Conlang Random Thread

[doctor shark]

In off-topic randomness, I was bored, so I whipped this up.

More: show

[TurkeySloth]

I'm, actually, strongly favouring the central vowels being marked. But, if I shift at the last minute, do you recommend changing all peripherals to being marked with circumflexes? As for the vowel system's weirdness, the language is spoken in a fictional galaxy, so that should afford some oddity. I plan to change the system for the setting's revamped sci-fi version. Additionally, I don't know, or want to figure out, the language's full phonology right now because playing with the setting's English-equivalent cost me story development during the last campaign I ran (planning a new one right now).

[bradrn]

I'm, actually, strongly favouring the central vowels being marked. But, if I shift at the last minute, do you recommend changing all peripherals to being marked with circumflexes?

I’m not sure I understand this question. What exactly do you mean by ‘shift’?

[Xwtek]

I'd romanize it as /ɐ ɑ ɪ ɨ ø̞ ɵ̞/ <a o i y e u>

[TurkeySloth]

I'm, actually, strongly favouring the central vowels being marked. But, if I shift at the last minute, do you recommend changing all peripherals to being marked with circumflexes?

I’m not sure I understand this question. What exactly do you mean by ‘shift’?

Change my mind. Considering the context I thought I could use shorthand.

[Nila_MadhaVa]

If I have contrasting C-Cʰ phonemes, could I analyze Cʰ as Ch clusters with cluster-induced allophony? I thought the answer was "no, if they contrast, they are phonemic", but then I read this at the top of the wikipedia page for Ubykh regarding clicks in Khoisan languages, and now I'm confused:

"although some analyses[1] view a large proportion of the clicks in these languages as clusters"

[bradrn]

If I have contrasting C-Cʰ phonemes, could I analyze Cʰ as Ch clusters with cluster-induced allophony? I thought the answer was "no, if they contrast, they are phonemic", but then I read this at the top of the wikipedia page for Ubykh regarding clicks in Khoisan languages, and now I'm confused:

"although some analyses[1] view a large proportion of the clicks in these languages as clusters"

I seem to remember reading on Wikipedia about a Taiwanese language which has no aspirates, but does have Ch clusters which are pronounced as aspirates. I can’t quite remember exactly why this analysis was chosen, but I believe it was because aspirated consonants are found only between vowels, so e.g. something like [atʰa] can be analysed as [at.ha].

I'm, actually, strongly favouring the central vowels being marked. But, if I shift at the last minute, do you recommend changing all peripherals to being marked with circumflexes?

I’m not sure I understand this question. What exactly do you mean by ‘shift’?

Change my mind. Considering the context I thought I could use shorthand.

Thanks for explaining! To answer your question now: I personally prefer to keep the peripherals unmarked, but it’s ultimately your aesthetic decision. If you’re not sure maybe you could write out a sample text in several romanizations, and see which one you like best.

[Nila_MadhaVa]

I seem to remember reading on Wikipedia about a Taiwanese language which has no aspirates, but does have Ch clusters which are pronounced as aspirates. I can’t quite remember exactly why this analysis was chosen, but I believe it was because aspirated consonants are found only between vowels, so e.g. something like [atʰa] can be analysed as [at.ha].

Yeah, my aspirates are fully contrasted with their plain counterparts, so I'm pretty sure I can't analyze them as clusters. The language in question is inspired by the Jadoonese dialogue from last week's Doctor Who, which is completely monosyllabic with a C(r/l)V structure. It uses English consonants and has two vowels, /əʊ̯/ and the occasional /ä/. I wanted to stick to that phonology/phonotactics as much as possible, but had to get creative to bring up the number of possible syllables. So the aspirates started out in my mind as Ch clusters, but obviously, given the syllable structure, they would be phonemic. The Khoisan clicks might all be initial, as your example is intervocalic, which would allow for analysis as clusters.

Being able to justify the aspirates as Ch clusters would have helped me feel better about deviating from Jadoonese, but since English has allophonic aspirated initial stops, I don't feel too bad about having an aspirated series. The prenasalized consonants on the other hand, initially conceived of as NC clusters, I'm finding harder to justify

EDIT: Is there anything else I could do based on English phonology to replace the prenazalized and prenasalized-aspirated series? I already have (s)Cw/l/r/y clusters.

[Xwtek]

If your aspirates behave like a cluster, I think it should be analyzed as a cluster.

In Khmer, it's because the infix -um- would split the phoneme to /C/ and /h/, like t-um-hum
If your language disallows cluster like khr or phl (English does), and there can only be a maximum of 2 consonants in the syllable-initial cluster, go analyze it as a cluster.

[Nila_MadhaVa]

If your aspirates behave like a cluster, I think it should be analyzed as a cluster.

In Khmer, it's because the infix -um- would split the phoneme to /C/ and /h/, like t-um-hum
If your language disallows cluster like khr or phl (English does), and there can only be a maximum of 2 consonants in the syllable-initial cluster, go analyze it as a cluster.

With a (s/N)C(h)(w/l/r/y)V syllable structure, only monosyllabic morphemes and no inflection, they don't behave like clusters. I was just thrown a bit by the wikipedia page I quoted earlier, but I think bradrn's post pointed me in the right direction; some factor (e.g. being word initial only or something) allows for the possibility for Khoisan clicks to be analyzed as clusters.

[bradrn]

If your aspirates behave like a cluster, I think it should be analyzed as a cluster.

In Khmer, it's because the infix -um- would split the phoneme to /C/ and /h/, like t-um-hum
If your language disallows cluster like khr or phl (English does), and there can only be a maximum of 2 consonants in the syllable-initial cluster, go analyze it as a cluster.

With a (s/N)C(h)(w/l/r/y)V syllable structure, only monosyllabic morphemes and no inflection, they don't behave like clusters. I was just thrown a bit by the wikipedia page I quoted earlier, but I think bradrn's post pointed me in the right direction; some factor (e.g. being word initial only or something) allows for the possibility for Khoisan clicks to be analyzed as clusters.

I think it’s worth noting that the cluster analysis may be controversial (or at least, this is what the Wikipedia article on ǃXóõ implies). I don’t think there’s any actual evidence one way or another: I believe clicks in these languages have been analysed as clusters simply because otherwise there are far, far too many consonants to be reasonable (e.g. Wikipedia lists 111 clicks for Western ǃXoon). And as for being word initial only, it’s not so much that clicks are forbidden outside initial position; rather, Khoisan languages generally only allow five or six consonants in non-initial position.

[chris_notts]

Funny this came up, since my current project is a bit Austroasiatic in style, at least in the phonology. And what AA language would be lacking in infixes? I have a grammar of Semelai and a few PDF grammars and it's impressive how messy the internal changes can get in some AA languages.

I have a few infixes, all basically V->N derivational, that do split Ch and C' clusters. Shapes are below, where C = reduplicated onset and V = reduplicated vowel. In all cases the infix occurs before or after the onset of the final stressed syllable, which may make it a prefix for CV roots:

-Ci- = diminutive, doesn't split onset clusters
e.g. pka -> pipka

-(C)aN- = to be decided, does split onset clusters
e.g.
ga -> gaŋga
kha -> kahŋa <-- splits cluster, metathesis of fricative and nasal

-r(n)- = instrumental~locative nominalisation, mixed behaviour. Can split onset, form part of complex onset, or form minor syllable depending on root shape and onset consonant
e.g.
na -> nra (1 syllable) <-- inserted into root as second member of cluster
ga -> gr.na (2 syllables) <-- -r.n- occurs after voiced obstruent
kha -> kr.ha (2 syllables) <-- splits cluster

(V)s- = to be decided, does split onset clusters, reduplicated V only surfaces when inserted into a complex onset cluster
na -> sna
kha -> kasha

I'm currently debating whether -Vs- should be a process~result nominalisation and -(C)aN- should be an actor nominalisation or vice versa. I know that length doesn't really matter much in the grand scheme of things, but I feel like -s-, which is a single consonant for all possible root-shapes but CCV(C), should be allocated to the most frequent / basic of the two. I'm just not sure which one is actually likely to be more frequent in discourse.

I guess if we look at English, probably the most reduced nominalisations are the patient~result ones, which are often zero derived or suppletively formed:

to hit ~ a hit
to run ~ a run
give ~ gift
sing ~ song
...

Whereas actor nominalisations take -er, which, while minimal (/ə/ for me), is more phonologically marked than zero, and there are also quite a few compounds ending in -man (like postman). Process nominalisation via the -ing gerund is the most phonologically marked in the sense that it contains two phonemes instead of zero or one.

Compare with Spanish:

cantar (to sing)
cantante (singer)
canción (song)
cantar (infinitive also used as a process noun)

golpear (to hit)
golpeador (hitter)
golpe (hit)
golpear (infinitive also used as a process noun)

Again, result and process seem to be more minimal forms that actor nominalisations. I'm not sure that the patterns of two SAE languages can be extrapolated, but since similar patterns of zero result~process nominalisation seem to recur in languages like Chinese I think may be the -s- form should go that way.

[Pabappa]

I believe clicks in these languages have been analysed as clusters simply because otherwise there are far, far too many consonants to be reasonable (e.g. Wikipedia lists 111 clicks for Western ǃXoon).

And the vowel inventory is almost as exquisite, if you treat the phonation qualities as specifying different vowels. I particularly love this paragraph:

A long, glottalized, murmured, nasalized o with falling tone is written ôʼhõ. A long, strident nasalized o with low tone is written òqhõ, since Traill analyzes stridency as phonemically pharyngealized murmur. (Note that phonetically these are distinct phonations.)

[Nila_MadhaVa]

I think it’s worth noting that the cluster analysis may be controversial (or at least, this is what the Wikipedia article on ǃXóõ implies). I don’t think there’s any actual evidence one way or another: I believe clicks in these languages have been analysed as clusters simply because otherwise there are far, far too many consonants to be reasonable (e.g. Wikipedia lists 111 clicks for Western ǃXoon). And as for being word initial only, it’s not so much that clicks are forbidden outside initial position; rather, Khoisan languages generally only allow five or six consonants in non-initial position.

Thanks, that clears things up for me.

[Xwtek]

Only monosyllabic morphemes and no inflection

Honestly, I think it is unrealistic. Even Chinese has disyllabic morpheme, like 蝴蝶 (which goes back to Middle Chinese) and 马达. While no inflection is doable, I think you mean word modification based on context, and there is no language without derivation.

[Nila_MadhaVa]

Only monosyllabic morphemes and no inflection

Honestly, I think it is unrealistic. Even Chinese has disyllabic morpheme, like 蝴蝶 (which goes back to Middle Chinese) and 马达. While no inflection is doable, I think you mean word modification based on context, and there is no language without derivation.

Oh, I'm aware of that, and I agree. But it is based on/inspired by a language as presented by a sci-fi tv show, and I wanted to see what I could come up with working within the limitations that imposes. Which means it is isolating, has no derivational morphology (the same "unmodified" morpheme can belong to various parts of speech), and has a small number of morphemes (by pushing the phonology and phonotactics to breaking point, I've managed to get around 4500 unique monosyllables). So definitely not realistic, but I am having fun with it.

[jal]

(the same "unmodified" morpheme can belong to various parts of speech)

The major problem with this is gardenpathing. I've had the same problems with Sajiwan, which does aim for realism (as a CEC), that e.g. has "dem" being the definite plural article, the generic article, the 3rd person pronoun (subject and object), the 3rd person possesive and the plural marker. I often wrote sentences that looked ok, but when rereading them later, I often got gardenpathed and had to rewrite or add commas to make more clear the meaning (in speech, intonation would do a lot of course).


JAL

[bradrn]

(the same "unmodified" morpheme can belong to various parts of speech)

The major problem with this is gardenpathing. I've had the same problems with Sajiwan, which does aim for realism (as a CEC), that e.g. has "dem" being the definite plural article, the generic article, the 3rd person pronoun (subject and object), the 3rd person possesive and the plural marker. I often wrote sentences that looked ok, but when rereading them later, I often got gardenpathed and had to rewrite or add commas to make more clear the meaning (in speech, intonation would do a lot of course).


JAL

What’s a ‘generic article’?

[jal]

What’s a ‘generic article’?

I don't know if that's the proper linguistic name, but it's used where English uses the plural, as in "birds can fly", Sajiwan "dem bah u flay". I think it works the same in Spanish.


JAL

[bradrn]

What’s a ‘generic article’?

I don't know if that's the proper linguistic name, but it's used where English uses the plural, as in "birds can fly", Sajiwan "dem bah u flay". I think it works the same in Spanish.

I think the most reasonable analysis is that that’s simply an indefinite article, realised as null. Note that all the other article/number combinations are fine, but that one is ungrammatical:

	s	p
indefinite	a bird flies	*a birds fly (but ‘∅ birds fly’ is fine)
definite[/cell]	the bird flies	the birds fly