Quick Go at a Northwest Caucasian Romlang

[sangi39]

Sound Changes

1) First, VL labialised velars merge with plain velars, e.g. aqua > /aka/. Final

2) Between vowels, in a two-consonant cluster where each sound is of the same MOA, the first sound assimilates with the second in POA, e.g. /nokte/ > /notte/.

3) Voiced plosives become fricatives, i.e. /b d dz dʒ g/ > /v ð z ʒ ɣ/, unless before another consonant. /ð/ later merged into /z/ and /s/ becomes /z/ between vowels.

4) Voiceless plosives become voiced after nasals

5) Nasals are lost when syllable final (or in a coda cluster before another consonant) and /s/ was similarly lost when syllable final

6) Plosives before /s/ reduce to /h/, lengthening the preceding vowel, unless that vowel was already long, being lost eventually.

7) Geminate voiceless plosives result in ejectives while geminates of other consonants simply shorten.

8) /i e/ before another vowel in hiatus became /j/.

9) Afterwards /t d s z k g x ɣ/ before /j i e/ became /tʃ dʒ ʃ ʒ c ɟ ç ʝ/ and eventually /ts dz ʃ ʒ tɕ dʑ ɕ ʑ/ with the following /j/ falling out. This places these sounds effectively between allophones and phonemes, appearing in complementary distribution with /t d s z k g x ɣ/ in some environments, but alternating with them in others.

10) Vowel changes:

Code: Select all

/i i: u u:/   > /ɨ i ʉ u/ > /ɨ i ɨ u/ > / ɨ ʲi  ɨ ʷu/ > / ɨ ʲɨ  ɨ ʷɨ/
/e e: o o:/   > /ɛ e ɔ o/ > /ɛ e ɔ o/ > /ʲɛ ʲe ʷo ʷɔ/ > /ʲa ʲə ʷa ʷə/
/a a:/        > /ɐ a/     > /ə a/     > /ə a/         > /ə a/
/ai au oi eu/ > /ɛ ɔ e o/ > /ɛ ɔ e o/ > /ʲɛ ʷɔ ʲe ʷo/ > /ʲa ʷa ʲə ʷə/

11) Following changes affecting consonants with secondary articulation:

Plain:

Code: Select all

/p  t  ts    tɕ  k / > /p  t  ts    tɕ  k / 
/b  d  dz    dʑ  g / > /b  d  dz    dʑ  g /
/p' t' ts'   tɕ' k'/ > /p' t' ts'   tɕ' k'/
/m  n              / > /m  n              /
/f     s   ʃ ɕ   x / > /f     s   ʃ ɕ   x /
/v     z   ʒ ʑ   ɣ / > /v     z   ʒ ʑ   ɣ /
/            j   w / > /            j   w /
/   r  l           / > /   r  l           /

Palatalised:

Code: Select all

/pʲ  tʲ  tsʲ     tɕʲ  kʲ / > /p  tʃ  tʃ    tɕ  kʲ /
/bʲ  dʲ  dzʲ     dʑʲ  gʲ / > /b  dʒ  dʒ    dʑ  gʲ /
/p'ʲ t'ʲ ts'ʲ    tɕ'ʲ k'ʲ/ > /p' tʃ' tʃ'   tɕ' k'ʲ/
/mʲ  nʲ                  / > /m  nʲ               /
/fʲ      sʲ   ʃʲ ɕʲ   xʲ / > /f      ʃ   ɕ ɕ   xʲ /
/vʲ      zʲ   ʒʲ ʑʲ   ɣʲ / > /v      ʒ   ʑ ʑ   ɣʲ /
/             jʲ      wʲ / > /             j   ɥ  /
/     rʲ lʲ              / > /   ʒ   j            /

Labialised:

Code: Select all

/pʷ  tʷ  tsʷ     tɕʷ  kʷ / > /p  tsʷ  tsʷ     tʃʷ  kʷ /
/bʷ  dʷ  dzʷ     dʑʷ  gʷ / > /b  dzʷ  dzʷ     dʒʷ  gʷ /
/p'ʷ t'ʷ ts'ʷ    tɕ'ʷ k'ʷ/ > /p' ts'ʷ ts'ʷ    tʃ'ʷ k'ʷ/
/mʷ  nʷ                  / > /m  nʷ                   /
/fʷ      sʷ   ʃʷ ɕʷ   xʷ / > /f       sʷ   ʃʷ ʃʷ   xʷ /
/vʷ      zʷ   ʒʷ ʑʷ   ɣʷ / > /w       zʷ   ʒʷ ʒʷ   ɣʷ /
/             jʷ      wʷ / > /                ɥ    w  /
/    rʷ  lʷ              / > /   w    w               /

Overall Phoneme Inventory

Code: Select all

                       Labial Alveolar     Post-Alveolar          Velar
                              Plain Labial Plain Palatal. Labial. Plain. Palatal. Labial
            Voiceless  p      t                                   k      kʲ       kʷ
Stop        Voiced     b      d                                   g      gʲ       gʷ
            Ejective   p'     t'                                  k'     k'ʲ      k'ʷ
            Voiceless         ts    tsʷ    tʃ    tɕ       tʃʷ                     
Affricate   Voiced            dz    dzʷ    dʒ    dʑ       dʒʷ                     
            Ejective          ts'   ts'ʷ   tʃ'   tɕ'      tʃ'ʷ                    
Fricative   Voiceless  f      s     sʷ     ʃ     ɕ        ʃʷ      x      xʲ       xʷ
            Voiced     v      z     zʷ     ʒ     ʑ        ʒʷ      ɣ      ɣʲ       ɣʷ
Nasal                  m      n     nʷ           nʲ
Approximant                   l                  j        ɥ                       w
Trill                         r

High: ɨ
Mid: ə
Low: a

Examples:

OMNIS > wanʲɨ - OMNEM > wanʲa
AQUA > aka - AQUAM > aka
EQUUS > jakʷɨ - EQUUM > jakʷɨ
GLŌRIA > gwəʒa~gʷəʒa - GLŌRIA > gwəʒa~gʷəʒa
CAESAR > kʲazar - CAESAREM > kʲazaʒa

That's all I have at the moment, and it's just a quick experiment in sound change really but any comments would be much appreciated nonetheless

[kanejam]

This all seems pretty good up to step nine, where this:

/t d s z k g x ɣ/ before /j i e/ became /tʃ dʒ ʃ ʒ c ɟ ç ʝ/ and eventually /ts dz ʃ ʒ tɕ dʑ ɕ ʑ/

would be better expressed as t, d > ts, dz. This avoids the awkward change tʃ, dʒ > ts, dz and in fact the change t, d > ts, dz is more common than t, d < tʃ dʒ. You can even have k and g just go straight to the affricates. This then reflects nicely the palatalisation in the East Romance languages.

Then when we hit the vowels it gets a little weird. I'm only familiar with a few Caucasian languages so I could be completely wrong about this, but I think that most of the languages with no front-back distinction in their phonemes have surface allophones of much more normal vowels and are made that way by the secondary articulations on the consonants. What you have seems to be the other way around, with the vowels giving their secondary articulation to the consonants and losing their distinctiveness, which might be plausible but, I think, unlikely.

Also your vowels seem to be slightly odd and not quite follow the generally reconstructed patterns of Vulgar Latin. Some of your consonantal allophones seem a bit odd as well; why not keep the iotacised dental consonants rather than making them postalveolar affricates? It would be interesting to see if you could develop some kind of laryngealisation somewhere.

EQUUS > jakʷɨ - EQUUM > jakʷɨ
GLŌRIA > gwəʒa~gʷəʒa - GLŌRIA > gwəʒa~gʷəʒa
CAESAR > kʲazar - CAESAREM > kʲazaʒa

Shouldn't these be jakɨ, gʷəʒə and tɕazar, as per your sound rules? Unless I missed some vowel lengthening rule. I don't think you mentioned what happened to <ae> and <oe> but it would be very weird if it didn't follow the rest of the romance languages. You can however do what you like with <au>.

Anyway, it seems pretty good and I like how it does get a bit weird. All in all, very good? But of course now I want to see more! Maybe the orthography and some morphology.

[Click]

This all seems pretty good up to step nine, where this:

/t d s z k g x ɣ/ before /j i e/ became /tʃ dʒ ʃ ʒ c ɟ ç ʝ/ and eventually /ts dz ʃ ʒ tɕ dʑ ɕ ʑ/

would be better expressed as t, d > ts, dz. This avoids the awkward change tʃ, dʒ > ts, dz and in fact the change t, d > ts, dz is more common than t, d < tʃ dʒ.

[tʃ dʒ] → [ts dz] is not awkward, except maybe for being an additional sound change to write down. If I recall correctly, it is plausible and attested in several natural languages.

Then when we hit the vowels it gets a little weird. I'm only familiar with a few Caucasian languages so I could be completely wrong about this, but I think that most of the languages with no front-back distinction in their phonemes have surface allophones of much more normal vowels and are made that way by the secondary articulations on the consonants. What you have seems to be the other way around, with the vowels giving their secondary articulation to the consonants and losing their distinctiveness, which might be plausible but, I think, unlikely.

The secondary articulations on consonants, if systematic and not only on consonants with a same PoA, usually arise by vowels giving their secondary articulation to the consonants.

[sangi39]

Shouldn't these be jakɨ, gʷəʒə and tɕazar, as per your sound rules?

You're right on all three accounts I must have missed my own sound changes. As I said, it was just a quick thing. As for this:

I don't think you mentioned what happened to <ae> and <oe> but it would be very weird if it didn't follow the rest of the romance languages. You can however do what you like with <au>.

... I did: /ai au oi eu/ > /ɛ ɔ e o/ > /ɛ ɔ e o/ > /ʲɛ ʷɔ ʲe ʷo/ > /ʲa ʷa ʲə ʷə/

would be better expressed as t, d > ts, dz. This avoids the awkward change tʃ, dʒ > ts, dz and in fact the change t, d > ts, dz is more common than t, d < tʃ dʒ. You can even have k and g just go straight to the affricates. This then reflects nicely the palatalisation in the East Romance languages.

Could do. Either way, the result is the same Click has also replied to this query.

Also your vowels seem to be slightly odd and not quite follow the generally reconstructed patterns of Vulgar Latin

True, but Vulgar Latin isn't a single language. The Vulgar Latin from which Sardinian derives, for example, doesn't seem to have developed a quality difference between Classical long and short vowels while in Romanian's antecedent the quality difference seems to have developed only in front vowels. It's entirely possibly, then, that different forms of Vulgar Latin could have had different vowel systems from the usually reconstructed /a ɛ e ɪ i ɔ o ʊ u/.

Vulgar Latin, on the other hand, is the actual speech of the common people during the late Roman Empire. As a result, it is not simply theoretical but actually attested (if thinly), and is not unitary, with differences over both time and space. Hence, it is possible to speak of, for example, the loss of initial /j/ in unstressed syllables in the Vulgar Latin of Cantabria (the area in northern Spain that gave birth to modern Spanish), while it is inaccurate to speak of a similar change in the "Proto-Romance of Cantabria".

Some of your consonantal allophones seem a bit odd as well; why not keep the iotacised dental consonants rather than making them postalveolar affricates?

None of the Northwest Caucasian languages have palatalised dental consonants, so, in this "experiment", neither does this particular Romance language.

Then when we hit the vowels it gets a little weird. I'm only familiar with a few Caucasian languages so I could be completely wrong about this, but I think that most of the languages with no front-back distinction in their phonemes have surface allophones of much more normal vowels and are made that way by the secondary articulations on the consonants. What you have seems to be the other way around, with the vowels giving their secondary articulation to the consonants and losing their distinctiveness, which might be plausible but, I think, unlikely.

I believe Click's last comment answered this comment, as has Wikipedia's entry on vertical vowel systems. While vowels in vertical vowel systems do have allophones reflecting more "usual" vowels, this appears to happen after the reassignment of rounding/frontness to adjacent consonants.

[atman]

Nice idea! It would be especially interesting to see what you come up with for morphology.

After all, I'm not in a very different position myself with Atlántika (Basque-ified Ancient Greek, among other influences).

[kanejam]

Okay I stand thoroughly corrected then! In that case ignore my qualms about the vowel changes. So for example would /gʷəʒə/ have /goʒa/ as a major allophone?

And I'm excited about morphology!

[Valkura]

And I'm excited about morphology!

Something no cool person has ever said.

Everyone knows that the young and hip really only care about syntax.

[sangi39]

Okay I stand thoroughly corrected then! In that case ignore my qualms about the vowel changes. So for example would /gʷəʒə/ have /goʒa/ as a major allophone?

And I'm excited about morphology!

I'd say something like [goʒa], yeah

I'm still looking at vocabulary right now, seeing, for example, whether I should delete word-initial and/or word-final vowels (and, further, whether such deletion should relate to stress). At the moment, however, it looks like the nominative and accusative would fall together in the singular, while the plural nominative would lead to different secondary articulation on word-final consonants, which I guess would likely be undone by means of analogy.

I'll have a quick go later on today and hopefully post something tomorrow

[Click]

And I'm excited about morphology!

Something no cool person has ever said.

Everyone knows that the young and hip really only care about syntax.

+1'd only because syntax is currently my main field of interest.

[Ear of the Sphinx]

Nice 'lang.

Why did not palatalized labials > labialized postalveolars / alveolar affricates?

[Omzinesý]

Which Caucasian languages have you studied for your lang, or do you have some paper of Caucasian sprachbund?

[sangi39]

Nice 'lang.

Why did not palatalized labials > labialized postalveolars / alveolar affricates?

I have been giving that possibility some thought since I started working on this, but for the moment, since I'm still undecided on that, I left them becoming /p/. Really, given the indecision, I should have left them at palatalised labials and the explained why

Which Caucasian languages have you studied for your lang, or do you have some paper of Caucasian sprachbund?

At the moment, I've literally looked at the phoneme inventories of, as the title suggests, Northwest Caucasian languages, so predominantly Ubykh, Abkhaz, Abaza and Kabardian. It's not something I'm working on in any serious capacity, I'm just trying to see if this sort of thing can be done well, having seen a couple of attempts at Caucasian Romlangs before that were generally lacking in consistent sound changes and seemed to "force" Classical Latin into a Caucasian form.

[kanejam]

I'm still looking at vocabulary right now, seeing, for example, whether I should delete word-initial and/or word-final vowels (and, further, whether such deletion should relate to stress).

Thinking about it, I like how word initial vowels become mostly word initial semivowels. However I realise with your current sound changes that no ejective consonants pop up word initially, so maybe some deletion based on stress would be good. Of course a real Caucasian romlang would probably borrow plenty of words starting with ejectives, so this could be a null problem.

[sangi39]

I'm still looking at vocabulary right now, seeing, for example, whether I should delete word-initial and/or word-final vowels (and, further, whether such deletion should relate to stress).

Thinking about it, I like how word initial vowels become mostly word initial semivowels. However I realise with your current sound changes that no ejective consonants pop up word initially, so maybe some deletion based on stress would be good. Of course a real Caucasian romlang would probably borrow plenty of words starting with ejectives, so this could be a null problem.

That was the main reason for my thinking about initial vowel deletion. Final vowel deletion would lead to secondary articulation in word-final consonants, something that right now doesn't happen either.