zompist bboard

Porphyrogenitos wrote:A very simple thing I came up with, somewhat inspired by Hebrew spirantization:

I like it! Reminds me also of the Tuscan gorgia.

The sequences /tj/ and /sj/ undergo palatalization, resulting in [t͜ʃ] and [ʃ], respectively. These allophones are also subject to voicing intervocalically, producing the further allophones [d͜ʒ] and [ʒ]. E.g. /tjasja/ > [t͜ʃaʒa]

Does /dj/ not palatalize? That's a bit odd, although I think it does have precedent somewhere. The World's Major Languages claims that /d/ in Hausa only sometimes palatalized in environments where /t s z/ regularly did, although I've had trouble finding further discussion of this claim. Also, does the /i/ in a /ji/ sequence still become [e] if the /j/ is absorbed by this palatalization?

Chengjiang wrote:
I like it! Reminds me also of the Tuscan gorgia.

Thanks! And you're right, it is similar, I didn't think of that.

Chengjiang wrote: Does /dj/ not palatalize? That's a bit odd, although I think it does have precedent somewhere.

Oh, wow, that's a big mess-up, I just realized. No, /dj/ definitely becomes [d͜ʒ], I don't know how I forgot about that. And I guess /tj/ wouldn't be voiced, then, either. And I think neither /tj/ nor /dj/ undergo spirantization, then, since they're underlyingly consonant clusters, not single intervocalic stops.

Chengjiang wrote: Also, does the /i/ in a /ji/ sequence still become [e] if the /j/ is absorbed by this palatalization?

Wow, that's a great question. Hmm, it would depend on diachronic factors, I suppose, but I'm just going to say that the lowering rule applies before the palatalization rule. So no /i/ after [t͜ʃ] [d͜ʒ] [ʃ] [ʒ], then. But probably the palatalized realizations would eventually become phonemic through loanwords and morphological analogy, and /i/ would thus gradually be introduced after them.

I'm proud to finally have a first more or less stable phonology for my PIE based conlang, currently named East Nisha. In terms of phonology and morphology it comes close to Anatolian languages, does not exhibit any fancy, unique features so that it is quite easy to recognize it as an IE lang.

Consonants

Nasals /m n/
Stops /p p' t t' k₁ k₁' k₁^w k₁^w' ʔ/
Affricates /t͡s t͡s' t͡ɬ/
Fricatives /θ s ʃ x h₁/
Approximants /ʋ~w j l~ɬ/
Trill /r/
Unknown /k₂ (k₂') h₂/

As for the two consonants of unknown position I currently have no idea how to render them - k₂ is a direct reflex of PIE k^j and is more of palatal nature. It is not clear yet whether there is a ejective variant of it or not, I personally prefer to use a non-ejective only - a possible outcome could be /ɕ/ or even /c/. h₂ on the other hand is quite difficult to explain, it results from PIE laryngeals h₂ and h₃ in certain environments (mostly in initial position), a rendering as /ħ/ is likely but also a merger with /x/. Not sure about that yet.

Vowels
Front /i y e/
Central /a/
Back /u/

Lots of allophony here. /i/ can appear as /ɨ/ in some environments, /e/ can range from /ɛ~æ/ to /ə/ depending on the context. /a/ may appear as /ɑ/. That's pretty much about it, those are first thoughts/works and quite unstable, haven't had the time yet to work on the vowels.

There are surely some logical flaws to figure out, but I guess I will use this phonological system as a basis.

Examples are rare at this moment, I only have an older one available:

arsinna θersteš areši tanahai innaθar nešmet' tlennai
we went through the dark streets in our world

CONSONANTS
Nasals: /m n/
Plosives: /p b t d k g/
Fricatives: /s z h/
Affricate: /ʦ/
Apprximant: /j/
Flap: /ɾ/

VOWELS
Close: /i i: u u:/
Close-mid: /e e: ø øː o o:/
Open: /a a:/

PHONOTACTICS
For short vowels, the general syllable structure is (C)V(C). For long vowels, the general syllable structure is CL(C). In both cases, the coda is restricted to voiceless plosives, nasals, /j/, and /ɾ/. The only permissible onset consonant clusters are /pɾ/, /tɾ/, and /kɾ/.

STRESS
Stress is placed on the heaviest syllable furthest to the right of the word.

(deleted redundant post)

Khásoitoi
/ˈkʰä.si.ti/
[ˈkʰɑ.sɪ̝.tɪ̝]

/pʲʰ pʲ pʼ pʰ p m b̰ tʲʰ tʲ tʼ tʰ t d̰ kʲʰ kʲ kʼ kʰ k ɡ̰ kʷʼ kʷʰ kʷ ɡ̰ʷ qʼ qʰ q ɢ̰/
/m n ŋ/
/ɸʲʰ ɸʲ ɸʰ ɸ sʰ s çʰ ç xʷʰ xʷ χʰ χ/
/p͡ɸʼ p͡ɸʰ p͡ɸ t͡θ̱ʼ t͡θ̱ʰ t͡θ̱ t͡ʃʼ t͡ʃʰ t͡ʃ k͡xʼ k͡xʰ k͡x k͡ʷxʷʼ k͡ʷxʷ k͡ʷxʷ/
/lʲ l/
Palatalised unaspirated stops are actually partially voiced. The creaky voiced consonants are referred to as glottalic or glottalised consonants, as their origin and function is similar to such, but this is phonetically incorrect. /çʰ ç/ act as the palatalised equivalents of /χʰ χ/ and /sʰ s/. /t͡ʃʼ t͡ʃʰ t͡ʃ/ act as the palatalised equivalents of /t͡θ̱ʼ t͡θ̱ʰ t͡θ̱/.
/i u ä/
Vowels are voiceless after ejectives.
Vowels after various sounds:
Palatalised bilabial: [i ʊ̝ ɶ]
Plain bilabial: [ɪ̝ u ɒ̈]
Palatalised alveolar/palato-alveolar: [i ɯ̝̈ a]
Plain alveolar: [ɪ̝ ɯ ä]
Palatalised velar/plain velar/uvular: [i u ɑ]
Labiovelar: [y u ɒ]

Stress is phonemic. There are two levels of stress. Secondary stress is realised by some speakers by pharyngealising the vowels. Secondary stress always and only appears in words with at least four syllables. Primary stress always appears in words with two or more syllables.

Syllables are C(C)V(C). Initial syllables are C(C)(C)V(C). The vast majority of syllables have only one consonant or one consonant followed by a lateral in their onsets. Other clusters are especially rare in non-initial syllables. If one sound is palatalised, all sounds following it within the same cluster are palatalised. Allowed onsets: Mutations affect only the first phoneme of a word. Postpositions and clitics are “invisible” to mutations, meaning that the next word or word the clitic is attached to undergoes the mutation while the particle does not. Particles never trigger mutations. Monosyllables and pronouns do not go through mutation, but are not “invisible” to it. Mutations do not occur when the affected phoneme is followed by a sound other than a vowel or lateral. They never carry over sentences or quotes. A sound may undergo multiple mutations at once. The final mutation is written preceding the mutated consonant in brackets. Depalatalisation before /i/ is indicated by adding <[o]>. The mutations are listed in the order that they are applied:
Hard mutation turns fricatives and nasals into stops. It is triggered by a preceding ejective stop. Uvular fricatives are turned into plain velars if followed by /u/.
Glottalic mutation causes ejectives and aspirated stops to become plain stops and plain stops to become glottalised. It is triggered by a preceding glottalised stop.
Palatal mutation causes all consonants which contrast with palatalised versions of themselves to become palatalised. It is triggered by a preceding palatalised consonant or /i/.
U-mutation causes consonants which are palatalised to become palatalised. It is triggered by a preceding /u/.
Nasal mutation causes palatalised aspirated stops to lose their aspiration and palatalised unaspirated stops as well as glottalised stops to become nasals. It is triggered by a preceding nasal.
Soft mutation causes all stops to turn into fricatives. It is triggered by a preceding tenius affricate. The plain velars turn into uvular fricatives.
Uvular mutation causes non-uvular consonants to shift to their POV towards uvulars. It is triggered by a uvular stop.

This has been sitting in a tab on my notepad for a few days while I've tinkered with it. I have a vague idea that the speakers are desert nomads living on the fringes of a major empire. This is very much embryonic, and I'm not taking it too seriously, so a lot of the ideas half-formed at best. /ə əː a aː/ are markedly more common than the other vowels. Probably what happened was that the ancestor language had a relatively complex European-style vowel system with a consonant inventory kind of like Finnish. The consonants developed a bunch of allophones which were then made phonemic when the vowels collapsed into /ə əː a aː/. This persists up until shortly before the present, when /i u e o/ become phonemes again. Probably if I worked out some formal sound changes I could do something interesting with ablaut. C₁V(C₂)

C₁: any consonant.
C₂: any consonant except /t͡s/.

Coda labialised consonants are realised as [wC].

I haven't really decided on how stress is going to work, but since I think this language is going to be isolating in the process of becoming agglutinating, probably it'll be fixed on one of the root's syllables (most of the roots will be monosyllabic).

/p b t d k/ <p b t d q>
/m n ŋ/ <m n g>
/ɽ/ <r>
/β s ʃ h/ <f s j h>
/ʋ r j/ <w r y>

/i u e ɘ o ɛ ɜ ɔ ɶ ɑ/ <i u e z o x v c k a>

(C)V

----

/p pʰ b t̻ t̻ʰ d̻ t̺ t̺ʰ d̺ k kʰ ɡ kʷ kʷʰ ɡʷ/ <p ph b t th d t̠ t̠h d̠ k kh g kw khw gw>
/m m̥ n̻ n̻̊ n̺ n̺̊ ȵ ȵ̊ ŋ ŋ̊ ŋʷ ŋ̊ʷ/ <m mh n nh n̠ n̠h ň ňh ŋ ŋh ŋw ŋhw>
/f v s̻ z̻ s̺ z̺ ɕ ʑ x ɣ xʷ ɣʷ/ <f v s z s̠ z̠ š ž x ġ xw ġw>
/l̻ l̺ j/ <l l̠ j>

/i e a o u/ <i e a o u>
/iː eː aː oː ɤː uː/ <í é á ó ɤ́ ú>

(C)V(F), where F = [p t̻ t̠ k m n̻ n̠ ȵ ŋ s̻ s̠ x l̻ l̠]

opipik wrote:/i e a o u/ <i e a o u>
/iː eː aː oː ɤː uː/ <í é á ó ɤ́ ú>

What, baby-gamma in long only? If it's only ever long as well, why the need for the additional (no doubt hard to produce) diacritic?

Is this better?

/iː eː aː oː ɤː uː/ <í é á ó ǿ ú>

I think Frislander is trying to say that you shouldn't use an accent mark at all.

The thing is using a diacritic is more consistent, and hence I agree with its use.

I agree consistency is important. I'd keep the accent. I definitely would use ø instead of ɤ, lest someone would assume you meant [ø]. That said, I'd find it surprising that a language would have ɤ as the only back unrounded vowel, contrasting with a rounded vowel at the same POA.


JAL

Isn't that what happens in Mandarin? Mandarin has barred-i, too, but no other back unrounded vowels, I don't think.

The thing that's bothering me here is, why analyze that back unrounded vowel as phonemically long, and why on Earth would you have a long phonemic vowel without a short one contrasting with it? I think I could believe that it happens to be phonetically long in this conlang, but that still doesn't mean it's underlyingly long or something.

Having /ɤ/ as the only back unrounded vowel in the system assumes now that the vowels cling tightly to their narrow IPA labels. I would assume that in a real and balanced system with the vowel values /a e i o u ɤ/ both /a/ and /ɤ/ would be somewhat centralised, /a/ probably ranging from central back to central front and /ɤ/ from back to central. Both front rounded and back unrounded vowels tend to be anyway more centralised than their front unrounded and back rounded counterparts.

A possible realistic development path to a /a(:) e(:) i(:) o(:) u(:) ɤ:/ vowel system could be from a standard /a e i o u ə/ system with length. The original short /ə/ could be dropped or merged to other vowels in all positions while at least some cases of the long /ə:/ could have been strengthened to yield /ɤ:/.

Vijay wrote:The thing that's bothering me here is, why analyze that back unrounded vowel as phonemically long

Well, if a language has a short/long contrast, then I would find it rather strange not analyzing all phonetically long vowels as phonemically long. It also depends on the diachronics, of course.

gach wrote:Having /ɤ/ as the only back unrounded vowel in the system assumes now that the vowels cling tightly to their narrow IPA labels. I would assume that in a real and balanced system with the vowel values /a e i o u ɤ/ both /a/ and /ɤ/ would be somewhat centralised

Wouldn't you rather choose /ʌ/ or /ə/ to represent it though?


JAL

jal wrote:Well, if a language has a short/long contrast, then I would find it rather strange not analyzing all phonetically long vowels as phonemically long.

Persian says hi.

Vijay wrote:The thing that's bothering me here is, why analyze that back unrounded vowel as phonemically long, and why on Earth would you have a long phonemic vowel without a short one contrasting with it? I think I could believe that it happens to be phonetically long in this conlang, but that still doesn't mean it's underlyingly long or something.

If the length distinction is important to the phonology of a language and there's some vowel quality that only appears long, I'd see it preferable to keep the length formalism intact and still describe the vowel as phonemically long. Saying that lengths is phonemic for some vowels but only an additional afterthought for others doesn't feel very consistent. Even if some of these pairless long vowels would appear short in some predictable shortening environments, while being long otherwise, I'd classify them as underlyingly long.

Here's one real world example. The Mansi of middle Konda had the short vowels /ɒ o æ e~i~ə y/ (5 in total) and long vowels (incl. diphthongs) /o: u: ɵa ɤ: æ: e:~i: y:/ (7 in total). This is a case of a long /ɤ:/ without a short phonemic counterpart. The long vowels /u: ɤ:/ had also shortened allophones, but the long versions were the primary forms. Looking at the system, labelling these two as unspecified for length would just be silly.

jal wrote:Wouldn't you rather choose /ʌ/ or /ə/ to represent it though?

Possibly, though choosing what IPA to use for labelling your phonemes is just a notational convenience. I wouldn't read too much into it. The choice of single symbols isn't anyway going to be enough for describing sufficiently well how your phonemes behave and are realised.

How did that Mansi system develop?

I can't say exactly for that dialect. The Ob-Ugric vowel history is in general a mess and there's a lot of interdialectal variation. You could for example look at some classic literature on the subject. Here are the four relevant pages from Kálmán's The history of Ob-Ugric languages [PDF] (from Sinor ed. 1988, The Uralic languages). Sammallahti's Historical phonology of the Uralic languages [PDF] (from the same edition) also includes a treatment of the Ob-Ugric vowel history (pp. 502-510), but annoyingly his reflex matrix for the Mansi dialects ignores quantity (he also states back then that "a re-evaluation of the vowel systems in the Vogul dialects would be highly desirable", so make of that what you will). The disparate accounts aren't all easy to bring together, but you can conclude that the low labial diphthong comes from older /æ:/ and the long non-low unrounded central or back vowel seems to have its origins in PFU/PU.

Chweyawa, a language spoken by the Mwitsem, humanoid creatures from another dimension that lack lips and whose vocal cords are in the nasal passage rather than the larynx.

/n ɳ ɲ ŋ ɦ̃/
/t ʈ c k/
/s ʂ ç x h/

/i i: ɯ ɯ: e̞ e̞: ɤ̞ ɤ̞: a a:/
NOTE: all of these vowels are voiceless

Allophony:
/VhV/ becomes [V:]
/ɤ̞ ɤ̞:/ become [ɰ̊] intervocalically (except before /ɯ/) and between a consonant and a vowel (except before /ɯ/)
/ɯ ɯ:/ become [ɰ̊] intervocalically and between a consonant and a vowel
/e̞ e̞:/ become [j̊] intervocalically (except before /i/) and between a consonant and a vowel (except before /i/)
/i i:/ become [j̊] intervocalically and between a consonant and a vowel

The language completely lacks number, gender, and tense and the only pronoun is māňso which is their word for themselves. However, the sentient ones who have an individual sense of self will instead refer to themselves and others by their names.
The language makes no distinction between adjectives and nouns and majority of the words are created by compounding.
Example: āūščusōīňtaek "tree" = āūšču "thing" + sōīň "tall" + taek "green"

Consonants:

Nasals: /m n ŋ/
Plosives: /p pʰ b t tʰ d k kʰ g k͡p ɡ͡b/
Affricates: /t͡s t͡sʰ t͡ʃ/
fricatives: /f s ʃ x~h/
sonorants: /ʋ ɾ~r l j/

[h] is the allophoneme of /x/ in syllable initial position; [r] is the allophoneme of /ɾ/ after vowels.

Consonants can be geminated, although /ʃ/ and /x/ are usually not geminated; besides, the geminated /ʋ/ is [v:], the geminated /ɾ/ is [r:].

Arguably, the aspirated plosives /pʰ tʰ kʰ t͡sʰ/ are actually allophonemes of the combinations /ph th kh t͡sh/.

Vowels:

short vowels: /a ə ɛ ɪ ɔ ʊ/
long vowels: /ɑ: ɛ: e: i: o: u:/

Syllable: (C)(C)(C)(C)(C)V(C)(C)(C)(C), at least up to five consonants can start a syllable and four consonants can be in the coda position, the syllable structure is perhaps more elaborate.

Inspired by Rotokas:

voiceless: /p t k/
voiced: /w~b ɾ~d j~g/

vowel: /a e~i o~u/

syllable structure: (C)V

----------------------------------------------------------------

I suggested this for someone else, but it is repealed, at least the consonant system is repealed:

nasals: /m n ŋ/
plosives: /p pʰ ɓ t tʰ ɗ k kʰ ʔ/
affricates: /ts tsʰ tɬ tɬʰ/
fricatives: /s z ɬ x ɣ/
sonorants: /w r l j/

vowels: /a ɛ e i ɔ o u/

there's an ATR harmony between /ɛ ɔ/ and /e o/

syllable structure: (C)(R)V(P), P is any of the nasals /m n ŋ/ or unaspirated voiceless plosives /p t k/, R is any sonorants /w r l j/

Some of you may remember a while back I asked a question in the C&C quickies section about a sci-fantasy gender system, later including a couple of short texts set in the world (a first draft of one is in the quote thread). This language is meant to be spoken in that world.

/p t t͡ʃ k ʔ/ <p t c k '>
/ⁿb~m ⁿd~n ⁿɖʳ~ɳ ⁿd͡ʒ~ɲ ⁿg~ŋ/ <mb~m nd~n ndr~nr nj ng>
/s~ʂ/ <s>
/w ɽ j/ <w r y>

/i e a o aɪ̯ aʊ̯ ɔɪ̯ ea̯ oa̯/ <i e a o ai ao oi ea oa>

The vowels and diphthongs occur in long (acute), short and nasalised (grave) versions.

The prenasalised stops are pronounced as their pure nasal allophones before the nasal vowels.

The /s/ phoneme freely varies with its retroflex counterpart, though the retroflex is more likely in certain contexts: when another retroflex consonant (<n(d)r> or <r>) is already present in the word; and with a few verbs where <s> mutates to <r> intervocalically.

Phonotactics are strictly CV, with roots being mostly CVCV, though there are a few CV roots in the language.

There is a marginally phonemic pitch accent, which is not marked in the orthography. This falls on the first vowel of the root(s) which make(s) up a word, ignoring affixes but including reduplication. In a few verbs this distinguishes the habitual (formed by initial CV- reduplication) from one of the infixes.

EDIT: Example text
/ⁿbaːɪ̯ kaː mãʊ̯ⁿbeːki toːɪ̯ⁿgeɽipomõa toːɪ̯ⁿgeɽi ⁿd͡ʒõ ⁿd͡ʒokaʊ̯. Sõa̯ ⁿbaːɪ̯ kaː ‘iːpoɽõ ‘iːsatẽⁿd͡ʒetaːɪ̯ wiː ‘iːjajaɽamõa̯./
<Mbái ká mãombéki tóingeripomõa tóingeri njõ njokao. Sõa mbái ká ‘íporõ ‘ísatẽnjetái wí ‘íyayaramõa.>

Go to Wikipedia and type the name of the accent and you'll find precomposed characters. Google Docs also has an insert special character feature, which allows you to find characters by Unicode value, by scrolling through character sets, by typing a search or by drawing an approximation.