Post your conlang's phonology

[WeepingElf]

A new phonology, which is a terrible rip off of something (mostly, sorta). Guess what it is!

Oh you. You even got the guts to include /ðʲ/.

Vowel harmony:

(not sure how to do vowel harmony with this vowel compliment)

That shouldn't be too hard. There are the natural pairs a~ä u~ü and ï~i. The remaining vowels e and o could be either neutral or determine the vowel harmony class of the word uniquely. Based on your example they are already rare, so whatever you choose to do shouldn't be much of a problem. You could also go for some labial harmony, but that would make the system quite a bit more complicated with these vowels.

The original had vowel harmony, but only four vowels in non-first syllables: a, ä, ï and i - or, if analyzed differently, only /a/ and /i/, each with a front and a back allophone according to the first vowel.

[Nortaneous]

Vowel harmony:

(not sure how to do vowel harmony with this vowel compliment)

easy. i/ï, ü/u, ä/a, then say e/o used to be split but ë merged into something (a?) and ö became e, so e/a and e/o.

alternatively, i is neutral and ï is the back equivalent of e.

Ćïngušïtïngïpïđućï šäpüdi pü lixü tüđi püđüpim. or
Ćïngušitingïpiđućï šäpüdi pü lixü tüđi püđüpem.

[Bristel]

OK, not sure if I got this right, or understand fully how VH works, but...

i~ï (I) ü~u (U) ä~a (A), neutral: e and o

tiš-ke-mun-tä > *tiš-ke-mün-tä
tï-đü-mïn-gïn-so-dü > *tïn-đu-mïn-gïn-so-dü

How's that?

Doesn't a neutral vowel "block" the harmonizing from preceding or following vowels depending on the type of spreading?

[Nortaneous]

I think /i/ is more likely to be neutral than /e/, but you could probably get away with that. Neutral vowels don't necessarily block harmony, but /e o/ could force front and back harmony respectively.

Note that I changed the examples to make them more illustrative.

Transparent neutral vowels:
tiš-ko-mUn-tA > tiškomüntä
tïn-đU-mIn-gIn-se-dU > tïnđumïngïnsedu

Opaque neutral vowels:
tiš-ko-mUn-tA > tiškomunta
tï-đU-mIn-gIn-se-dU > *tïnđumïngïnsedü

I'd go with opaque to better make sense of the fact that *mid* vowels are neutral.

[Basilius]

[e o] can be neutral e. g. if they result from recent mergers.

aj = äj > e
aw = äw > o

Or something.

[Bristel]

[e o] can be neutral e. g. if they result from recent mergers.

aj = äj > e
aw = äw > o

Or something.

I used [e o] as neutral vowels, but I haven't decided fully, so I've only updated a few things on my post that were missing. (allophony of plosives)

[zyxw59]

My current proto-language:
Consonants

Code: Select all

             Labial   Alveolar Velar    Labiovelar Glottal
Nasal        m        n        ŋ        ŋʷ
VL Plosive   p        t        k        kʷ         ʔ
V Plosive    b        d        g        gʷ
Implosive    ɓ        ɗ        ɠ        ɠʷ
VL Fricative          s                            h
V Fricative           z
Approximant           l  r     j        w

Vowels
The vowels are unspecified for backness and roundedness, and differ only by height (is this plausible?)
/i~ɨ~u/
/e~ɘ~o/
/ɛ~ɜ~ɔ/
/æ~a~ɑ/
Syllable
Onsets:
C, C + fricative, C + approximant
word initially only: nothing
Nuclei:
V, V + approximant
Codas:
nothing, C
Forbidden Clusters
labiovelar + w
w + labiovelar
approximant + approximant
word initial fricative + fricative
glottal + glottal

[Bristel]

Very cool. I like it. (and yes, afaik, that vowel height distinction is plausible, but is likely to have allophonic processes with glottal (?) consonants for backing, or the labiovelar series rounding the vowels)

[zyxw59]

Very cool. I like it. (and yes, afaik, that vowel height distinction is plausible, but is likely to have allophonic processes with glottal (?) consonants for backing, or the labiovelar series rounding the vowels)

yeah, I'm doing a lot of that in the daughter langs.

[gach]

I used [e o] as neutral vowels, but I haven't decided fully, so I've only updated a few things on my post that were missing. (allophony of plosives)

You can do a lot of tricks with vowel harmony so feel free to play around a bit.

I'll add a few real world examples of transparent vs. opaque neutral vowels. In Finnish we have the two neutral vowels /e i/. On the first (stressed) syllable they group with front vowels by prototypically triggering front vowel harmony (etä, itä) while on further (unstressed) syllables they freely coexist with any vowel and let both front and bask vocalisms pass through unaltered: kapea vs. häpeä and ojitus vs. ylitys.

Hill Mari has the same neutral vowels /e i/, but whenever these occur, they trigger front harmony that spreads rightwards from them. Thus you get words like ošemənäm and kugilä that start with back vowels but end in front ones.

Yet another possibility that hasn't been covered is to do things a bit like Chukchi does. It has a dominant-recessive harmony where adding any affix containing some of the dominant vowels /a o e/ causes all recessive vowels /e u i/ to change into their dominant counterparts: e > a, u > o, i > e (note that /e/ in included in both of the sets; these are underlyingly two totally different vowels). Inspired by this you could have a system where /o/ is a back vowel without a front counterpart and /e/ a front vowel with no back counterpart. Adding either of these onto a word would then be enough to trigger back and front vowel harmony respectively on the entire word so that

tiš-ko-mUn-tA > tïškomunta but
tiš-ke-mUn-tA > tiškemüntä

You'd still have to deal with what happens when both /e/ and /o/ occur within the same word. The two most likely solutions would be that one is dominant over the other (likely /o/ over /e/) or, since both of the vowels appear to be rare, just treat each of the cases irregularly. A likely alternation that would fit into either of the cases would be /e/ backing into /a/ when forced to include a back feature.

Don't worry if you don't take any of these ideas. Just writing them down has given me ideas of how to spice up my own dull vowel harmonies.

The vowels are unspecified for backness and roundedness, and differ only by height (is this plausible?)
/i~ɨ~u/
/e~ɘ~o/
/ɛ~ɜ~ɔ/
/æ~a~ɑ/

It's unusual but attested.

[Bristel]

SNIP

I really like the Chukchi idea. If the vowel harmony affects the entire word, that could be an interesting way of having words become very different very quickly in daughter languages, even though they are the same root and possibly similar morphology.

[CatDoom]

I've been tinkering around with Riktas and I've pretty much overhauled the phonology. I've got some ideas about a system of consonant mutation more complex than what I had before, so I figured I'd post it and see if it makes sense to anybody else.

Consonants

Nasal /m, n, ŋ/
Plosive /p, b, p', t̪, d̪, t̪', ʈ, ɖ, ʈ', k, g, k', ʔ/
Affricate /t͡s, d͡z, t͡s', t͡ʃ, (d͡ʒ), t͡ʃ'/
Fricative /ɸ, β, s, z, ʃ, (ʒ), h/
Approximant /l, j, w/
Tap /ɾ/

/ʒ/ and /d͡ʒ/ are marginal allophones of /z/ and /d͡z/ occurring in diminuitive forms (see below).

Glottalized stops and affricates were inherited from a linguistic substrate, occur most frequently in loanwords. However, they have been "naturalized" and introduced into the core vocabulary of Riktas Rammay based on onomatopoeia and sound symbolism. In particular, voiceless stops have become ejectives in diminuitive nouns and many verb roots describing sudden, brief, and often involuntary actions.

In addition, Riktas distinguishes between "short" and geminiated forms of its nasals, pulmonic (non-ejective) stops, affricates, and /l/ between vowels.

Vowels

/i, u, a/

No phonemic vowel length distinctions.

Phonotactics

The canonical syllable is CV(C). Onset consonants are required and the glottal consonants /ʔ, h/ never occur as codas. Nasal consonants never occur in word final position, and word final /t̪/ and /tʲ/ are replaced with the open syllables /t͡si/ and /t͡ʃi/, respectively.

Consonant Mutation

Many Riktas consonants occur in regularly alternating pairs of fricative and plosive allophones, specifically:

/ɸ, β/ ~ /p, b/
/s, z/ ~ /t͡s, d͡z/
/ʃ, ʒ/ ~ /t͡ʃ, d͡ʒ/
/h/ ~ /k/
/ɾ/ ~ /d/

(/ɾ/ isn't a fricative, and /h/ is a borderline case, but they follow a similar pattern of allophony)

The fricative allophones are considered primary, and with the exception of /k/, which may contrast with /h/ in any position, the plosive allophones may only occur in noun stems and particles as the second segment of a consonant cluster following a nasal, fricative, or liquid (/l/ or /ɾ/), for instance in the words /t̪aɸdaj/, "above" and /sumpa/, "mountain, high island."

Many roots in Riktas can form stems for nouns or verbs, while others are normally exclusively nominal but may take one of several possible verbalizing suffixes. When these roots are inflected as verbs, its consonants uniformly shift to their secondary, plosive form. For instance, when the noun /haɸa/, "father," takes the verbalizing suffix /-t̪i/, the resulting verb stem is realized as /kapat̪i/, "to be or behave as a father."

Roots that normally form exclusively verbal stems are somewhat less regular in form; in at least some conjugations they may contain consonant segments in their "primary" form, though complementary allophones still won't co-occur except in clusters, as described above. However, when a verb stem is inflected with a nominalizing aspect, all consonants will be realized as their primary, non-plosive allophones. For instance, when the nominaizing suffix /-zmi/ is applied to the verb stem /pat͡ʃi/, "to freeze, to be icy cold," it becomes /ɸaʃizmi/, "frozen one, cold one." Note that, in this case, the resulting word must still obey the rule regarding consonant clusters, and that geminated stops and affricates in verb stems are treated as a cluster of two identical consonants, such that the verb stem /ʔit͡ʃːu/, "to go out" corresponds to the noun stem /ʔiʃt͡ʃu/, "an outing."

Diminuitive Forms

Riktas nouns, and less frequently verbs, may take on a "diminuitive" form, which may express smallness, insignificance, affection, or contempt, depending on the context in which it is used and the particular word affected. The diminuitive is marked by a set of regular consonant shifts in the stem, which are as follows:

/l/ /n/
/s/ /ʃ/
/z/ /ʒ/
voiced plosives voiceless
voiceless plosives ejective

Nouns may also take the diminutive suffix /-niʃ/, but in many cases the sound shifts alone are sufficient. For instance, the word for "dog" is /guk/, while either /kuk'/ or /kuk'niʃ/ may be used for "puppy."

Edit: Changed notation for postalveolar stops from /tʲ, dʲ, t'ʲ/ to /ʈ, ɖ, ʈ'/. The contrast between these and the corresponding dental stops is essentially the same as in Hindustani, with the "retroflex" series being more specifically apical-postalveolar, sometimes shifting allophonically to become alveolar.

[Tropylium]

ɠʷ

Sexy.

[CatDoom]

Here's what I've got for my most recent project, tentatively called Mita, which is essentially a fanciful "reconstruction" of Pre-Proto-Chumash.

Consonants

nasal /m n/
stop /p t k q ʔ/
fricative /s/
approximant /l j w/

/s/ varies fairly freely between [s] and [ʃ], with the notable restriction that all sibilants in a word will generally have the same place of articulation

Geminate clusters of stops or fricatives (e.g. [pp], [ss]) created through morphological processes are resolved as a single aspirated consonant ([pʰ], [sʰ]). A similar pattern is found in non-stop consonants, where the second identical consonant will become a voiceless glottal transition [h] (e.g. [mm] > [mh], [ll] > [lh], etc).

A similar process occurs in clusters of any consonant with a glotal stop, which are resolved as the glottalized allophones [ˀm ˀn p' t' k' q' s' ˀl ˀj ˀw]. Because these allophones appear only in consonant clusters formed at morpheme boundaries they are not considered to be phomnemic.

Vowels

Four vowel qualities with no phonemic distinctions in thength: /a i o u/

The vowel system is essentially square, with binary distinctions in closeness and backness. The pronunciation of /o/ is usually fairly close to [ɒ], but all of the vowels permit a fairly wide range of pronunciations.

Mita has a fairly strict system of vowel harmony. In particles, any of these vowels may appear relatively freely, save that vowels of similar closeness must also conform in backness. So CiCo could occur, but CaCo would become CoCo.

The vowels in nouns and verbs are governed by a productive system of ablaut in addition to the above vowel harmony rule. Vowels in stems are always either /a/ or /o/, depending on the inflection, while vowels in prefixes may be /i/ or /u/.

Phonotactics

The following syllable structures are permitted: V, CV, VC, CVC, CCV, VCC, CCVC, and CVCC.

Within word roots, consonant clusters are limited to (Stop or Nasal) + /s/, (Stop or nasal) + Approximant, or two dissimilar stops, and never include /ʔ/. The order of consonants in a root cluster conforms relatively strictly to the sonority scale, so that one finds the root snV but not nsV.

[Ser]

Yet another inventory. Comes with a list of possible diphthongs and hiatuses (which are not really very many).

[Ambrisio]

Nama minus the clicks:

m n
p ~ β t ~ ɾ k ʔ
t͜sʰ k͜xʰ
s x h
/i e a o u/ and nasal /ĩ ã ũ/

<m n p~b t~d k ' c q s x h i e a o u ĩ ã ũ>

I think I should also try ~Xoon minus the clicks (and strident vowels). Sounds great for a European-ish conlang. The Nama one has a nice Polynesian feel to it.

A few example words:

nii 'one'
teda 'two'
sema 'three'
soe 'four'
qee 'five'
hade 'six'
hade 'e nii 'seven'
...

Conversely, Arabic plus clicks could yield a very interesting phoneme inventory for a conlang!

[CatDoom]

ʔoslehmos/C'ihuzi, a language I'm working on over on the Akana wiki, has the following inventory:

Consonants:

Voiceless plosive /p t k ʔ/
Aspirated plosive /pʰ tʰ kʰ/
Ejective plosive /p' t' k'/
Voiced plosive /b d g/

Voiceless affricate /t͡s t͡ʃ/
Ejective affricate /t͡s' t͡ʃ'/
Voiced affricate /d͡z d͡ʒ/

Plain nasal /m n ŋ/
Glottalized nasal /ˀm ˀn ˀŋ/

Voiceless fricative /s ʃ h/
Voiced fricative /z ʒ ɦ/

Plain flap /ɾ/
Glottalized flap /ˀɾ/

Plain approximant /l j w/
Glottalized approximant /ˀl ˀj ˀw/

Vowels

/a, ɒ, e, o, i, u/

All vowels may be phonemically short or long. Vowels in compounds and inflected stems observe height harmony, assimilating to the height of the vowel in the stressed syllable of the stem.

[Nortaneous]

Considering redoing Miar and adding a proto-stage.

Proto-Miar:
/p b t d c j k g/
/f~pʰ s~tʰ ɕ~cʰ x~kʰ/
/m n ñ ŋ/
/r l y/
/: º/
/h/
/a ə i u/ + syllabic nasals and /r l/ (and /y~w/)

/:/ is a chroneme that can attach to either initial consonants or vowels. No two adjacent syllables can have the chroneme.
/º/ is a backing suprasegmental that applies on the syllable level: it turns alveolars to dentals, palatals to retroflexes, and high vowels to mid. Spreads to high vowel of preceding syllable, which also becomes mid, if non-labial consonant between them.

Roots typically consist of two consonants, with an optional chroneme or backing suprasegmental, and with a sonorant optionally preceding the second consonant. Affixes typically consist of C(:)(R)(V)(: º).

All vowels neutralize to schwa before the stress, which typically appears on the tense-marker but can move backwards or (along with an added chroneme) forwards. Roots typically take schwa as the vowel, but other vowel infixes can occur. Schwa deletes before a vowel or consonant that can become syllabic.

Example verb: n-yt 'to walk'
xºə-m-ya:-nəyt χḿ̩ya:nit 'I am walking'
cə-ñ-kə-n<a>yt cñ̩́kn̩ayt 'you habitually walked long distances'
ml̩-y-cʰaº-gwə-ya:-n<u>yt mlec̣ʰɑgwiya:nuyt 'I will not make him walk'

Don't know if I'm actually going to do this but w/e here it is

[clawgrip]

I'm in the initial stages of designing a new logographic script, but, like Japanese or Akkadian, I want the script to be used for a language other than the one it was designed for. As a result, I need to design a kind of skeleton language to base the script on before exporting it to another language. What I've got for the original language's phonology is this:

Code: Select all

Consonants:
             stops              nasals      trills    fric  approximants
             asp    unvoc  voc  unvoc  voc  reg  stop       unvoc  voc
labial       pʰ      p      b     m̥     m                    w̥      w
dental/etc.  tʰ      t      d     n̥     n    r    t͡r   s     j̊      j
velar        kʰ      k      g                          x     

Vowels:
             i   u
             æ   ɑ

Syllable structure is simple:

(C)V(/n, r, s, x, w, j/)

I figure there may be some sort of tone or phonation or something on the vowels that is not recorded in the script. This way I can get more apparent homophones to assist in rebus writing and so on.

Phuta khey yay tun akumu.
pʰutɑ kʰæj jɑj tun ɑkumu
The dog bit the man's leg.

[2+3 clusivity]

Clawgrip, for a phonology of that size and diversity, it seems a bit odd to not have a lateral. Especially when the only rhotic(s) have a trill rather than an approximant or flap~tap.

[clawgrip]

You may be right. Perhaps I will add one. The reason I did it was specifically to cause difficulty in adapting the writing system to another language, which does have a lateral. Maybe I will just screw up something else instead.

[Nortaneous]

nah it's fine. dahalo doesn't have /j/ outside one word

[ObsequiousNewt]

nah it's fine. dahalo doesn't have /j/ outside one word

What's the word?

[cromulant]

Clawgrip, for a phonology of that size and diversity, it seems a bit odd to not have a lateral. Especially when the only rhotic(s) have a trill rather than an approximant or flap~tap.

It's really not that big or diverse an inventory. On WALS it would score "Average" for Consonant Inventory; languages this size with no laterals abound. Juǀ'hoan has no laterals; the phonemic status of /l/ in Taa is debatable.

But you're right, it still is "a bit odd." Just not impossible/unattested by any stretch (not that you were saying it was).

nah it's fine. dahalo doesn't have /j/ outside one word

What's the word?

/jáːjo/ 'mother'.

[Nortaneous]

i would expect it to have /h/ though, are there any languages with aspirates or voiceless nasals and no /h/