Case in Romance

[GoshDiggityDangit]

Hello! I'm once again working on a Romlang, and just like last time it is set in Britain. I have a few goals for this project, now that I am armed with the knowledge that a British Romance language might be more conservative and less innovative:

- Retention of geminate consonants
- B & V do not merge
- Retention of case

This last goal is the most vexing to me now, and so I come to you all for help. I'm thinking that the language, named Britto, should sound and look kinda like Italian. However, I'm fascinated as of late with the idea of Old French's case system. Of course, French and Italian are... not very similar. I was wondering if y'all might have some ideas as to how to mend these disparate parts?

Thanks for your time, Gabby

[VaptuantaDoi]

Hello! I'm once again working on a Romlang, and just like last time it is set in Britain. I have a few goals for this project, now that I am armed with the knowledge that a British Romance language might be more conservative and less innovative:

- Retention of geminate consonants
- B & V do not merge
- Retention of case

This last goal is the most vexing to me now, and so I come to you all for help. I'm thinking that the language, named Britto, should sound and look kinda like Italian. However, I'm fascinated as of late with the idea of Old French's case system. Of course, French and Italian are... not very similar. I was wondering if y'all might have some ideas as to how to mend these disparate parts?

Thanks for your time, Gabby

I reckon retention of geminates is not unrealistic. Some very western dialects like High Aragonese and Gascon dialects in the Pyrenees have retained Latin voiceless stops intervocalically – although some people have claimed this was a reversal, whereby TT T D → */t d ð/ which was then interpreted by early Basque speakers as /t t ð/, with early Basque having only a /t : ð/ contrast intervocalically, but it's still a conservative feature. There is some merit to the idea of a "peripheral" area in a language family, where early isolation leads to conservatism; compare how case was only retained in Gallo-Romance and Romanian.

I think retention of case is certainly possible; one of the driving factors for its loss in Gallo-Romance was phonological erosion to basically nothing (masculine nouns were -Ø in the obl.sg. and nom.pl. and -s, later becoming silent, in the nom.sg. and obl.pl.; feminine nouns tended to just be -e sg. and -es pl. with no case marking; and no nouns distinguished more than three forms*). If you retain final vowels, then there'd be a stronger phonological marker of case. Italian does retain some aspects of the nominative in the masculine plural -i (some argue this is from -ŌS which is plausible phonologically, but does not explain why many northern dialects have both plurals in -s, which must be from -ŌS, and palatal plurals which very strongly suggest -Ī).

If you have a Sardinian or Eastern vowel system (at least one of which appears in African Romance, and both of which appear in Southern Italy, suggesting that their current range is much smaller than their historical range), the distinction between nom.sg. -US and obl.pl. -ŌS would even be maintained, which would further bolster the case distinction.

There's also the question of articles; in Francoprovençal, less vowel reduction occurred than in Oïl, meaning that nominative and oblique articles were better distinguished. Some dialects still maintain a two-case system marked only on the articles:

lʏ ˈmamɐ ʃ ɛ aʃø̝ˈtaeø ʃʊg lɔ bɔr dɛ la ˈkuᵏsʏ
the.F.NOM mother REFL has sat on the edge of the.F.OBL bed
"The mother sat on the edge of the bed." (Lens Valais)

This also has the benefit of preserving case in the feminine, which was entirely lost in Gallo-Romance.

Say you have an Eastern vowel system, and retain final -S unlike in Romanian or Italian, and also generalise the masculine nominative article to the feminine like in Valais Francoprovençal. Then you produce this system:

Code: Select all

Masculine:
    most nouns:   article:
    sg    pl      sg    pl
nom -us   -i      le    li
obl -u    -os     lu    los

Feminine:
    most nouns:   article:
    sg    pl      sg    pl
nom -a    -as     le    las
obl -a    -as     la    las

That looks sufficient material to retain case. Another option would be to preserve the final nasal in ILLAM (fem.obl.sg.); French does retain it in MEUM → mon, REM → rien, so it's not too great of a stretch. Then you get:

Code: Select all

Feminine:
    most nouns:   article:
    sg    pl      sg    pl
nom -a    -as     la    las
obl -a    -as     lan   las

I'm not sure how realistic retaining feminine nom.pl. -AE is; it's pretty thoroughly lost in Gallo- and Italo-romance.


tl;dr

- Retaining geminates is probably fine
- A conservative Italian style vowel system and keeping -S makes case retention more realistic

---

* This is turning into a tangent, but there is at least one example of a noun with four distinct forms in Old Provençal; "child" nom.sg. enfas, obl.sg. enfan, nom.pl. enfanh, obl.pl. enfans. But that's a very rare case.

[GoshDiggityDangit]

I don't really understand what metaphony is, but would case endings be affected by this process?

[Creyeditor]

Metaphony is basically Romance Unlaut, IINM. Just with more vowel height and less vowel backness involved.

[Salmoneus]

Metaphony is basically Romance Unlaut, IINM. Just with more vowel height and less vowel backness involved.

Correct. It's basically raising of vowels before /i/ and /u/. It can also be connected to breaking of low mid vowels, but I've seen different opinions on whether that was originally a different thing or not*. Someone who knows more than me will have to explain.

But in any case, no, metaphony won't change the case endings. However, metaphony could preserve case.

You could for instance have (sg., pl.):

nom. sèrvo, sérve
acc. sérvo, sèrvo
gen. sérve, sèrvo
dat. sèrvo, sérve

The two problems, however, are that a) this is a really confusing paradigm!, and b) it would only apply in a subset of 2nd declension nouns, not elsewhere. I presume that the combination of these two problems is a big part of why this system did not survive anywhere!



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


* I've seen a suggestion that actually what happened was the metaphony resulted in vowel breaking only in LONG (i.e. open-syllable) mid vowels (with ê: > ei, è > ie). With subsequent ei>ê: reversal (or ei>i in some dialects), this left vowel breaking only in long low mid vowels. Different romance language then generalised either a) vowel breaking in metaphonous circumstances regardless of length; b) vowel breaking in long low mid vowels regardless of metaphony; or c) vowel breaking in low mid vowels regardlss of either metaphony OR length. [worth pointing out that the contexts for both metaphony and length were breaking down, that metaphony only survived in some dialects, and that vowel length only survived in even fewer dialects].

I can't remember why this seemed like a good idea, though. I think it had something to do with rhaeto-romance, where long vowels survived.

[GoshDiggityDangit]

How do the Latin declensions play into this? This has been a very confusing part of developing case in a modern Romance language in particular: Do declensions remain distinct? Do some fall into each other? Which ones?

[WeepingElf]

The dative and ablative cases were already identical in form in most declensions in Classical Latin; and the loss of final -m led to the accusative merging into this as well. On top of this, there was a syntactic change: the genitive was ousted by de + ablative, and the dative by ad + accusative. The result was a two case system with only a direct (< nominative) and an oblique case (< dative, ablative and accusative). This system was still present in Old French; but with paradigms like he following:

direct sg. murs, pl. mur
oblique sf. mur, pl. murs

the collapse of this case system was only a question of time.

[GoshDiggityDangit]

And so by the time that this system developed, the declensions were no longer meaningfully distinct from each other?

[GoshDiggityDangit]

And another thing, would the noun endings (I think I will just take VapuantaDoi's example) put Britto into a pluralization class of its own? As feminine nouns affix -s to the end, so do masculine oblique plurals, but the masculine subject have the opposite effect.

[WeepingElf]

And so by the time that this system developed, the declensions were no longer meaningfully distinct from each other?

No. The declensions underwent some mergers - 4th merged into 2nd, 5th into 1st - but one can say that there are still three declensions in languages such as Italian.

[VaptuantaDoi]

And so by the time that this system developed, the declensions were no longer meaningfully distinct from each other?

No. The declensions underwent some mergers - 4th merged into 2nd, 5th into 1st - but one can say that there are still three declensions in languages such as Italian.

And even the fourth is still marginally distinct in Italo-Romance (although not via direct inheritance of the endings).

And another thing, would the noun endings (I think I will just take VapuantaDoi's example) put Britto into a pluralization class of its own? As feminine nouns affix -s to the end, so do masculine oblique plurals, but the masculine subject have the opposite effect.

It would put it in a class with Old Gallo-Romance (basically Old French and Old Provençal).

[GoshDiggityDangit]

I'm not sure how realistic retaining feminine nom.pl. -AE is; it's pretty thoroughly lost in Gallo- and Italo-romance.

Wouldn't the ending be more likely to stick around given a Sardesque vowel system?

[VaptuantaDoi]

I'm not sure how realistic retaining feminine nom.pl. -AE is; it's pretty thoroughly lost in Gallo- and Italo-romance.

Wouldn't the ending be more likely to stick around given a Sardesque vowel system?

I don't know. It was lost in Gallo and Italo-romance, even where it would have allowed an unambiguous case distinction in the former. It may be that already by the time of proto-Romance -AE was lost; afaict there are no vestiges of it in any branch of Romance (maybe in some pronouns but it's hard to tell cos a lot of Italian languages had /as/ → */aj/ → /e/ anyway). Proto-Italic had *-ās as a 1st declension nom. pl. ending, which was analogically replaced with *-āī > -AE by analogy with the second declension. It's possible that rural Latin retained -ĀS in the nominative, which spread to all of Proto-Romance.

[GoshDiggityDangit]

What mergers took place in particular? It's looking like the 1st declension remained distinct, 2nd and 4th merged or could've been distinct, and 3rd and 5th merged.

[VaptuantaDoi]

What mergers took place in particular? It's looking like the 1st declension remained distinct, 2nd and 4th merged or could've been distinct, and 3rd and 5th merged.

In Gallo-Romance, the 1st and 2nd were continued as the default feminine and masculine declensions. The 3rd was the next most populous, containing both masculine and feminine nouns, having generally no number distinction in the plural, and sometimes having irregular nom.sg. forms (like Old Provençal neps "nephew nom.sg.", nebot "obl.sg." < NEPŌS, NEPŌTEM). The fifth, which was tiny anyway, merged into the 2nd (e.g. DIĒS, DIEM > Old French dis "day nom.sg.", di "obl.sg.") and the fourth merged into the 2nd (CORNŪ, CORNŪ > OFr cors "horn nom.sg.", cor "obl.sg.") or 3rd (MANUS, MANŪS > OFr main "hand nom./obl.sg.", mainz "nom./obl.pl.", albeit an irregular one).