still found as a regional form, and not about the dubious ultimate etymology. Catonif (talk) 09:20, 17 June 2022 (UTC)

Yes it can be an example of /-o/ > /-e/, or the reverse, if anche is the original form. Treccani cites Dante as using anche, so it's not clear to me that anco must be the original. Nicodene (talk) 03:05, 18 June 2022 (UTC)

---