U and i before the penultimate syllable vowel

meisenimverbis

Civis Illustris
Whenever I'm reading, it feels natural to stress in the third last syllable in situations I've seen grammars say it's stressed on the u or i... But shouldn't the u and i be thought as consonants in these situations?

Examples:

crebuĕrat - grammar: crebúerat - why not crébuerat?
muliĕres - grammar: mulíeres - why not múlieres?
potuĕrat - grammar: potúerat - ... pótuerat
debuĕrat - grammar: debúerat - ... débuerat
tenuĕrat - grammar: tenúerat - ... ténuerat
tribuĕrit - grammar: tribúerit - ... tríbuerit
promenuĕris - promenúeris - ... proménueris
insinuĕris - insinúeris - ... insínueris

(Instead of thinking syllable division as cre-bú-ĕ-rat, it would think it as cré-buĕ-rat.)

Here is another case. Shouldn't we consider the syllable with a u or i as not countable in terms of stress?

aperuit - grammar: apéruit - why not áperuit?
patefiĕrent - grammar: patefíerent - ... patéfierent

(Instead of a-pé-rŭ-it, we would have á-pĕ-ruit.)

Here on the contrary, shouldn't we think it not stressable?

fuĕrit - grammar: fúerit - why not fuérit?
fiĕri - grammar: fíeri - ... fiéri
puĕris - grammar: púeris - ... puéris

(Instead of fú-ĕ-rit, we would have fué-rit.)

It feels more likely to do the stress as though the u and i were part of the preceding consonant, in these cases... And it seems to make sense.

(As I was writing the examples, it came to mind that it would work like the qu-, the u or i would make part of the consonant, and not be treated separately...)

Does this make any sense? Has any of you ever faced this tendency when reading?
 

Quasus

Civis Illustris
> But shouldn't the u and i be thought as consonants in these situations?

Why, they are vowels and they take part in syllabification like all the rest vowels. If we stuck to good old ijuv, that would be more obvious.

Actually, things like that happened in Vulgar Latin, like muliere(m) > muljere, but I don't remember the details.
 

meisenimverbis

Civis Illustris
Actually, things like that happened in Vulgar Latin, like muliere(m) > muljere, but I don't remember the details.

I was thinking about this... Yes it makes sense that it does, especially on the way to the Romance languages, late vulgar.
 

Pacifica

grammaticissima
Staff member
There are even instances of such consonantization in classical Latin, at least with i — I can't remember if I've seen any example with u but it wouldn't surprise me.

Vergil, for instance, has this line:

parietibusque premunt artis et quattuor addunt

Which only scans if you read parietibus as par-je-ti-bus.

However, this is the exception rather than the rule.
 

Pacifica

grammaticissima
Staff member
Poetry...
Maybe it didn't happen only in poetry, but poetry is the only place where we can see it happen (thanks to the meter). In prose, there's no way for you to know how the author would have divided these words.
 

Pacifica

grammaticissima
Staff member
However, I'm not sure the i or u is very likely to be consonantized if it would normally have received the stress.

There are several examples of parjetibus, but the stress would be on the e anyway, even in pa-ri-e-ti-bus. I'm not sure a change from pa-RI-e-te to PAR-je-te is equally likely. Now, you never know how people can distort their own language.
 

meisenimverbis

Civis Illustris
However, I'm not sure the i or u is very likely to be consonantized if it would normally have received the stress.

There are several examples of parjetibus, but the stress would be on the e anyway, even in pa-ri-e-ti-bus. I'm not sure a change from pa-RI-e-te to PAR-je-te is equally likely. Now, you never know how people can distort their own language.

(Pariete lost the i when it passed to Portuguese, we have parede. Our natural stress though is on the second last syllable, unlike in Latin)

Doesn't 'pa-rie-te make more sense than 'par-ie-te? Like not because of the ie (r-ie), but because of the ri (ri-e), as though the j were joined with the consonant...
 

Iáson

Cívis Illústris
Doesn't 'pa-rie-te make more sense than 'par-ie-te? Like not because of the ie (r-ie), but because of the ri (ri-e), as though the j were joined with the consonant...
Well, Vergil clearly didn't pronounce it as such, because in that case it wouldn't scan.
 

Bitmap

Civis Illustris
There are even instances of such consonantization in classical Latin, at least with i — I can't remember if I've seen any example with u but it wouldn't surprise me.

Vergil, for instance, has this line:

parietibusque premunt artis et quattuor addunt

Which only scans if you read parietibus as par-je-ti-bus.

However, this is the exception rather than the rule.

I remember when I first learnt about that rule (that i can also be read as /j/) -- I thought it must be a pretty regular thing to do ... but then I was surprised how rarely this could be found in poetry – and it's something poets don't just do freely (unlike things like the varying length in the last i in "tibi"), but most of the time only as an act of metrical necessity.

It can even work the other way round, actually: Semi vowels like /w/ can become /u/, e.g. volvisse -> voluisse, as in these lines from tris. 2:

Mirer in hoc igitur tantarum pondere rerum
te numquam nostros evolvisse iocos?
 

Anbrutal Russicus

Active Member
There is a short yet comprehensive answer to this: the Latin syllable structure maximises the onset and minimises the coda. In other words the optimal syllabification of a word is one where as many consonants as possible begin syllables, and as few as possible close them. If you compare the two inputs /pa.ri.e.tem/ and /par.je.tem/, you'll find that the former is the syllabification that better satisfies these conditions.

The stress that the OP feels is more natural was coincidentally what passed into all the Romance languages (Sp. mujér, Sic. muggh(i)ɛ́ri, It. figli(u)ɔ́lo), which is explained by a change in preferred syllable structure due to many more closed syllables arising from syncope. Hey, who just whispered "chicken and egg"?! Notice how syllabification was computed before stress, so even a normally-stressed vowel could drop/become a consonant. Also, the vowel obligatorily lengthened (or else was borrowed from the Nominative, cf. It. paréte, abéte < -iēs, iĕtem), but the poets generally continued to follow the classical syllabification rules. The resulting /mu.li.ɛ̄.rem/ (which should really be spelt with AE) etc is to be found widely in Late and Medieval Latin, and for those words that classically have three short syllables in a row, this was the only non-artificial solution to using them in a hexameter, so you get self-selection.

For the fate of the perfect forms cf. tenuī > *ten.nvī > It. tenni - here the problem of how to syllabify that cluster was solved by doubling, same as in faccia, occhio etc. Come to think of it, the /li+open vowel/ cases also have doubling, but not the close-vowel ones, and some reflexes of mulier (eg. Spanish) point to /mujjere/.
 
Last edited:

Serenus

Civis Illustris
There is a short yet comprehensive answer to this: the Latin syllable structure maximises the onset and minimises the coda. In other words the optimal syllabification of a word is one where as many consonants as possible begin syllables, and as few as possible close them. If you compare the two inputs /pa.ri.e.tem/ and /par.je.tem/, you'll find that the former is the syllabification that better satisfies these conditions.

The stress that the OP feels is more natural was coincidentally what passed into all the Romance languages (Sp. mujér, Sic. muggh(i)ɛ́ri, It. figli(u)ɔ́lo), which is explained by a change in preferred syllable structure due to many more closed syllables arising from syncope. Hey, who just whispered "chicken and egg"?! Notice how syllabification was computed before stress, so even a normally-stressed vowel could drop/become a consonant. Also, the vowel obligatorily lengthened (or else was borrowed from the Nominative, cf. It. paréte, abéte < -iēs, iĕtem), but the poets generally continued to follow the classical syllabification rules. The resulting /mu.li.ɛ̄.rem/ (which should really be spelt with AE) etc is to be found widely in Late and Medieval Latin, and for those words that classically have three short syllables in a row, this was the only non-artificial solution to using them in a hexameter, so you get self-selection.
Have you read of this somewhere in particular, some work where I could read more of it? It seems pretty interesting and at least my first impression right now is it feels more intuitively correct than what explanations I've heard before of this; I have to meditate and digest it more... Just wanted to also comment on this:
For the fate of the perfect forms cf. tenuī > *ten.nvī > It. tenni - here the problem of how to syllabify that cluster was solved by doubling, same as in faccia, occhio etc. Come to think of it, the /li+open vowel/ cases also have doubling, but not the close-vowel ones, and some reflexes of mulier (eg. Spanish) point to /mujjere/.
Ancient Latin /jj/ evolves into Old Spanish y /j/ though, e.g. plagia > playa 'beach', habeam > haya, vādam > *vād-iam > vaya. It happens even before a back vowel, maiōrem > mayor, radius > rayo (in contrast to its usual word-initial outcome as /ʒ/ before a back vowel, iocus > juego, iūdicāre > judgar (now juzgar), iunctus > junto).

So I don't think O.Sp. muger/muier *[muˈʒeɾ] points to /mujjére/. But I do like this idea you mention that it may point to /mulljere/... It's easy to see that possibility considering the regular fortition of Late Latin /lj/ > Old Spanish /ʒ/.
 

Anbrutal Russicus

Active Member
Have you read of this somewhere in particular, some work where I could read more of it? It seems pretty interesting and at least my first impression right now is it feels more intuitively correct than what explanations I've heard before of this; I have to meditate and digest it more...
Which part in particular? I think most of this follows automatically from understanding the Latin syllable structure, its synchronic variation and the changes to it over time, which are also observable directly in the syllabification of stop-liquid clusters, which incidentally also display change of stress showing it was post-lexical.

Any way, I don't have any particular reference. The part about the vowels of parete etc could also be interpreted as /pa.ri.je.te/ with j-deletion and contraction of [eɛ~ɪɛ > e:], instead of the usual syncope of the vowel preceding /j/ as in -ius, if we take /ī/ to be underlyingly /ĭj/, which it probably was in the classical times.

Even more likely is that it started in the nominative in the dialects that merged iV and eV, so the same vowels /ēē/ contracted, just as happened everywhere with /īī/ - which still makes it an extension from the nominative! :-] In fact this seems to be already attested by Varro.
So I don't think O.Sp. muger/muier *[muˈʒeɾ] points to /mujjére/. But I do like this idea you mention that it may point to /mulljere/... It's easy to see that possibility considering the regular fortition of Late Latin /lj/ > Old Spanish /ʒ/.
I remembered reading about a considerable variation in its reflexes so /jj/ looked plausible - but now I see that in all the languages the reflex of mulier shows the same basic development as in fīlius, so yeah they must all go back to the same form in /llj/.
 
Last edited:

Anbrutal Russicus

Active Member
One more thing I didn't address:
(Instead of thinking syllable division as cre-bú-ĕ-rat, it would think it as cré-buĕ-rat.)
[...]
aperuit - grammar: apéruit - why not áperuit?
patefiĕrent - grammar: patefíerent - ... patéfierent
[...]
fuĕrit - grammar: fúerit - why not fuérit?
fiĕri - grammar: fíeri - ... fiéri
puĕris - grammar: púeris - ... puéris
This is because these aren't possible syllable onsets in Latin - no syllable in that language is allowed to begin in /bw/ or /fj/; nor in /wb/ or /jf/; in fact a glide consonant can never be part of a cluster, with the exception of /sw/, but only when a long vowel follows, e.g. suāvis (treating it as a vowel is also allowed, cf. Italian soave).
(As I was writing the examples, it came to mind that it would work like the qu-, the u or i would make part of the consonant, and not be treated separately...)
You've hit on the right thing there - /kw/ (and /gw/ which only occurs after /n/) is the one major exception to the above rule, and this is one reason it's been traditionally treated as one single coarticulated consonant /kʷ/.

The Brazilian Portuguese syllable structure is of course different and does allow those clusters to begin syllables. European Portuguese allows much more combinations, and Georgian is totally chill about things like /mt͡svrtʰ/.
 

Serenus

Civis Illustris
(treating it as a vowel is also allowed, cf. Italian soave).
So interesting; I hadn't noticed that. Old French "soef" also suggests a Late Latin / pre-French *[soˈa:βe], then undergoing *a: > *e: raising and final devoicing, as in caput > analogized capum > chief.

You've hit on the right thing there - /kw/ (and /gw/ which only occurs after /n/) is the one major exception to the above rule, and this is one reason it's been traditionally treated as one single coarticulated consonant /kʷ/.
By the way, do you happen to know of a good reason why people don't classify the /sw/ of suāvis/suādeō/suēscō as /sʷ/ as well? Same goes for the /dw/ of Archaic Latin duenos > bonus. I notice Wikipedia says "The sequences *ḱw, *ǵw, *ǵʰw develop identically to *kʷ, *gʷ, *gʷʰ", giving *h₁éḱwos as an example.
 

Anbrutal Russicus

Active Member
So interesting; I hadn't noticed that. Old French "soef" also suggests a Late Latin / pre-French *[soˈa:βe], then undergoing *a: > *e: raising and final devoicing, as in caput > analogized capum > chief.
Very nice, I was wondering what that French form was all about, but don't know enough of the latter's historical phonology.
By the way, do you happen to know of a good reason why people don't classify the /sw/ of suāvis/suādeō/suēscō as /sʷ/ as well? Same goes for the /dw/ of Archaic Latin duenos > bonus. I notice Wikipedia says "The sequences *ḱw, *ǵw, *ǵʰw develop identically to *kʷ, *gʷ, *gʷʰ", giving *h₁éḱwos as an example.
I think this depends on the particular people. For instance, De Vaan 2008: 550 says "there's no hard evidence for *sw- in Italic" - tho I personally can't comment on this. There's also the form sāvium which either spells /sʷ/ or must go back to it, reinterpreted by sʷ-less speakers as starting in simple /s/. For /dw/ I've never read analyses either way, but certainly the shift itself is evidence that a lot of folks heard it as a single phoneme - otherwise I'd expect the same syllabic resolution as in quoniam. I'd like to find a good treatment of the glide situation in pre-Latin.
 
Last edited:
Top