Two questions about Romance and Vulgar Latin

[Ars Lande]

I'm trying to figure out an odd thing in Romance languages, especially in the paradigm of habere.

ego -> io (yo, je, eu)
habeō -> Sp. he, Port. ai, Rom. am, It. ho, Fr. ai
habēs -> Sp. has, It. ai, Fr. as, and so on.
magis -> Sp. mas, It. ma/mai, Fr. mais, Rom. mai.
but videō -> Sp. veo, Fr. vois, but vedo, Rom văd.
bibō -> It. bevo, Rom. beau

So it seems that:
- Western Romance languages lose these voiced stops. So far so good.
- Eastern Romance languages though, sometimes lose them, and sometimes don't.

1) It looks a bit like a change deleted all voiced stops in certain positions, applying both in Eastern and Western Romance. If I'm not completely wrong about that, what would it be? (I'm guessing in posttonic position, but might be wrong).
2) What bothers is why Italian sometimes kept the voiced sound. (Romanian too, but not always in the same place). Were the verb forms regularized?

Do any of you have any idea? Or a good resource that might have an answer? (I have some good references on Old French, but nothing on Italian).

Also, I read contradictory things about vowel+nasal sequences (as in -um). Occasionnaly I read that they actually represent nasal vowels (later merged with oral ones), sometimes that the nasal was simply deleted. Do you know what's the current opinion on the matter?

[Lyra]

videō Cat. veig [be̞t͡ʃ]

Because more wood to the fire.


~Lyra

[Ser]

Yes, sound change rules don't apply universally.

- Western Romance languages lose these voiced stops. So far so good.

Latin intervocalic /b/ (ignoring vowel syncope next to /r l n/, as in nebula > niebla névoa) was generally conserved in Spanish and Portuguese—you're ignoring bibo > bebo (in Spanish at least); caballum > caballo, cavalo; cibum > cebo (in Spanish); ciba > ceba, ceva; nūbem > nube, nuvem; dēbēre > deber, dever, etc.

Intervocalic /b/ was only lost in function words: ibi > hi/y (Medieval Spanish adverbial pronoun, still found in the compound ad ibi > ahí [a.ˈi]), hi/y (Medieval Portuguese adverbial pronoun), ubi > o (Medieval Spanish interrogative/relative pronoun).

So, at least as far as Latin intervocalic /b/ in Spanish/Portuguese is concerned, it was preserved depending on how grammaticalized a word was.

[Ars Lande]

Ooops, my bad about intervocalic /b/!
There's still an odd,, irritating irregularity there, though.

Spanish also dropped /b/ in the imperfect of certain classes of verbs too: Lat. timēbat > Sp. temía vs It. temevo. Has any explanation been proposed?

[Legion]

There are different things going on.

Intervocalic [b] is generally preserved as [v/β] in all romance languages, but there might be sporadic loss

habeo > *ayyo > ai/ho/he (but habere > avoir/avere/haber; habere habeo > aurai/avrò/habré)
tenebam > tenía (but cantabam > cantaba)

Intervocalic [d] is lost in Gallo-romance and Ibero-Romance.

Intervocalic [g] is complicated, various things can happen to it, it is often reduced to a yod then subsequently lost (magis is an example, the forms mai(s) of French, Occitan and Portuguese have the -i- being a former yodified -g-).

However, ego > *eo is an irregular reduction just like habeo > *ayyo

[Drydic]

Also since I apparently never submitted my answer to the original second question


Latin VN# sequences were nasal vowels, even in Classical Latin. We can tell this because among other evidence they scanned as long vowels in iambic poetry, even though they were written as short vowel-consonant sequences.

[Ser]

Also since I apparently never submitted my answer to the original second question


Latin VN# sequences were nasal vowels, even in Classical Latin. We can tell this because among other evidence they scanned as long vowels in iambic poetry, even though they were written as short vowel-consonant sequences.

That's evidence that those vowels were long because of compensatory lengthening—it's not evidence that they were nasalized.

Allen in his book quotes Quintilian and Priscian saying that it was hardly pronounced (not deleted, but hardly pronounced), and from there he seems to simply guess they were nasalized. Somebody ought to travel back in time and record them.

[Ars Lande]

Thanks, Legion, for crushing my hopes of finding a regular pattern there

That's evidence that those vowels were long because of compensatory lengthening—it's not evidence that they were nasalized.

Allen in his book quotes Quintilian and Priscian saying that it was hardly pronounced (not deleted, but hardly pronounced), and from there he seems to simply guess they were nasalized. Somebody ought to travel back in time and record them.

Before I get the DeLorean out of the garage, there's still one thing I'm not so sure about. We have Latin -um > -o in various assorted Romance languages, which implies that either:
- the vowel wasn't just lenghtened, it also changed in quality somehow. (/o/ isn't the expected outcome of long u).
- or final -m was dropped without compensatory lengthening (which appears to be ruled out by scansion rules).

[Salmoneus]

I don't know what actually happened... but it's possible for it to have been a 'long vowel' for the purposes of scansion but 'underlyingly' a short vowel followed by a consonant for the purposes of sound change. That is, a long vowel could have been an allophone of a VN sequence when final, but not treated as phonemically a long vowel when it came to the vowel shifts.

[Legion]

Another possibility, iirc, may be that Vulgar Latin simply never underwent the -om > -um change that affected Classical Latin, and thus instead retained the Old Latin final -om up until the dialect split into Romance languages.

[Dewrad]

Another possibility, iirc, may be that Vulgar Latin simply never underwent the -om > -um change that affected Classical Latin, and thus instead retained the Old Latin final -om up until the dialect split into Romance languages.

Evidence from metaphonic alternations in several Romance languages rather contradicts this argument.

[Basilius]

I don't know what actually happened... but it's possible for it to have been a 'long vowel' for the purposes of scansion but 'underlyingly' a short vowel followed by a consonant for the purposes of sound change. That is, a long vowel could have been an allophone of a VN sequence when final, but not treated as phonemically a long vowel when it came to the vowel shifts.

Sound changes operate on surface forms.

Another possibility, iirc, may be that Vulgar Latin simply never underwent the -om > -um change that affected Classical Latin, and thus instead retained the Old Latin final -om up until the dialect split into Romance languages.

Evidence from metaphonic alternations in several Romance languages rather contradicts this argument.

You mean, e. g. in Portuguese?

AFAIK, it clearly shows that the vowel was different from short-o, but I don't know of any Romance material showing unambiguously that it was different from long-u. Any pointers?

OTOH forms like (Old) French <i>mien</i> <- meum (disyllabic in Classical) show that the final nasal was indeed pronounced.

[Dewrad]

I don't know what actually happened... but it's possible for it to have been a 'long vowel' for the purposes of scansion but 'underlyingly' a short vowel followed by a consonant for the purposes of sound change. That is, a long vowel could have been an allophone of a VN sequence when final, but not treated as phonemically a long vowel when it came to the vowel shifts.

Sound changes operate on surface forms.

Another possibility, iirc, may be that Vulgar Latin simply never underwent the -om > -um change that affected Classical Latin, and thus instead retained the Old Latin final -om up until the dialect split into Romance languages.

Evidence from metaphonic alternations in several Romance languages rather contradicts this argument.

You mean, e. g. in Portuguese?

Portuguese and Asturian immediately come to mind, yes, but the source I have to hand describes metaphony in the Gallo-Italian varieties.

AFAIK, it clearly shows that the vowel was different from short-o, but I don't know of any Romance material showing unambiguously that it was different from long-u. Any pointers?

The evidence suggests that the vowel reflexes of Classical -um and -us were distinct from that of -os (accusative plural), and that like final -i they triggered metaphony (in Gallo-Italian at least, this involved diphthongisation of the VL mid open vowels *ę *ǫ- I don't have appropriate sources on the effects in Ibero-Romance with me). This suggests that the outcome of these vowels was a high vowel, like -i, rather than a mid vowel. Hull (1984) provides the following reconstructed paradigm:

Code: Select all

apęrtus apęrti  > apięrtus apięrti
apęrtu  apęrtos > apięrtu  apęrtos

NE: I just realised that all of the above does not actually address the question of the quantity of final -um in CL (i.e. was it something like [ũː]). I don't think it actually has any evidence to provide here. I cannot for the life of me think of any forms in the inherited Romance lexicon which show a reflex of an unambiguously long post-tonic auslaut ū, so claiming that /o/ is an unexpected reflex is to my mind jumping the gun somewhat.

[Basilius]

AFAIK, it clearly shows that the vowel was different from short-o, but I don't know of any Romance material showing unambiguously that it was different from long-u. Any pointers?

The evidence suggests that the vowel reflexes of Classical -um and -us were distinct from that of -os (accusative plural), and that like final -i they triggered metaphony (in Gallo-Italian at least, this involved diphthongisation of the VL mid open vowels *ę *ǫ- I don't have appropriate sources on the effects in Ibero-Romance with me). This suggests that the outcome of these vowels was a high vowel, like -i, rather than a mid vowel. Hull (1984) provides the following reconstructed paradigm:

Code: Select all

apęrtus apęrti  > apięrtus apięrti
apęrtu  apęrtos > apięrtu  apęrtos

NE: I just realised that all of the above does not actually address the question of the quantity of final -um in CL (i.e. was it something like [ũː]). I don't think it actually has any evidence to provide here. I cannot for the life of me think of any forms in the inherited Romance lexicon which show a reflex of an unambiguously long post-tonic auslaut ū, so claiming that /o/ is an unexpected reflex is to my mind jumping the gun somewhat.

Yep, that's what I meant.

Also, front and back vowels needn't be parallel in terms of effects on the preceding vowel: IIRC in Portuguese short-o was "widening" (like a), -um was "narrowing", but both e's were neutral.

Sure, we have words like cornū and plurals like manūs, but in all of them the endings could be affected by analogy.

[Ser]

Before I get the DeLorean out of the garage, there's still one thing I'm not so sure about. We have Latin -um > -o in various assorted Romance languages, which implies that either:
- the vowel wasn't just lenghtened, it also changed in quality somehow. (/o/ isn't the expected outcome of long u).
- or final -m was dropped without compensatory lengthening (which appears to be ruled out by scansion rules).

I thought the [uː] > Portuguese/Spanish/Italian thing was meant for stressed syllables only, and that unstressed vowels are generally more drastic in the mergers. Though maybe I'm thinking too much about Spanish, where all of unstressed [uː ʊ oː ɔ] > [o], with a small number of exceptions.

[Salmoneus]

I don't know what actually happened... but it's possible for it to have been a 'long vowel' for the purposes of scansion but 'underlyingly' a short vowel followed by a consonant for the purposes of sound change. That is, a long vowel could have been an allophone of a VN sequence when final, but not treated as phonemically a long vowel when it came to the vowel shifts.

Sound changes operate on surface forms.

Only in reality... not.

I mean, obviously they sometimes can, which is why allophones sometimes split into new phonemes... but sometimes not, which is why they don't always.
[Or, if you insist upon a certain ideology: you can say that 'proper' sound changes always operate on surface forms, but only if you then admit that 'improper' sound changes can operate 'through analogy' at a phonemic level, including restoring phonemes that have been changed by the 'proper' sound changes. In reality, however, it's not clear that the change always happens clearly before the analogy... and if the analogous change-blocking happens at the same time as the change, this is logically equivalent to a change whose conditions depend on phonemes rather than only on phones.]
To give an example: the raising of /aI/ to /VI/ in some north american dialects has taken place AFTER the neutralisation of the intervocalic T-D contrast that conditions that change! On the surface level, there's just a flap, but speakers are able to recognise the underlying phoneme and raise or not raise the preceding vowel accordingly. Or, in Britain, look at sound changes affecting /T/. In Glasgow, where in uneducated speech there has been a soundchange /T/>[h], there is now a growing soundchange /T/>[f] - even though [h] from /T/ has exactly the same realisation as [h] from /h/. So for instance /hIN/ becomes /fIN/ while /h{N/ does not become /f{N/ - because although the surface realisation has merged, speakers still know which phoneme is which. Similarly, in London, th-fronting once spread a way into the afrocaribbean th-stopped community, so that some [t] would become [f] while other [t] did not become [f] - now the influence is more the other way, so that in some communities are shifting [f] to [t] - but only where the [f] 'represents underlying' or 'is derived from previous' /T/. Apparently the phoneme-discriminating [t]>[f] change is now also happening for some black american speakers. One last diachronic /T/-fronting example: speakers in the west country traditionally voice /T/ to /D/, but when they then apply th-fronting, they apply the rules appropriate to the original phoneme, so a speaking may change their [DIN] tor [vIN] or [fIN] while keeping their [D{n] as [D{n], because in the older change initial /D/ doesn't front. Another, synchronic example is the allophonic change /D/>[z], which applies even for many speakers with /D/>[v] in all other locations: so even people who say 'that' as [v{?], identical with 'vat', may still say "who's that?" as [huz.z{?] - while pronouncing "whose vat?" as [huz.v{?].
A higher class example of this is the epenthetic [n] inserted into the indefinite article before vowels - because many speakers also insert it before [h] where the [h] is itself a recent insertion due to spelling. [These speakers are now a minority, but they're still around].

Of course, this sort of phoneme-distinguishing change tends to be most notable where you have multiple dialects in contact, and where many speakers speak very distinct registers (so, eg, they can apply phonemic knowledge from one register into another register where a distinction has been neutralised) (though this is not necessary: dialectical speakers may simply have passive knowledge of phoneme distinctions in other dialects (this is probably what is happening with the writer-rider split in America), or they may import changes from another dialect on a lexical basis. Or it can be a mixture of all of the above - eg in glaswegian lower-class th-fronting both knowledge of higher registers and lexical import are clearly factors.

But then: that's exactly the sort of situation we probably have with vulgar latin. So I don't think it's any more problematic to have plebs say [u:] but 'know' that it's 'meant' to be [u~] or [um], and discriminate accordingly when it comes to vowel changes, than it is to have plebs say [f] or [h] or [t] or [D] but 'know' that it's 'meant' to be [T], and discriminate accordingly.

(((And then there are the cases that could be written out in terms of phones but where it's a pointless hassle to do it. For instance, there's no point saying that a change in English affects [p_h] and post-sibilant [p] or voiceless , plus [p_}] and [p?], when you could just say it affects /p/...)))

[Salmoneus]

Short version: 'sound changes affect surface realisations' probably belongs in the same neogrammarian museum as 'sound changes apply universally to all words where the conditions are shared'. And not coincidentally, since the two assumptions are closely linked.

[linguoboy]

Intervocalic [d] is lost in Gallo-romance and Ibero-Romance.

There are numerous exceptions in Ibero-Romance (e.g. CRUDU(M) > Sp. crudo, PEDE(M) > Astur. piet). In Catalan, the resulting fricative was often vocalised rather than simply being lost (e.g. PEDE(M) > peu).

[Travis B.]

To give an example: the raising of /aI/ to /VI/ in some north american dialects has taken place AFTER the neutralisation of the intervocalic T-D contrast that conditions that change! On the surface level, there's just a flap, but speakers are able to recognise the underlying phoneme and raise or not raise the preceding vowel accordingly.

This is slightly off-topic, but it may have been somewhat different. I am used to there still being a remaining preceding vowel length difference between unstressed intervocalic /t/ and /d/, indicating that the two phonemes are not truly neutralized, even when unstressed intervocalic /t/ is fully voiced. One of the changes splitting /aɪ̯/ from /əɪ̯/ could have been simply that all short /aɪ̯/ not followed by a morpheme boundary became /əɪ̯/, which would make all /aɪ̯/ before unstressed intervocalic /t/ become /əɪ̯/ without need for any remembering what phonemes are present after seemingly neutralizing sound changes.

In the other examples you provided, one thing that could likely be at work is the influence of other unchanged or differently-changed varieties to help remember what a given phone was before a historical sound change or to introduce different outcomes of sound change to replace historical sound changes. This could include influence of other registers existing in parallel to a given register, which would help keep memory of what was present before alive.

[Ars Lande]

NE: I just realised that all of the above does not actually address the question of the quantity of final -um in CL (i.e. was it something like [ũː]). I don't think it actually has any evidence to provide here. I cannot for the life of me think of any forms in the inherited Romance lexicon which show a reflex of an unambiguously long post-tonic auslaut ū, so claiming that /o/ is an unexpected reflex is to my mind jumping the gun somewhat.

You're absolutely right, of course. I was posting late at night, and after checking could not find any such example.

I thought the [uː] > Portuguese/Spanish/Italian thing was meant for stressed syllables only, and that unstressed vowels are generally more drastic in the mergers. Though maybe I'm thinking too much about Spanish, where all of unstressed [uː ʊ oː ɔ] > [o], with a small number of exceptions.

No, no that's perfectly correct, it applies to stressed syllables.

[Nortaneous]

To give an example: the raising of /aI/ to /VI/ in some north american dialects has taken place AFTER the neutralisation of the intervocalic T-D contrast that conditions that change! On the surface level, there's just a flap, but speakers are able to recognise the underlying phoneme and raise or not raise the preceding vowel accordingly.

This is slightly off-topic, but it may have been somewhat different. I am used to there still being a remaining preceding vowel length difference between unstressed intervocalic /t/ and /d/, indicating that the two phonemes are not truly neutralized, even when unstressed intervocalic /t/ is fully voiced.

I'm not.

[Travis B.]

To give an example: the raising of /aI/ to /VI/ in some north american dialects has taken place AFTER the neutralisation of the intervocalic T-D contrast that conditions that change! On the surface level, there's just a flap, but speakers are able to recognise the underlying phoneme and raise or not raise the preceding vowel accordingly.

This is slightly off-topic, but it may have been somewhat different. I am used to there still being a remaining preceding vowel length difference between unstressed intervocalic /t/ and /d/, indicating that the two phonemes are not truly neutralized, even when unstressed intervocalic /t/ is fully voiced.

I'm not.

I was not necessarily saying that there is synchronically a vowel length distinction in all present dialects that split /ai̯/ into /aɪ̯/ and /əɪ̯/, just that such may have existed at the time the split originally occurred, and that it would explain away any need to remember historical forms without surface realizations.

Of course this would be harder to explain if we could find dialects that both have this split and preserve historical vowel length, i.e. have never developed allophonic vowel length in the first place. Then we would have to go back to what Sal was saying.

[Basilius]

To give an example: the raising of /aI/ to /VI/ in some north american dialects has taken place AFTER the neutralisation of the intervocalic T-D contrast that conditions that change! On the surface level, there's just a flap, but speakers are able to recognise the underlying phoneme and raise or not raise the preceding vowel accordingly.

This is slightly off-topic, but it may have been somewhat different. I am used to there still being a remaining preceding vowel length difference between unstressed intervocalic /t/ and /d/, indicating that the two phonemes are not truly neutralized, even when unstressed intervocalic /t/ is fully voiced.

I'm not.

Nortaneous, how do you analyze this case then?

Not to hijack Ars Lande's thread, I started a new one.

[AnTeallach]

One last diachronic /T/-fronting example: speakers in the west country traditionally voice /T/ to /D/, but when they then apply th-fronting, they apply the rules appropriate to the original phoneme, so a speaking may change their [DIN] tor [vIN] or [fIN] while keeping their [D{n] as [D{n], because in the older change initial /D/ doesn't front.

Do you have a citation for this? I'd have thought most West Country speakers with th-fronting would have lost initial fricative voicing, which is a very recessive feature. (I don't think I've ever heard a real live West Country speaker with initial fricative voicing, not that I've spent much time in its heartlands.)

Anyway, word initial /ð/ in English is a strange beast. For me (and I'm sure for many others) it tends to be more a weak plosive or affricate than a fricative, so I actually get something different if I voice initial /θ/.