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

[Zekoslav]

The only thing that seems unrealistic to me is that some sound changes seem to reverse, for example [pʰʊːʔɴ̩̊] > [pʰɯˀᶰɢə] > [pʰʊʔɴ̩̊]. Otherwise it looks fine. As far as similarity goes, even if this word remains relatively similar, others may become unrecognizable, depending on the exact sound changes.

[yangfiretiger121]

Is [e, i, o, ø, ɶ, ɔ, ɯ] or [a~æ, ɑ, e, o, ø, u, y] the more natural inventory?

[Ælfwine]

Is [e, i, o, ø, ɶ, ɔ, ɯ] or [a~æ, ɑ, e, o, ø, u, y] the more natural inventory?

I would say the latter, if only because all spaces are equally distributed by phonemes.

[Nortaneous]

Is [e, i, o, ø, ɶ, ɔ, ɯ] or [a~æ, ɑ, e, o, ø, u, y] the more natural inventory?

Code: Select all

i   ɯ
e ø o
ɶ   ɔ

  y u
e ø o
æ ɑ

neither - almost all languages have /a/ and either /i/ or (in the case of VVSes) a high vowel unmarked for [+/-front]. /y/ necessarily implies /i/, and /ɶ/ is an artifact of the IPA trying to be logical, not a real phoneme that can actually happen

/a ɔ e ø o i u/ and /æ ɑ e ø o i y u/ are the closest inventories to those that are natural, and in the first one you could maybe have /ɯ/ instead of /u/ but I think that would lead to a shift of ɔ > o > u, ɯ > i or schwa or something, maybe also ø > y but if you only have one front rounded vowel it can just vary between mid and high (cf. Kentish y > e, Souletin Basque y~ø, Polish central unrounded ɨ~ə... are there phonetic studies of Hopi?)

[yangfiretiger121]

Is [e, i, o, ø, ɶ, ɔ, ɯ] or [a~æ, ɑ, e, o, ø, u, y] the more natural inventory?

Code: Select all

i   ɯ
e ø o
ɶ   ɔ

  y u
e ø o
æ ɑ

neither - almost all languages have /a/ and either /i/ or (in the case of VVSes) a high vowel unmarked for [+/-front]. /y/ necessarily implies /i/, and /ɶ/ is an artifact of the IPA trying to be logical, not a real phoneme that can actually happen

/a ɔ e ø o i u/ and /æ ɑ e ø o i y u/ are the closest inventories to those that are natural, and in the first one you could maybe have /ɯ/ instead of /u/ but I think that would lead to a shift of ɔ > o > u, ɯ > i or schwa or something, maybe also ø > y but if you only have one front rounded vowel it can just vary between mid and high (cf. Kentish y > e, Souletin Basque y~ø, Polish central unrounded ɨ~ə... are there phonetic studies of Hopi?)

Ah. Okay. However—regarding the second inventory, this Gaelic hybrid isolate wouldn't contrast the low front unrounded vowels. The low front vowel, /æ̞/, will be Romanized <æ/ä>, whereas the low back vowel, /ɑ/, will be Romanized <a>, if I use that one. Native linguists will have, simply, used the chosen an unambiguous transcription for the phone not wanting to confuse native speakers.

Is there a way to get [fˠ → ʍ], with [ʍ] being a true fricative?

[Nloki]

The only thing that seems unrealistic to me is that some sound changes seem to reverse, for example [pʰʊːʔɴ̩̊] > [pʰɯˀᶰɢə] > [pʰʊʔɴ̩̊]. Otherwise it looks fine. As far as similarity goes, even if this word remains relatively similar, others may become unrecognizable, depending on the exact sound changes.

Thank you Zekoslav. Besides this matter, I've come up with another doubt (I mean, for everyone):

•Does any natlang feature (a set of) preglottalized (voiceless) nasals? I don't know how did I come up these phonemes, but unexplainably I've fell in love with them. In fact, within both the Jehoeian and Qatsewan branches of the Jhekhu-Qatsaic language family there is a phonemic set of these consonants (/ˀm̥ ˀn̪̊ ˀɲ̊ ˀŋ̊/) throughout all the phonological evolution on each branch (with some variations during phonological history of course), despite their respective speaker communities being separated from each other by the mountainous ice cap to the north and a vast ocean to the south (diachronically speaking of my conworlding project).

[Omzinesý]

I'm wondering subject properties of chain verb constructions. In (1), come and give share the subject and subject marking is straightforward, but in (2), the subject of die is the object of hit. How do languages with serial verbs mark it or should it just be understood from the context? I understand different languages do it differently but what strategies there are?

(1)
[ X ] [ [ come ] [ give Y ] ]
'X brought Y.'

(2)
[ X hit ( Y ] die )
'X killed Y.'

[Nloki]

The only thing that seems unrealistic to me is that some sound changes seem to reverse, for example [pʰʊːʔɴ̩̊] > [pʰɯˀᶰɢə] > [pʰʊʔɴ̩̊]. Otherwise it looks fine. As far as similarity goes, even if this word remains relatively similar, others may become unrecognizable, depending on the exact sound changes.

Thank you Zekoslav. Besides this matter, I've come up with another doubt (I mean, for everyone):

•Does any natlang feature (a set of) preglottalized (voiceless) nasals? I don't know how did I come up these phonemes, but unexplainably I've fallen in love with them. In fact, within both the Jehoeian and Qatsewan branches of the Jhekhu-Qatsaic language family there is a phonemic set of these consonants (/ˀm̥ ˀn̪̊ ˀɲ̊ ˀŋ̊/) throughout all the phonological evolution on each branch (with some variations during phonological history of course), despite their respective speaker communities being separated from each other by the mountainous ice cap to the north and a vast ocean to the south (diachronically speaking of my conworlding project).

I'm wondering subject properties of chain verb constructions. In (1), come and give share the subject and subject marking is straightforward, but in (2), the subject of die is the object of hit. How do languages with serial verbs mark it or should it just be understood from the context? I understand different languages do it differently but what strategies there are?

(1)
[ X ] [ [ come ] [ give Y ] ]
'X brought Y.'

(2)
[ X hit ( Y ] die )
'X killed Y.'

Despite worthless half an hour research I couldn't really find anything on how natural languages deal with such issues on rather semantically "catenative" verb constructions. However, there could be some approaches to this matter independently from naturalism:

•If thinking on a moderately high degree of sythesis, I would rather have some sort of switch reference marking on the Y argument’s verb, in this case die.

•If your ideas are to make it isolating, even if slightly, use a causative. If barely isolating however, then split the construction into two clauses instead, linking them with "and" (which I'd never do for my grammatical taste's sake but... Well, that's up to you).

[Ahzoh]

Can a language have /ʝ ɣʷ/ without /j w/?

[LinguoFranco]

How are prenasalized consonants treated in between vowels? Are they considered separate or the same. For example, let's say you have /kamba/, would it be /ka.mba/ or /kam.ba/? I know that prenasalized consonants are treated as one phoneme at the beginning of the word instead of a syllabic nasal followed by a cosnont (/mba/ and not /m.ba/), but could/does the rule change under other circumstances?

[Zekoslav]

My understanding of prenasalised consonants is that /kamba/ would be syllabified as /ka.mba/. If for example your only consonant clusters were intervocallic /mb/, /nd/ and /ŋg/, then it would be feasible to analyse these as prenasalised consonants and your syllable structure as CV

[LinguistCat]

How are prenasalized consonants treated in between vowels? Are they considered separate or the same. For example, let's say you have /kamba/, would it be /ka.mba/ or /kam.ba/? I know that prenasalized consonants are treated as one phoneme at the beginning of the word instead of a syllabic nasal followed by a consonant (/mba/ and not /m.ba/), but could/does the rule change under other circumstances?

This is usually the difference between prenasalized consonants and homorganic nasal/consonant clusters. In all cases, prenasalized consonants are treated as phonemes, while nasal/consonant clusters are treated as, well, clusters. While I was researching prenasalized consonants for my own conlang project, someone even pointed me to evidence that, in languages with prenasalized consonants and gemination, geminated prenasalized consonants most often become nasal/consonant clusters.

[Nortaneous]

Is there a way to get [fˠ → ʍ], with [ʍ] being a true fricative?

Woleaian had *f *pʷ > f ɸʷ (and Proto-Micronesian *f patterned as palatalized - cf. *f > ɦʲ in Marshallese), and Irish Gaelic has vʲ/w - so yes, you can just do that

Does any natlang feature (a set of) preglottalized (voiceless) nasals?

Preglottalized nasals are attested, but I don't know if there are any attestations of a voicing contrast in preglottalized nasals. Partial voicelessness could be a phonetic detail, maybe - 'voiceless' nasals usually still have negative VOT.

My understanding of prenasalised consonants is that /kamba/ would be syllabified as /ka.mba/. If for example your only consonant clusters were intervocallic /mb/, /nd/ and /ŋg/, then it would be feasible to analyse these as prenasalised consonants and your syllable structure as CV

In some Papuan languages, the nasal component of a prenasalized stop can close a preceding syllable where possible, except in particularly slow speech.

Can a language have /ʝ ɣʷ/ without /j w/?

This isn't attested in PHOIBLE, but /ʝ β/ without /j w/ is (Pisamira, Chiriguano, Koryak), and /ɣʷ/ without /w/ is (Hiw). Analyzing these as fricatives would probably come down to phonotactic detail - Hiw is conventionally analyzed as having /w/, but it patterns phonotactically with the voiced fricatives rather than with /j/.

[Nloki]

Does any natlang feature (a set of) preglottalized (voiceless) nasals?

Preglottalized nasals are attested, but I don't know if there are any attestations of a voicing contrast in preglottalized nasals. Partial voicelessness could be a phonetic detail, maybe - 'voiceless' nasals usually still have negative VOT.

I meant a set of preglottalized nasals alone, not a contrastive difference among phonemic voiceless and voiced ones. For example: /ˀm̥ ˀn̪̊ ˀɲ̊ ˀŋ̊/ which would turn into plain voiced nasal allophones between vowels, and also a phonemic set of the latter ones turning into approximants intervocallically (again).

[Sequor]

How are prenasalized consonants treated in between vowels? Are they considered separate or the same. For example, let's say you have /kamba/, would it be /ka.mba/ or /kam.ba/? I know that prenasalized consonants are treated as one phoneme at the beginning of the word instead of a syllabic nasal followed by a cosnont (/mba/ and not /m.ba/), but could/does the rule change under other circumstances?

The Wikipedia article "prenasalized consonant" contains a pair of beautiful spectrograms showing how Sri Lankan Malay distinguishes /a.mba/ from /am.ba/ (it has something like /a.mba/ [amba] with a very short [m] versus /am.ba/ [amba] with a very short [a] before the pre-nasalized consonant).

[yangfiretiger121]

Is {i, y, ɨ → y} but {j, j˗ → ʝ̞˗} with separate [ɥ] plausible, or would {j, ɥ, j˗ → ɥ} happen as well?

[Davush]

Can a language have /ʝ ɣʷ/ without /j w/?

Arguably some dialects of Spanish have this, where initial /w/ is fortified to /gw/ or /ɣw/ which is most notable in loans but also things like huevo as /gwevo/.

[Zekoslav]

Can a language have /ʝ ɣʷ/ without /j w/?

Arguably some dialects of Spanish have this, where initial /w/ is fortified to /gw/ or /ɣw/ which is most notable in loans but also things like huevo as /gwevo/.

/w/ and /j/ could easily fortify and become /ɣʷ/ and /ʝ/, which would lead to the exact situation you're describing. Although, the languages I know of that fortified /w/ and /j/ usually go further and end up with stops and affricates. Maybe if you have preexisting /ɣʷ/ and /ʝ/ the development would stop at that point.

[holbuzvala]

@Ahzoh

Not sure if this helps, but in one of my langs I diachronically generated /ʝ ɣʷ/ like so:

/ɣiV ɣuV/ -> /ɣʲV ɣʷV/ -> /ʝV ɣʷV/

Might work without those final vowels.

---

So I have /tʰ/ which changes to /r̥/ when bounded by vowels or a word boundary. e.g.: /tʰatʰ/-> /r̥ar̥/, /atʰa/ -> /ar̥a/.

However, I am having trouble deciding what /tʰ/ should become when adjacent to other consonants, particularly stops and voiced sonorants. The current consonant inventory is as follows:

/p t k pʰ tʰ kʰ pʲ tʲ kʲ pʷ tʷ kʷ/
/s t͡s/
/m n ŋ/
/r l ɣ/

Any ideas? /pʰ/ becomes /f/ everywhere, so maybe /tʰ/ could become /s/ in clusters?

[Nortaneous]

the only language I know of with *tʰ > r̥ is Nivkh, and Nivkh doesn't object at all to /r̥/ in clusters