Project Mammoth Hunter - Help me make my first conlang

[Xwtek]

-velar vs. uvular plosives: [k]-[q], [g]-[ɢ], [x]-[X] and [ɣ]-[ʁ]. I can pronounce the velars more back in my throat but they doesn't sound very different. Also, I wouldn't be able to tell them apart if I hear them in speech. Maybe I'm just not used to this distinction yet, idk.

for uvular, use the velar equivalent, but drag the tongue far as possible. Note that [ɢ] is quite difficult to pronounce, but it doesn't matter since there is just a few of languages contrasting it anyway. Also, velar-uvular fricatives are pretty rare. Just use the french R (and the voiceless version), and you'll be actually fine.

-the vowel [ɯ]: I want to include this badly in my conlang but every time I try to pronounce it, it sounds either like [y] or [ø]. But I won't give up yet and countinue practising

if you can pronounce (and distinguish) [ɨ], use that. I think it's present in some English dialects. Probably your English dialect is too. Hardly any language distinguish [ɨ] and [ɯ]. Otherwise, just say [u], but force the lips to take shape of [i]. (The exact opposite process is used to form [y]). Despite the method of production, it's actually phonetically quite similar. I read that Acehnese associate their back unrounded vowel with German rounded vowel.

[Këkkytir]

Phonetics

General

• CV(C) syllable structure (more about this later)
• Stress is always on the first syllable of the root
• Agglutinative morphology, lots of suffixes

Vowels

• Progressive, stem-controlled, backness vowel-harmony (plus a secondary roundedness agreement of affixes, as in hungarian)
• No diphthongs, no phonemic vowel length
• Evolved from a proto-language with vowel inventory similar to today's hungarian and turkish one
• Should end up to be a variation on the classic 5 vowels + schwa system, that is dominant among today's paleo-siberian languages (Wikipedia)
• Include /ɯ/ or /ɨ/ (which are common among paleo-siberian languages according to http://citeseerx.ist.psu.edu/viewdoc/do ... 1&type=pdf)

Proto-language vowel inventory:

	Front	Central	Back
Close	/i/, /y/		/ɯ/, /u/
Mid	/e̞/, /ø̞/		/o̞/
Open		/ä/

Sound changes:

1. /ø̞/, /o̞/ -> /ə/ in unstressed syllables
2. /u/ -> /ɯ/, /ø̞/ -> /e̞/
3. some /y/ -> /i/ (recently ongoing) but preserved in affixes

Mammoth hunter's language vowel inventory:

	Front	Central	Back
Close	/i/, /y/		/ɯ/
Mid	/e̞/	/ə/	/o̞/
Open		/ä/

[Këkkytir]

Romanization
I'll follow the turkish notation, because I think that's the most logical for this system:
a /ä/, o /o̞/, e /e̞/, ö /ø̞/, ı /ɯ/, u /u/, i /i/, ü /y/ and ō /ə/.

Note that you get the front version of a back vowel by adding diacritic (umlaut or tittle) to it. Since e is not exactly the front pair of a, they are represented by unrelated symbols. As for the schwa, I wanted a symbol that represents that it was originally o or ö.

[Këkkytir]

Vowel-harmony

Proto-language:

• Vowels are grouped by backness and roundedness
• Back vowels: a, o, ı, u, front vowels: e, ö, i, ü
• Unrounded vowels: a, e, ı, i, rounded vowels: o, ö, u, ü
• A morpheme contains all-front or all-back vowels, but roundedness can mix
• Vowels in an affix must agree with the root in backness
• Additionally, if the root contains an unrounded vowel, all vowels in the affix must be unrounded. Otherwise, all vowels in the affix must be rounded.
• A specific vowel in an affix either follows an open scheme or a close scheme, which means that it either alternates between a/o/e/ö (=V₁) or between ı/u/i/ü (=V₂) (according to the above rules of vowel harmony). The vowel's scheme doesn't depend on the root's vowel quality; it is inherent to the affix and to the position of the vowel in the affix.
• Different vowels in the same affix can follow different schemes

Exmple (with gibberish):
A suffix -kV₁tV₂r can have the forms -katır/-kotur/-ketir/-kötür. The words gemök and unjpo would be conjugated as gemökketir and unjpokotur.

Effect of sound changes:

• Vowel-harmony is maintained
• ō counts as a neutral rounded vowel
• Affixes follow the sound changes with slight delay

1. (o, ö -> ō in unstressed syllables): this affects the vowel schemes: V₁=a/ō/e/ō, V₂=ı/u/i/ü. As the first syllable of the root is always stressed and front vowels cannot mix with back vowels, this change does not cause ambiguity. Roundedness is also preserved. (E.g. gemōkketir, unjpōkōtur)
2. (some rounded vowels became unrounded: u -> ı, ö -> e): V₁=(a, ō)/ō/(e, ō)/ō, V₂=ı/ı/(i, ü)/ü. Backness is preserved, but roundedness isn't. Roundedness (ō and ü) is retained in the affixes (E.g. gemōkketir, ınjpōkōtır)
   • 2'. (Front rounded suffixes of front unrounded roots become front unrounded (= roundedness lost in step 2. is only preserved in case of back affixes)): V₁=(a, ō)/ō/e/ō, V₂=ı/ı/i/ü.(E.g. gemōkketir, ınjpōkōtır unchanged)
3. (ü -> i in some words (recently ongoing) but preserved in affixes): V₁=(a, ō)/ō/(e, ō)/ō, V₂=ı/ı/(i, ü)/ü. Same as in step 2., backness is preserved, roundedness isn't. Roundedness is retained in affixes. (E.g. gemōkketir, ınjpōkōtır unchanged)

[bradrn]

Looks great! I don’t think I’ve seen a conlang with vowel harmony before, so it’s nice to see that you’ve worked that out in detail. I do have one critique though: it is very, very odd to have a language with /ɯ/ but not /u/. Unconditional /ɯ/ → /u/ as a sound change is also unrealistic, so I would recommend removing that sound change. EDIT: I forgot that Japanese has a vowel system of /a e i o ɯ/, so this is more plausible than I thought. Still, it’s not a common vowel system at all. Also, I’ve just noticed that you list both /u/ and /ɯ/ in your romanisation, so you may have a typo somewhere.

Do you have any idea about the consonants yet?

[Këkkytir]

-velar vs. uvular plosives: [k]-[q], [g]-[ɢ], [x]-[X] and [ɣ]-[ʁ]. I can pronounce the velars more back in my throat but they doesn't sound very different. Also, I wouldn't be able to tell them apart if I hear them in speech. Maybe I'm just not used to this distinction yet, idk.

for uvular, use the velar equivalent, but drag the tongue far as possible. Note that [ɢ] is quite difficult to pronounce, but it doesn't matter since there is just a few of languages contrasting it anyway. Also, velar-uvular fricatives are pretty rare. Just use the french R (and the voiceless version), and you'll be actually fine.

Thanks for the help! A velar-uvular contrast is present in many paleosiberian languages, which I've found strange. I'm doing better with the [q] now, but I'll probably get rid of the velar-uvular fricatives with a clever sound change :)

-the vowel [ɯ]: I want to include this badly in my conlang but every time I try to pronounce it, it sounds either like [y] or [ø]. But I won't give up yet and countinue practising ;)

if you can pronounce (and distinguish) [ɨ], use that. I think it's present in some English dialects. Probably your English dialect is too. Hardly any language distinguish [ɨ] and [ɯ]. Otherwise, just say [u], but force the lips to take shape of [i]. (The exact opposite process is used to form [y]). Despite the method of production, it's actually phonetically quite similar. I read that Acehnese associate their back unrounded vowel with German rounded vowel.

Good to know that the Acehnese find [ɯ] and [y] similar, too. I can't pronounce [ɨ] either, so I'll just go with [ɯ] because it sounds more distinctive to me.

[Këkkytir]

Looks great! I don’t think I’ve seen a conlang with vowel harmony before, so it’s nice to see that you’ve worked that out in detail. I do have one critique though: it is very, very odd to have a language with /ɯ/ but not /u/. Unconditional /ɯ/ → /u/ as a sound change is also unrealistic, so I would recommend removing that sound change. EDIT: I forgot that Japanese has a vowel system of /a e i o ɯ/, so this is more plausible than I thought. Still, it’s not a common vowel system at all. Also, I’ve just noticed that you list both /u/ and /ɯ/ in your romanisation, so you may have a typo somewhere.

I'm glad you like it! :D
Yeah, I guess an /u/ -> /ɯ/ sound change it's a little unrealistic as /u/ is much more common in languages than /ɯ/. But do you think it'd be ok to do a conditional /u/ -> /ɯ/ sound change instead? Maybe it should depend on the surrounding consonants.

Do you have any idea about the consonants yet?

Yes, I have a consonant chart, but I don't have very good ideas for the sound changes yet. Let me think about it a little more.

[bradrn]

Looks great! I don’t think I’ve seen a conlang with vowel harmony before, so it’s nice to see that you’ve worked that out in detail. I do have one critique though: it is very, very odd to have a language with /ɯ/ but not /u/. Unconditional /ɯ/ → /u/ as a sound change is also unrealistic, so I would recommend removing that sound change. EDIT: I forgot that Japanese has a vowel system of /a e i o ɯ/, so this is more plausible than I thought. Still, it’s not a common vowel system at all. Also, I’ve just noticed that you list both /u/ and /ɯ/ in your romanisation, so you may have a typo somewhere.

I'm glad you like it!
Yeah, I guess an /u/ -> /ɯ/ sound change it's a little unrealistic as /u/ is much more common in languages than /ɯ/. But do you think it'd be ok to do a conditional /u/ -> /ɯ/ sound change instead? Maybe it should depend on the surrounding consonants

Conditional /u/→/ɯ/ sounds fine.

[Kuchigakatai]

Looks great! I don’t think I’ve seen a conlang with vowel harmony before, so it’s nice to see that you’ve worked that out in detail.

Some other conlangs with vowel harmony by ZBBers:
- Vædty Qys, by akam chinjir (first conlang in his scratchpad) [sporadic backness-based harmony]
- Dasaro, by DePaw [backness-and-height-based harmony]
- Luworese, by Chengjiang [rounding-based harmony]
- Kannow, by Nortaneous and Hallow XIII (description taken down, but you can see some of it here) [height-based harmony]

[Këkkytir]

Looks great! I don’t think I’ve seen a conlang with vowel harmony before, so it’s nice to see that you’ve worked that out in detail.

Some other conlangs with vowel harmony by ZBBers:
- Vædty Qys, by akam chinjir (first conlang in his scratchpad) [sporadic backness-based harmony]
- Dasaro, by DePaw [backness-and-height-based harmony]
- Luworese, by Chengjiang [rounding-based harmony]
- Kannow, by Nortaneus and Hallow XIII (description taken down, but you can see some of it here) [height-based harmony]

Wow these are all very interesting sources! I'll look into them.

[Këkkytir]

Just posting some consonant charts. I'm still not satisfied with the logic of sound-changes I've came up with, so I'll write about them later.


Proto-language (no phonemic voicing contrast; sounds in parentheses are under consideration):

Consonants	Bilabial	Alveolar	Retroflex	Palatal	Velar	Uvular	Glottal
Nasal	/m/	/n/	/ɳ/		/ŋ/
Stop	/p~b/	/t~d/	/ʈ~ɖ/		/k~g/	/q~ɢ/	/ʔ/
Fricative	/ɸ~β/	/s~z/, (/ɬ~ɮ/)	/ʂ~ʐ/		/x~ɣ/	(/X~ʁ/)	(/h/)
Approximant		/l/	(/ɭ/)	/j/
Trill		/r/

Mammoth hunters' language:

Consonants	Bilabial	Alveolar	Retroflex	Palatal	Velar	Uvular	Glottal
Nasal	/m/	/n/		/ɲ/	/ŋ/
Stop	/p/	/t/, /d/		/c/	/k/, /g/	(/q/)	(/ʔ/)
Fricative		/s/, (/ɬ/)		/ɕ/
Approximant		/l/		/j/, (/ʎ/)
Tap		/ɾ/

[bradrn]

Just posting some consonant charts.

These look good!

I'm still not satisfied with the logic of sound-changes I've came up with, so I'll write about them later.

Note that you can always ask questions about sound change on the Sound Change Quickie Thread. Also, I quite struggle with putting together sound changes as well, so I created the Sound Change Critique Thread, where you can post a whole set of sound changes and get a critique of the whole thing.

[Këkkytir]

Sound changes for consonants (still very uncertain):

1. Fricatives, except for /ɬ~ɮ/, became stops
2. Retroflexes palatalize before front unrounded vowels (i, e), merge with alveolars anywhere else (/ʈ~ɖ/ -> t, /t~d/ -> d introduces voicing contrast)
3. Palatalization of /ɬ~ɮ/: /ɬ~ɮ/ -> /ɕ~ʑ/ before i and e, /ɬ~ɮ/ -> /s~z/ anywhere else; /r/ becomes a tap /ɾ/
4. /q~ɢ/ -> /k/, /k~g/ -> /g/ (introduces voicing contrast again)
5. Stops and fricatives without voicing distinction become uniformly unvoiced
6. Glottal stop is not phonemic anymore, only present in word-inital position (before vowel), and between vowels

Other ideas:
-At the beginning, the proto-language had a strict CV syllable structure
-Loss of certain/all unstressed word-final vowels -> introduces word-final CVC syllables
-Loss of certain/all unstressed vowels between two stops or between a nasal and a stop -> more CVC syllables
-Fortition of word-final consonants: word-final nasals, fricatives, voiced stops become voiceless stops; word-final /l/ becomes /ɾ/
-Lenition of word-initial consonants, e.g. /p/, /b/ -> /m/ (maybe only before high unrounded vowels)
-In consonant clusters, assimilation of the coda happens by place of articulation (e.g. /mt/ -> /nt/, /tk/ -> /k:/)
-Glottal stop assimilates to the following voiceless top: /ʔ/ + /p/, /t/, /k/ -> /p:/, /t:/, /k:/

[bradrn]

Sound changes for consonants (still very uncertain):

These look much more interesting than what I usually come up with! But I do have a couple of critiques:

Fricatives, except for /ɬ~ɮ/, became stops

It is very weird for all fricatives (or even most of them) to become stops, unconditionally. In our world, fricativelessness is an extremely uncommon areal feature confined mainly to Australia. Either make this change conditional, or exclude some more fricatives from this change (or do both).

Retroflexes palatalize before front unrounded vowels (i, e), merge with alveolars anywhere else (/ʈ~ɖ/ -> t, /t~d/ -> d introduces voicing contrast)

I’m not quite sure what you’re saying about the voicing contrast here?

Glottal stop is not phonemic anymore, only present in word-inital position (before vowel), and between vowels

Could you rephrase this as a phonological rule? At the moment this gives very little information about exactly what happens to glottal stops.

Other ideas:
-At the beginning, the proto-language had a strict CV syllable structure
-Loss of certain/all unstressed word-final vowels -> introduces word-final CVC syllables
-Loss of certain/all unstressed vowels between two stops or between a nasal and a stop -> more CVC syllables
-Fortition of word-final consonants: word-final nasals, fricatives, voiced stops become voiceless stops; word-final /l/ becomes /ɾ/
-Lenition of word-initial consonants, e.g. /p/, /b/ -> /m/ (maybe only before high unrounded vowels)
-In consonant clusters, assimilation of the coda happens by place of articulation (e.g. /mt/ -> /nt/, /tk/ -> /k:/)
-Glottal stop assimilates to the following voiceless top: /ʔ/ + /p/, /t/, /k/ -> /p:/, /t:/, /k:/

These all look fine.

[Darren]

Fricatives, except for /ɬ~ɮ/, became stops

It is very weird for all fricatives (or even most of them) to become stops, unconditionally. In our world, fricativelessness is an extremely uncommon areal feature confined mainly to Australia. Either make this change conditional, or exclude some more fricatives from this change (or do both).

There is a very close example in Kokama-Kokamilla, where the only fricative is /x/, and the historical fricatives /s ʃ/ became /ts tʃ/ unconditionally. I think in this case /x/ was exempt because /kx/ is pretty rare, but /ɬ~ɮ/ could easily be exempt because it is more like an approximant. According to the Index Diachronica, x → k is also attested, so there's nothing wrong with doing that.

Basically, it's a weird sound change, but not unprecedented.

Edit: There's an article specifically about the sound change here

[Këkkytir]

Sound changes for consonants (still very uncertain):

These look much more interesting than what I usually come up with! But I do have a couple of critiques:

Fricatives, except for /ɬ~ɮ/, became stops

It is very weird for all fricatives (or even most of them) to become stops, unconditionally. In our world, fricativelessness is an extremely uncommon areal feature confined mainly to Australia. Either make this change conditional, or exclude some more fricatives from this change (or do both).

I think you have a good point. I guess I can't kill off all the fricatives. I'll ask about this in the Sound Change Quickie Thread.

Retroflexes palatalize before front unrounded vowels (i, e), merge with alveolars anywhere else (/ʈ~ɖ/ -> t, /t~d/ -> d introduces voicing contrast)

I’m not quite sure what you’re saying about the voicing contrast here?

Well, this was my idea of introducing voicing contrast in individual stops, like t/d. Let's say, that you have two stops without phonemic voicing contrast: /t~d/ and /ʈ~ɖ/. In my case, this means that voicing of these consonants is fully dependent on the context. But what happens, if the two stops start to merge because of a sound change. Of course, I could simply say that /ʈ~ɖ/ becomes /t~d/ and the language loses distinction between some pairs of words, like [da] and [ɖa]. But what if /ʈ~ɖ/ keeps to be pronounced with a little bit more stress than /t~d/ (like an aspirated consonant) and this slowly turns into a voicing contrast, so the end result is /ʈ~ɖ/ -> t and /t~d/ -> d.
I'm not basing this sound change on a natural language, so it might be very unlikey, but that's what I came up with. Now I feel like I should ask about this topic in the Sound Change Quickie Thread, too.

Glottal stop is not phonemic anymore, only present in word-inital position (before vowel), and between vowels

Could you rephrase this as a phonological rule? At the moment this gives very little information about exactly what happens to glottal stops.

I have no exact idea about this change yet, but I want the glottal stop to be present only in word-inital position and between vowels.

Other ideas:
-At the beginning, the proto-language had a strict CV syllable structure
-Loss of certain/all unstressed word-final vowels -> introduces word-final CVC syllables
-Loss of certain/all unstressed vowels between two stops or between a nasal and a stop -> more CVC syllables
-Fortition of word-final consonants: word-final nasals, fricatives, voiced stops become voiceless stops; word-final /l/ becomes /ɾ/
-Lenition of word-initial consonants, e.g. /p/, /b/ -> /m/ (maybe only before high unrounded vowels)
-In consonant clusters, assimilation of the coda happens by place of articulation (e.g. /mt/ -> /nt/, /tk/ -> /k:/)
-Glottal stop assimilates to the following voiceless top: /ʔ/ + /p/, /t/, /k/ -> /p:/, /t:/, /k:/

These all look fine.

Thanks! :) I still have a lot of work to do in the phonetics but I want to create some basic vocabulary for a change next week.

[Këkkytir]

Fricatives, except for /ɬ~ɮ/, became stops

It is very weird for all fricatives (or even most of them) to become stops, unconditionally. In our world, fricativelessness is an extremely uncommon areal feature confined mainly to Australia. Either make this change conditional, or exclude some more fricatives from this change (or do both).

There is a very close example in Kokama-Kokamilla, where the only fricative is /x/, and the historical fricatives /s ʃ/ became /ts tʃ/ unconditionally. I think in this case /x/ was exempt because /kx/ is pretty rare, but /ɬ~ɮ/ could easily be exempt because it is more like an approximant. According to the Index Diachronica, x → k is also attested, so there's nothing wrong with doing that.

Basically, it's a weird sound change, but not unprecedented.

Edit: There's an article specifically about the sound change here

Thanks for the info! I really like the Omagua consonant inventory, it is very close to what I wanted to achieve.
I was thinking the same about /ɬ~ɮ/ being an exempt to the rule.

[Darren]

Well, this was my idea of introducing voicing contrast in individual stops, like t/d. Let's say, that you have two stops without phonemic voicing contrast: /t~d/ and /ʈ~ɖ/. In my case, this means that voicing of these consonants is fully dependent on the context. But what happens, if the two stops start to merge because of a sound change. Of course, I could simply say that /ʈ~ɖ/ becomes /t~d/ and the language loses distinction between some pairs of words, like [da] and [ɖa]. But what if /ʈ~ɖ/ keeps to be pronounced with a little bit more stress than /t~d/ (like an aspirated consonant) and this slowly turns into a voicing contrast, so the end result is /ʈ~ɖ/ -> t and /t~d/ -> d.
I'm not basing this sound change on a natural language, so it might be very unlikey, but that's what I came up with. Now I feel like I should ask about this topic in the Sound Change Quickie Thread, too.

I'm not sure whether this is completely pertinent, but it's not uncommon for languages to have the voiceless coronal articulated further forward than the voiced one, or to have a voiced palatal but no voiceless*. A couple of examples (most likely bad ones):

Girawa /p t dʒ ʔ/
Onondaga /t dʒ k ʔ/
Jarawara /b t ɟ k/
Dawan /p b t dʒ k ʔ/
Jamamadi /b t ɟ k ʔ/
Kuikúro-Kalapálo /p t dʲ k/
Paraguayan Guaraní /p t dʒ k kʷ ʔ/ (and /mb nd ŋg ŋgʷ/)
Kaiwá /p t dʒ k ʔ/
Nhadeva /p t dʒ k g kʷ gʷ ʔ/

So perhaps the change could be justified as

1. t~d → t
2. ʈ~ɖ → ʈ
3. ʈ → ɖ
4. ɖ → d

I suspect in a lot of these inventories the further back /d/ doesn't correspond diachronically to /t/, but these show that the state at stage 3 would be possible.


*Usually in Papuan languages it's been replaced by /s/.

[bradrn]

Retroflexes palatalize before front unrounded vowels (i, e), merge with alveolars anywhere else (/ʈ~ɖ/ -> t, /t~d/ -> d introduces voicing contrast)

I’m not quite sure what you’re saying about the voicing contrast here?

Well, this was my idea of introducing voicing contrast in individual stops, like t/d. Let's say, that you have two stops without phonemic voicing contrast: /t~d/ and /ʈ~ɖ/. In my case, this means that voicing of these consonants is fully dependent on the context. But what happens, if the two stops start to merge because of a sound change. Of course, I could simply say that /ʈ~ɖ/ becomes /t~d/ and the language loses distinction between some pairs of words, like [da] and [ɖa]. But what if /ʈ~ɖ/ keeps to be pronounced with a little bit more stress than /t~d/ (like an aspirated consonant) and this slowly turns into a voicing contrast, so the end result is /ʈ~ɖ/ -> t and /t~d/ -> d.
I'm not basing this sound change on a natural language, so it might be very unlikey, but that's what I came up with. Now I feel like I should ask about this topic in the Sound Change Quickie Thread, too.

This makes perfect sense, and sounds realistic as-is. But it would be nice to know the exact conditions where /t~d/ and /ʈ~ɖ/ are voiceless and voiced respectively.

[Qwynegold]

Hardly any language distinguish [ɨ] and [ɯ]. Otherwise, just say [u], but force the lips to take shape of [i]. (The exact opposite process is used to form [y]). Despite the method of production, it's actually phonetically quite similar. I read that Acehnese associate their back unrounded vowel with German rounded vowel.

That's interesting. When I studied Chinese we were taught to pronounce [ɤ], and I was like "wtf this sounds almost completely like [ø] even though they're supposed to be total opposites!"