(Conlangs) Q&A Thread - Quick questions go here [2010-2020]

[Porphyrogenitos]

I’m trying to reverse-engineer a situation where some suffixes give roots initial stress and others shift the stress to the right. Like say you have a root kawm, and the suffix -ta produces káwmta, but the suffix -su produces kawmsú. I know that if I initially have a regular stress assignment rule, and then apply various sound changes, I can produce this kind of situation, but my mind is frazzled cause I’m up at 1 in the morning - Can someone suggest a rule and sound change combination that could produce this?

[Dormouse559]

One of the (many, many) possibilities is to have the original pattern be penultimate stress by default, shifted to ultimate stress when the final syllable is heavy. In that case, -su might have originally been -suC. Then, final consonants would have been elided.

EDIT: This solution is limited to your example, but I'm sure it could be adapted to handle longer roots and suffixes.

[gestaltist]

One of the (many, many) possibilities is to have the original pattern be penultimate stress by default, shifted to ultimate stress when the final syllable is heavy. In that case, -su might have originally been -suC. Then, final consonants would have been elided.

EDIT: This solution is limited to your example, but I'm sure it could be adapted to handle longer roots and suffixes.

That would be my first suggestion, as well. Another possibility would be an earlier stage that had tone or pitch where the suffix had high tone that got reanalyzed as stress when tone was lost.

[Ahzoh]

I'm looking to do some extensive vowel coalescence but I'm not sure what vowels could result from these combinations:

/a/ + /e/
/a/ + /o/

/e/ + /a/
/e/ + /o/
/e/ + /u/

/i/ + /a/
/i/ + /e/
/i/ + /o/
/i/ + /u/

/o/ + /a/
/o/ + /e/
/o/ + /i/

/u/ + /a/
/u/ + /e/
/u/ + /i/
/u/ + /o/

[sangi39]

Are you looking for them to collapse down to vowels already in the language, or could new vowels arise from them (for example, if these vowels most often come together at morpheme boundaries rather than within morphemes then the results might be different)? Would you prefer that the results be short vowels, long vowels, or diphthongs?

[Ahzoh]

Are you looking for them to collapse down to vowels already in the language, or could new vowels arise from them (for example, if these vowels most often come together at morpheme boundaries rather than within morphemes then the results might be different)? Would you prefer that the results be short vowels, long vowels, or diphthongs?

Yea I'm looking to make them collapse down to single-length vowels already in the language that arise mostly across morpheme boundaries.
It's mostly a morphophonemic change.

[gestaltist]

I'm looking to do some extensive vowel coalescence but I'm not sure what vowels could result from these combinations:

/a/ + /e/
/a/ + /o/

/e/ + /a/
/e/ + /o/
/e/ + /u/

/i/ + /a/
/i/ + /e/
/i/ + /o/
/i/ + /u/

/o/ + /a/
/o/ + /e/
/o/ + /i/

/u/ + /a/
/u/ + /e/
/u/ + /i/
/u/ + /o/

Assuming you have /a e i o u/ and don't want length distinction, I would probably go with something like this:
a+e > ai > e
a+o > au > o
e+a > i@ > i~e
e+o > o
e+u > u
i+a > i@ > i~e
i+e > e
i+o > iw > y > i
i+u > iw > y > i
o+a > o@ > o
o+e > we > (v)e
o+i > oj > ø > e
u+a > u@ > u~o
u+e > ø > e
u+i > y > i
u+o > wo > (v)o

[Ahzoh]

That looks great, thanks!

Is there some way for /eo/ and /oe/ to become /a/, possibly from /ə/?

[LinguistCat]

Is there some way for /eo/ and /oe/ to become /a/, possibly from /ə/?

I mean, that's possible. You might not even need to change what the suggested combos with @ in them would become either, depending what you're going for.

My questions
If I have a very limited consonant inventory like proposed for Pre-Old Japanese or Proto Japonic (basically /p t s k n m w r j/ with a lot of allophones for /p t s k/ in particular),
A. what could I do with /p/ that wouldn't be following the path it took historically in Japanese?
B. If I went k > x > h conditionally (or something similar), what else could I do feeding into this or branching off?
C. What could nasal+unvoiced consonant become besides voiced (and/or prenasalized) consonants at the same point of articulation?

[DesEsseintes]

If I have a very limited consonant inventory like proposed for Pre-Old Japanese or Proto Japonic (basically /p t s k n m w r j/ with a lot of allophones for /p t s k/ in particular),
A. what could I do with /p/ that wouldn't be following the path it took historically in Japanese?
B. If I went k > x > h conditionally (or something similar), what else could I do feeding into this or branching off?
C. What could nasal+unvoiced consonant become besides voiced (and/or prenasalized) consonants at the same point of articulation?

A. & B. You could have p k > f x > hʷ h in the some conditions. The rounding on hʷ could then transfer onto the following vowel. Alternatively, have the rounding on hʷ metathesise to ʷh and affect the preceding vowel.
C. Geminate nasals mp nt nk → mm nn ŋŋ. If standard nasals m n debuccalise to w̃ j̃, mm nn ŋŋ could provide a new three-way plain nasal series through degemination.

Just the first things that pop to mind.

[Creyeditor]

I'm looking to do some extensive vowel coalescence but I'm not sure what vowels could result from these combinations:
[...]

Sorry for the late answer, but I did some reading on vowel hiatus resolution for other reasons, and I just want to share some universal tendencies. Vowel coalescence often takes the height of the lower of the two vowels. This is not true if a low vowel is involved, here we get mid vowels as a result as often as low vowels. Backness and frontness are equally probably to survive, but if you coalesce a rounded vowel with an unrounded one, the result is a tad more likely to also be rounded. Based on these, I propose the following. Of course you can pick and choose from the suggestions, you get. This is not meant as a correction or some kind of 'uhm, actually'. I decided to let order decide between low and mid vowels. So in general: the result is rounded, if one of the vowels is a round vowel (unless the result would be a low round vowel). The result is mid, if at least one of the vowels is a non-high vowel (unless the first vowel is a low vowel). The result is low only if the first vowel is a low vowel.


/a/ + /e/ = a
/a/ + /o/ = a

/e/ + /a/ = e
/e/ + /o/ = o
/e/ + /u/ = o

/i/ + /a/ = e
/i/ + /e/ = e
/i/ + /o/ = o
/i/ + /u/ = u

/o/ + /a/ = o
/o/ + /e/ = o
/o/ + /i/ = o

/u/ + /a/ = o
/u/ + /e/ = o
/u/ + /i/ = u
/u/ + /o/ = o

[sangi39]

If I have a very limited consonant inventory like proposed for Pre-Old Japanese or Proto Japonic (basically /p t s k n m w r j/ with a lot of allophones for /p t s k/ in particular),
A. what could I do with /p/ that wouldn't be following the path it took historically in Japanese?

/p/ could remain /p/, as it did in the Ryukyuan languages, IIRC, or it could become /w/ or /ʔ/. That's aside from the fact that it could become voiced, between vowels for example.

B. If I went k > x > h conditionally (or something similar), what else could I do feeding into this or branching off?

If /k/ became /h/ (even unconditionally) then you could have any instances of /g/ that might have arisen devoice, filling the gap.

C. What could nasal+unvoiced consonant become besides voiced (and/or prenasalized) consonants at the same point of articulation?

IIRC, prenasalised stops are likely to become either voiced stops or nasal, but if you had /mp nt nk/ clusters, you could steal a page out of West Norse and have them become geminate stops, i.e. /pp tt kk/, which could then go on to either shorten, become preaspirated, affect preceding vowel length, and a host of other things.

[Creyeditor]

My questions
If I have a very limited consonant inventory like proposed for Pre-Old Japanese or Proto Japonic (basically /p t s k n m w r j/ with a lot of allophones for /p t s k/ in particular),
A. what could I do with /p/ that wouldn't be following the path it took historically in Japanese?
B. If I went k > x > h conditionally (or something similar), what else could I do feeding into this or branching off?
C. What could nasal+unvoiced consonant become besides voiced (and/or prenasalized) consonants at the same point of articulation?

A. some ideas that come to mind: p>f, p>b,p> pf,p>pʰ, p>p̚, p>kʷ. All of these can of course be conditional. p>pʰ is the only change where I would suspect it might be more likely word-initially. p>p̚ is more probably word-final or syllable-final and p>kʷ is probably conditioned by vowel harmony. The others are mostly intervocalic, but of course these are not mutually exclusive. My suggestion. Try to get several sounds from *p by splitting it up a bit.
B. Intervocalic deletion of *k would look cool maybe. This then could lead to *g and * merging into one phoneme, if g devoiced word in all other contexts.
C. You could look at Austronesian languages, they have a wide variety of relexes. Deletion of one or the other is probably the most obvious solution (apart fro prenasalization), but you could also fuse the sounds, e.g. a nasal with the place of articulation of the unvoiced consonant. Another result of the fusion could be a voiceles nasal, maybe?

[Ælfwine]

Having not given up the "polysynthetic romlang" idea, I've been toying with the idea of a romlang heavily influenced by Abkhaz with a large consonant and small vowel inventory typical of such languages. Props to Isfendil who helped me establish some of the background of the language as spoken by expatriate Jews who fled Rome after it was sacked by Alaric, eventually ending up in the Caucus.

Whilst I am not the first one to do this idea, Sangi39 had in this thread, I don't want to particularly rip off of him or steal any of his ideas (without permission at least.) While the vowel derivations he has works, they are somewhat atypical of developments in vulgar latin — for example, it is unlikely that VL *ɪ and *ʊ would head off it completely opposite directions than in VL, nor do I know of any Romance language or dialect that treated the outcomes of a and ā differently. But I don't know.

Essentially what I want to do is get a similar result by more realistic means. However I am wondering what the peanut gallery might think is a believable path for this romlang to take, or if anyone has their own ideas they'd like to share? Or are Sangi's original ideas realistic enough to handwave?

[sangi39]

Having not given up the "polysynthetic romlang" idea, I've been toying with the idea of a romlang heavily influenced by Abkhaz with a large consonant and small vowel inventory typical of such languages. Props to Isfendil who helped me establish some of the background of the language as spoken by expatriate Jews who fled Rome after it was sacked by Alaric, eventually ending up in the Caucus.

Whilst I am not the first one to do this idea, Sangi39 had in this thread, I don't want to particularly rip off of him or steal any of his ideas (without permission at least.) While the vowel derivations he has works, they are somewhat atypical of developments in vulgar latin — for example, it is unlikely that VL *ɪ and *ʊ would head off it completely opposite directions than in VL, nor do I know of any Romance language or dialect that treated the outcomes of a and ā differently. But I don't know.

Essentially what I want to do is get a similar result by more realistic means. However I am wondering what the peanut gallery might think is a believable path for this romlang to take, or if anyone has their own ideas they'd like to share? Or are Sangi's original ideas realistic enough to handwave?

Eh, if you want to rip any of it off, please do. As I said in the thread, it was just an experiment to see whether it could be done with somewhat consistent sound changes.

As for the development of VL *ɪ, *ʊ and short *a, I didn't particularly see a reason why they couldn't develop differently. The short high vowels, for example, developed in at least three different ways in the extant Romance languages, so I didn't see any harm in having them change in some fourth way, i.e. centralisation, rather than a) lowering with some centralisation or b) remaining high. And while *a did merge with its long counterpart literally everywhere as far as I can tell as well, *a > [ɐ] isn't a hugely unreasonable change to occur, especially for a romlang situated outside of the Empire (although I suppose it would depend on timing). It's mostly there to get instances of plain [ə] in the phoneme inventory because all of the other sound changes resulting in [ə] also result in secondary articulation.

[Dormouse559]

I think the unusual changes are fine. Especially given this conlang is set in the Caucasus, where the languages do phonological BASE jumps from the mountain peaks. You can only get those Romance sprachbund effects if you're in contact with Romance speakers. But if you wanted to, maybe you could put off the Abkhaz-like sound changes until later. French and Italian vowels evolved somewhat similarly for a little while, until French started throwing in front-rounded vowels and phonemic length.

[Ælfwine]

Having not given up the "polysynthetic romlang" idea, I've been toying with the idea of a romlang heavily influenced by Abkhaz with a large consonant and small vowel inventory typical of such languages. Props to Isfendil who helped me establish some of the background of the language as spoken by expatriate Jews who fled Rome after it was sacked by Alaric, eventually ending up in the Caucus.

Whilst I am not the first one to do this idea, Sangi39 had in this thread, I don't want to particularly rip off of him or steal any of his ideas (without permission at least.) While the vowel derivations he has works, they are somewhat atypical of developments in vulgar latin — for example, it is unlikely that VL *ɪ and *ʊ would head off it completely opposite directions than in VL, nor do I know of any Romance language or dialect that treated the outcomes of a and ā differently. But I don't know.

Essentially what I want to do is get a similar result by more realistic means. However I am wondering what the peanut gallery might think is a believable path for this romlang to take, or if anyone has their own ideas they'd like to share? Or are Sangi's original ideas realistic enough to handwave?

Eh, if you want to rip any of it off, please do. As I said in the thread, it was just an experiment to see whether it could be done with somewhat consistent sound changes.

As for the development of VL *ɪ, *ʊ and short *a, I didn't particularly see a reason why they couldn't develop differently. The short high vowels, for example, developed in at least three different ways in the extant Romance languages, so I didn't see any harm in having them change in some fourth way, i.e. centralisation, rather than a) lowering with some centralisation or b) remaining high. And while *a did merge with its long counterpart literally everywhere as far as I can tell as well, *a > [ɐ] isn't a hugely unreasonable change to occur, especially for a romlang situated outside of the Empire (although I suppose it would depend on timing). It's mostly there to get instances of plain [ə] in the phoneme inventory because all of the other sound changes resulting in [ə] also result in secondary articulation.

The point about the vowels centralizing is fair, though I don't know how realistic it would be for my idea of expatriate Jews than a native population in the Caucus, but I assume that even the Jews probably had a different way of speaking Vulgar Latin than the rest of the populace.

As for /a/, I recall reading somewhere that that vowel was the first to merge in the Roman Empire, given its consistency.

I think the unusual changes are fine. Especially given this conlang is set in the Caucasus, where the languages do phonological BASE jumps from the mountain peaks. You can only get those Romance sprachbund effects if you're in contact with Romance speakers. But if you wanted to, maybe you could put off the Abkhaz-like sound changes until later. French and Italian vowels evolved somewhat similarly for a while, until French started throwing in front-rounded vowels and phonemic length.

That was my idea, as my romance speakers were migrants to the Caucus. (I wasn't sure how realistic it would be to have a surviving remnant of Latin speakers in an area which never natively spoken Latin, but possibly Greek)

Actually, Isf says that in a lot of areas Latin short vowels developed to schwa in Romanian, so perhaps Sangi's vowel developments have more plausibility than I though. I'm thinking that Romanian has a lot of potenital to give me ideas as it developed both phonemic schwa and an array of palatalized consonants.

[Vlürch]

A. what could I do with /p/ that wouldn't be following the path it took historically in Japanese?

VpV > VbV > VʙV > VbrV
pV > pV̥ > p (allowing word-initial pC clusters)
pp > pəp > pub
pt > t͡p > t̼
ps > p͡s > t͡sʷ
pk > k͡p > kʷ
pn > pən > pɯ̃n
pm > pəm > pũm
pw >pʷ > p
pr > r̥ʷ > ʃʷ
pj > pʲ > cʷ

Not necessarily the most realistic, at least all together, but well.

B. If I went k > x > h conditionally (or something similar), what else could I do feeding into this or branching off?

I don't really understand the question since if it's unconditional, it's... well... unconditional.

C. What could nasal+unvoiced consonant become besides voiced (and/or prenasalized) consonants at the same point of articulation?

Maybe something like:
Vmp Vnt Vŋk > Ṽɸ Ṽθ Ṽx > Vv Vð Vɣ
...but I don't know if that's realistic enough, at least without additional intermediate steps.

[Omzinesý]

Does it happen somewhere that if a language does not allow word-initial consonant clusters, s+t clusters undergo metathesis and become geminate ts ?

[Sumelic]

Does it happen somewhere that if a language does not allow word-initial consonant clusters, s+t clusters undergo metathesis and become geminate ts ?

I've never heard of that. The opposite type of metathesis, where an affricate changes to a cluster of sibilant+ stop, seems to have occured (unconditionally, not word-initially) in some Slavic and Greek languages.