Random phonology/phonemic inventory thread [2011–2018]

A forum for all topics related to constructed languages
User avatar
Omzinesý
mongolian
mongolian
Posts: 4079
Joined: 27 Aug 2010 08:17
Location: nowhere [naʊhɪɚ]

Re: Random phonology/phonemic inventory thread

Post by Omzinesý »

p t k
b d g
m n ŋ
s ʃ
z ʒ
l ʟ
r

Semivowels
ʋ j

and prestopped resonants:
bm dn gŋ
dl gʟ
dr

vowels
i u
ɛ ɑ

vowels can be short or long. There are three phonations: normal, creaky, and breathy voice.

Syllable:
(C)(semivowel)V(liquid or /n/ or /ŋ/)

Coda nasals nasalize the preceding vowel like in Mandarin.

Writing:
There are different symbols for onset and coda consonants, so syllable patterns are clear. There are also markers for semivowels vowel length and phonation. Vowel qualities are not normally marked.
My meta-thread: viewtopic.php?f=6&t=5760
Porphyrogenitos
sinic
sinic
Posts: 401
Joined: 21 Jul 2012 08:01
Location: Buffalo, NY

Re: Random phonology/phonemic inventory thread

Post by Porphyrogenitos »

gestaltist wrote: 04 Jan 2018 10:57 I get p > /N_ but why would /g/ get devoiced? It doesn't make much sense unless there is some context to this I'm missing.
The second rule is also a bit weird. It's probably g p > ɣ ɸ > h h, right? I guess it makes sense in a minimalist phonology with few fricatives. But I'm still a bit iffy about it being unconditional.

r w > g b makes more sense to me with some intermediate stages. Maybe r w > l w > ʟ w > ɡ b
g > k /N_ and g > h, except with /ħ ~ x ~ h/ instead of just /h/, is actually what happens with the Galician gheada IRL. I don't know the exact phonetic motivation for it, but I assume that it's something like, as [g] exits the phonology, remaining [g] preserved by preceding nasals are merged into the closest similar phoneme.

And yes, g p > h h would probably involve some intermediate steps, but shifts like that are straightforward and common enough that I didn't feel the need to list any. Likewise with r w > g b, but both are attested as more or less unconditional changes, so I also just listed them as simple changes.
User avatar
gestaltist
mayan
mayan
Posts: 1617
Joined: 11 Feb 2015 11:23

Re: Random phonology/phonemic inventory thread

Post by gestaltist »

Porphyrogenitos wrote: 04 Jan 2018 17:00
gestaltist wrote: 04 Jan 2018 10:57 I get p > /N_ but why would /g/ get devoiced? It doesn't make much sense unless there is some context to this I'm missing.
The second rule is also a bit weird. It's probably g p > ɣ ɸ > h h, right? I guess it makes sense in a minimalist phonology with few fricatives. But I'm still a bit iffy about it being unconditional.

r w > g b makes more sense to me with some intermediate stages. Maybe r w > l w > ʟ w > ɡ b
g > k /N_ and g > h, except with /ħ ~ x ~ h/ instead of just /h/, is actually what happens with the Galician gheada IRL. I don't know the exact phonetic motivation for it, but I assume that it's something like, as [g] exits the phonology, remaining [g] preserved by preceding nasals are merged into the closest similar phoneme.

And yes, g p > h h would probably involve some intermediate steps, but shifts like that are straightforward and common enough that I didn't feel the need to list any. Likewise with r w > g b, but both are attested as more or less unconditional changes, so I also just listed them as simple changes.
Thanks for the explanation. Do you perhaps remember where you found attestation for r > g?
User avatar
Frislander
mayan
mayan
Posts: 2088
Joined: 14 May 2016 18:47
Location: The North

Re: Random phonology/phonemic inventory thread

Post by Frislander »

This is the phonology of my latest conlang project.

/(p) t (t͡ɬ) (t͡ʃ) k q/ <(p) t (tl) (ts) k q>
/pː tː t͡ɬː t͡ʃː kː qː/ <pp tt tl ts kk qq>
/b d d͡ʒ g/ <b d j g>
/bː dː d͡ʒː gː/ <bb dd jj gg>
/ɬ ʃ h/ <l s h>
/m n ɲ ŋ/ <m n nj ng>
/w/ <w>

Single /p t͡ɬ t͡ʃ/ only occur in the clusters /mp nt͡ɬ nt͡ʃ/.

/i u/ <i u>
/ə/ <ə>
/ɛ ɑ/ <e a>

/˥ ˨ ˧˨ˀ/ <a á a'>

There are no tone contrasts on the reduced vowel /ə/, which generally functions as an unstressed epenthetic vowel.

Syllable structure is (C)(w)V(n)(C), where the glide /w/ does not occur after /w ɲ/ or before the vowel /u/. Coda consonants are restricted to /t k q b d d͡ʒ g ɬ ʃ h m n ɲ ŋ w/ and clusters are restricted to /mp nt nt͡ɬ nt͡ʃ ŋk ɴq/. Furthermore when a coda stop (including when in a cluster) would occur before another consonant a schwa /ə/ is inserted. Furthermore coda nasals assimilate to the POA of a following obstruent. As an additional restriction word-initial consonants are restricted to /t k q b d d͡ʒ g ɬ ʃ h w/, again of which all but the last two may be followed by /w/.

There are a series of consonant mutations which operate on the set of word-initial consonants. There are two mutations: the T- and N-mutations. These are summarised on the table below:

Code: Select all

  T  N
h pp mp
t tt nt
k kk nk
q qq nq
b bb m
d dd n
j jj nj
g gg ng
s ts nts
l tl ntl
0 t  n
Clio
sinic
sinic
Posts: 228
Joined: 27 Dec 2012 23:45

Re: Random phonology/phonemic inventory thread

Post by Clio »

A more finished version of the phonology that I started sketching here:

/m n/ <m n>
/pː tː cː kː/ <p t kj k> ("fortis" plosives)
/p t c k/ <b d gj g> ("lenis" plosives)
/ts̪ ts̠/ <tz ts> ("fortis" affricates)
/s̪ s̠/ <z s> ("lenis" fricatives)
/l ʎ j/ <l lj j>
/i e a u o ʌ (ɨ)/ <i e a o u ë (ë)>
/iː eː aː uː oː ʌː/ <ii ee aa uu oo ëë>
/ei ai ui ou ʌi/ <ei ai ui ou ëi>
/˥ ˧˥ ˥˩/ (unwritten)

/ɨ/ occurs as an epenthetic vowel to break up illegal consonant clusters and as an allophone of toneless /ʌ/. When consonant clusters are broken up, the coda of the leftmost syllable is maximized.

Up to two syllables in a polysyllabic word can carry tone, and tone is minimally contrastive. It is possible to determine the pitch of every syllable in a word by applying at most three rules after all epenthetic vowels have been inserted; for these rules, diphthongs pattern with long vowels, and a syllable is heavy if and only if it either contains a long vowel or ends in a consonant cluster (where the fortis consonants are considered singletons). Words of one and two syllables require only rule (1), as do words of three syllables with a light penult; other words require all three rules.
  1. The first syllable of every word is accented. If the vowel in that syllable is short, it takes high tone. If the vowel is long, it can take either rising or falling tone.
  2. If the penult is heavy, then it also is accented: a short vowel receives high tone, and a long vowel takes falling tone.
  3. If the penult is light, then it is accented: a short vowel receives high tone, and a long vowel take rising tone.
After all the tones in a word are assigned, then the tones of adjacent syllables are adjusted as follows:
  1. If two vowels with rising tone are adjacent, the first tone becomes falling; similarly, if two vowels with falling tone are adjacent, the first tone becomes rising.
  2. If a high tone is followed by a falling tone, the first vowel loses its tone.
Otherwise, all tones remain as they were.

Long vowels do not occur before fortis consonants, but they can occur before clusters.

The maximal syllable has the form CCVCCC. Permitted word-initial clusters are: /mn/; any plosive followed by /l/ or /n/; and any of /p(ː) k(ː)/ followed by /t(ː)/. Word-finally, clusters can consist of: /l ʎ n/ followed by any plosive, affricate, or fricative; /m/ followed by a plosive; a fricative followed by a plosive; or /l ʎ n/ followed by a fricative then a plosive. Between vowels, clusters can be any of the above. In addition, a fricative may precede any cluster that would be legal word-initially; and /l ʎ m n/ may follow any plosive or fricative, even if they are preceded by one of /l ʎ m n/.

Within clusters and in adjacent syllables, coronals harmonize right-to-left between the classes /t(ː) (t)s̪ l/ and /c(ː) (t)s̠ ʎ/. /n/ also has an allophone [ɲ] before a consonant of the second set (but not in an adjacent syllable).

Before /k(ː)/, /m/ has an allophone [ŋ].

Lenis consonants undergo voicing in word-initial position, in intervocalic position, and in clusters with voiced consonants such that /p t c k s̪ s̠/ become [b d ɟ g z̪ ɻ] respectively. Fortis consonants undergo a different form of weakening in word-initial position and in all clusters such that /pː tː cː kː ts̪ ts̠/ become [p t c k s̪ s̠] respectively.
Niûro nCora
Getic: longum Getico murmur in ore fuit
scratchpad
User avatar
Vlürch
greek
greek
Posts: 452
Joined: 09 Mar 2016 21:19
Location: Finland
Contact:

Re: Random phonology/phonemic inventory thread

Post by Vlürch »

I kinda like this vowel inventory, but I'm not sure if I'll ever make even a quick shitty start of a conlang that uses it.

/ɑ ə i ʊ/
/aˑ ɛˑ ɤˑ ɞˑ/
/ɑː æː eː iː ɵː uː/
User avatar
eldin raigmore
korean
korean
Posts: 6352
Joined: 14 Aug 2010 19:38
Location: SouthEast Michigan

Re: Random phonology/phonemic inventory thread

Post by eldin raigmore »

Vlürch wrote: 08 Jan 2018 08:20 I kinda like this vowel inventory, but I'm not sure if I'll ever make even a quick shitty start of a conlang that uses it.

/ɑ ə i ʊ/
/aˑ ɛˑ ɤˑ ɞˑ/
/ɑː æː eː iː ɵː uː/
What do you like about it?
User avatar
Frislander
mayan
mayan
Posts: 2088
Joined: 14 May 2016 18:47
Location: The North

Re: Random phonology/phonemic inventory thread

Post by Frislander »

The proto-lang:

*p *t *c *k *ʔ
*h
*m *n *ŋ
*l

*i *u
*e *a

Syllable structure was CV(*ʔ, *h, *n), where syllable codas only appeared word-finally. The language was agglutinative-isolating with disyllabic roots.

The daughter-lang:

/p t tɬ tʂ tʃ k ʔ/
/p’ t’ tɬ’ tʂ’ tʃ’ k’/
/ⁿb ⁿd ⁿdɮ ⁿdʐ ⁿdʒ ⁿg/
/ɓ ɗ/
/f s ɬ ʂ ʃ x h/
/m n ŋ/
/m̥ n̥/
/l ɽ j w/

/i u/
/e ɤ o/
/ɛ ɑ/

Vowels may occur in nasal, creaky or breathy voice.

Syllable structure is CV.

The language is strongly polysynthetic.

Example words:

*late > ɽe
*ʔakuh > k’o̤
*himun > m̥ũ
*niʔaʔ > ɗɤ̰
*kuten > tʂẽ
*melu > ⁿdɮo
*tike > te
User avatar
Vlürch
greek
greek
Posts: 452
Joined: 09 Mar 2016 21:19
Location: Finland
Contact:

Re: Random phonology/phonemic inventory thread

Post by Vlürch »

eldin raigmore wrote: 08 Jan 2018 23:57
Vlürch wrote: 08 Jan 2018 08:20 I kinda like this vowel inventory, but I'm not sure if I'll ever make even a quick shitty start of a conlang that uses it.

/ɑ ə i ʊ/
/aˑ ɛˑ ɤˑ ɞˑ/
/ɑː æː eː iː ɵː uː/
What do you like about it?
I'm not sure, but I guess the idea of three vowel lengths with a distinct quality as mid-length and "splitting up" is appealing to me. You know, like the short and long /ɑ/ and /ɑː/ being back while the half-long one is front /aˑ/, the half-long /ɛˑ/ being intermediate in quality between the long /æː/ and /eː/, etc.

Something like this could also be interesting to work with:

/ɑ ɘ ɯ/
/ʌˑ ɔˑ ɛˑ œˑ iˑ yˑ/
/ɒː ɵː uː/

...ie. something where short vowels are unrounded while their long counterparts are rounded (or vice versa), and half-long ones are both (in addition to having some quality difference).
User avatar
DesEsseintes
mongolian
mongolian
Posts: 4331
Joined: 31 Mar 2013 13:16

Re: Random phonology/phonemic inventory thread

Post by DesEsseintes »

/m n/
/t k ʔ/
/h/
/r w/

/a i/

Circumflexes should feature somehow.
User avatar
Thrice Xandvii
runic
runic
Posts: 2698
Joined: 25 Nov 2012 10:13
Location: Carnassus

Re: Random phonology/phonemic inventory thread

Post by Thrice Xandvii »

That looks surprisingly familiar, Des...
Image
User avatar
DesEsseintes
mongolian
mongolian
Posts: 4331
Joined: 31 Mar 2013 13:16

Re: Random phonology/phonemic inventory thread

Post by DesEsseintes »

Thrice Xandvii wrote: 12 Jan 2018 18:27 That looks surprisingly familiar, Des...
You mean to say that it looks like all my other inventories? If so, you’re probably right. [:)]
User avatar
Thrice Xandvii
runic
runic
Posts: 2698
Joined: 25 Nov 2012 10:13
Location: Carnassus

Re: Random phonology/phonemic inventory thread

Post by Thrice Xandvii »

Didn't you post that in the chat as well? Cuz that's what I meant.
Image
User avatar
Adarain
greek
greek
Posts: 511
Joined: 03 Jul 2015 15:36
Location: Switzerland, usually

Re: Random phonology/phonemic inventory thread

Post by Adarain »

DesEsseintes wrote: 12 Jan 2018 14:31 /m n/
/t k ʔ/
/h/
/r w/

/a i/

Circumflexes should feature somehow.
Obviously /k ʔ/ ⟨k k̂⟩ ← that’s a combining diacritic, this forum’s font is bad.
At kveldi skal dag lęyfa,
Konu es bręnnd es,
Mæki es ręyndr es,
Męy es gefin es,
Ís es yfir kømr,
Ǫl es drukkit es.
User avatar
DesEsseintes
mongolian
mongolian
Posts: 4331
Joined: 31 Mar 2013 13:16

Re: Random phonology/phonemic inventory thread

Post by DesEsseintes »

Adarain wrote: 13 Jan 2018 10:36
DesEsseintes wrote: 12 Jan 2018 14:31 /m n/
/t k ʔ/
/h/
/r w/

/a i/

Circumflexes should feature somehow.
Obviously /k ʔ/ ⟨k k̂⟩ ← that’s a combining diacritic, this forum’s font is bad.
Interesting. Thanks for the idea, but I’m almost certainly going to be using â î.
Thrice Xandvii wrote: 13 Jan 2018 09:13 Didn't you post that in the chat as well? Cuz that's what I meant.
Yes, I posted it in both places at the same time.
User avatar
Frislander
mayan
mayan
Posts: 2088
Joined: 14 May 2016 18:47
Location: The North

Re: Random phonology/phonemic inventory thread

Post by Frislander »

Make of this what you will.

The proto-language

*p *t *k
*s
*m *n *ŋ
*w *j

*e *a *o
*z̩ *ɣ̩ *v̩

Syllable structure is CV. The syllabic fricatives only occur after obstruents.

The daughter

/p t t͡s k/
/x h/
/m n ŋ/
/w ɻ j/

/i iː u uː/
/eː oː/
/a aː/

Syllable structure is (C)V.

Sound changes:

*p, *t, *k, *s > ɸ, t͡s, x, ʐ before syllabic fricatives.
*z̩, *ɣ̩, *v̩ > i, a, u
e, o > i, u
ɸ > 0
ai, ia > e:
au, ua > o:
iu, ii > i:
ui, uu > u:
aa > a:
s > h
User avatar
Frislander
mayan
mayan
Posts: 2088
Joined: 14 May 2016 18:47
Location: The North

Re: Random phonology/phonemic inventory thread

Post by Frislander »

Proto-language

*p *t~ɾ *k
*s *h
*m *n

*i *a *u

*t was realised as a flap intervocalically.

Syllable structure was CV(C), where any consonant could appear in either the onset or coda.

Sound changes:

*p, *t > *w, *j before another consonant.
*h > 0
*k > ʔ
*ɾ > *ʒ / _i
*s > *ʃ / _i
*t > ts / _i
*t > k
*s > h
V > 0 / VC_CV, VC_#
Vowel - coda consonant resolution:

Code: Select all

   w   j    ʔ    s  m    n
i (j)u i   (j)e  s̩ (j)õ (j)ẽ
a  o   e    a    a  õ    ẽ
u  u  (w)i (w)o  s̩ (w)õ (w)ẽ
ẽ > a
õ > o
*ʃ > s
*ʒ > ɻ

The daughter language:

/p t͡s k ʔ/ <p c k ‘>
/s h/ <s h>
/m n/ <m n>
/w ɻ j/ <w r y>
/r/ <rr>

/i u/ <i u>
/e o/ <e o>
/a s̩/ <a s>

Syllable structure is (C)V(C), where any consonant may occur in the onset or coda, with the additional restriction that rhotics may not occur word-initially. When the two rhotics come into contact the glide is deleted.
User avatar
Creyeditor
MVP
MVP
Posts: 5091
Joined: 14 Aug 2012 19:32

Re: Random phonology/phonemic inventory thread

Post by Creyeditor »

I thought about a phonology were every segment corresponds to a semantic primitive in Natural Semantic Metalanguage. This means I need about 50 segments. I also thought that prosodic constituents, like words or syllable correspond to sentences or clauses in NSM. This means that probably every segment should have a vowel and a consonant allophone, so that it easy to build syllables. This is the basic vowel inventory.

/i u/
/e ə o/
/ɛ a ɔ/

Every vowel can occur in different versions, exemplified for /i/. Nasalized, breathy voiced, long vowels all contrast.
/i iː ĩ ĩː i̤ i̤ː ĩ̤ ĩ̤ː/

Consonants allophones of each segment are determined by the phonological features. Every vowel position corresponds to a place of articulation.
/i/ = palatal
/e/ = dental
/ə/ = alveolar
/ɛ/ = velar
/u/= bilabial
/o/ = retroflex
/ɔ/ = uvular/pharyngeal
/a/ = glottal

Long vowels have continuant/resonant allophones, nasalized vowels have prenasalized stops/nasal stops allophones, breathy vowels have voiced consonant allophones. And I am getting tired, so no more details...
Creyeditor
"Thoughts are free."
Produce, Analyze, Manipulate
1 :deu: 2 :eng: 3 :idn: 4 :fra: 4 :esp:
:con: Ook & Omlűt & Nautli languages & Sperenjas
[<3] Papuan languages, Morphophonology, Lexical Semantics [<3]
User avatar
gestaltist
mayan
mayan
Posts: 1617
Joined: 11 Feb 2015 11:23

Re: Random phonology/phonemic inventory thread

Post by gestaltist »

Creyeditor wrote: 22 Jan 2018 22:50 I thought about a phonology were every segment corresponds to a semantic primitive in Natural Semantic Metalanguage. This means I need about 50 segments. I also thought that prosodic constituents, like words or syllable correspond to sentences or clauses in NSM. This means that probably every segment should have a vowel and a consonant allophone, so that it easy to build syllables. This is the basic vowel inventory.

/i u/
/e ə o/
/ɛ a ɔ/

Every vowel can occur in different versions, exemplified for /i/. Nasalized, breathy voiced, long vowels all contrast.
/i iː ĩ ĩː i̤ i̤ː ĩ̤ ĩ̤ː/

Consonants allophones of each segment are determined by the phonological features. Every vowel position corresponds to a place of articulation.
/i/ = palatal
/e/ = dental
/ə/ = alveolar
/ɛ/ = velar
/u/= bilabial
/o/ = retroflex
/ɔ/ = uvular/pharyngeal
/a/ = glottal

Long vowels have continuant/resonant allophones, nasalized vowels have prenasalized stops/nasal stops allophones, breathy vowels have voiced consonant allophones. And I am getting tired, so no more details...
I feel like it'd be easier to find 50 consonant phonemes and have vowel placement be epenthetic and predictable. A cool idea BTW.
User avatar
Vlürch
greek
greek
Posts: 452
Joined: 09 Mar 2016 21:19
Location: Finland
Contact:

Re: Random phonology/phonemic inventory thread

Post by Vlürch »

Something ugly, once again...

/m n̪ ɳ ɲ ŋ/
/n̪͡m ɳ͡m ɲ͡m ŋ͡m/
/p b t̼ d̼ t̪ d̪ ʈ ɖ k g q ʔ/
/pː bː t̼ː t̪ː ʈː kː qː/
/pʰ t̼ʰ t̪ʰ ʈʰ kʰ qʰ/
/pʰː t̼ʰː t̪ʰː ʈʰː kʰː qʰː/
/ʰp ʰt̼ ʰt̪ ʰʈ ʰk ʰq/
/ʰpː ʰt̼ː ʰt̪ː ʰʈː ʰkː ʰqː/
/pʼ t̼ʼ tʼ kʼ qʼ/
/pʼː t̼ʼː tʼː kʼː qʼː/
/t̪͡p d̪͡b ʈ͡p ɖ͡b k͡p g͡b k͡t g͡d/
/t͡pʼ k͡pʼ q͡t̪ʼ q͡ʈʼ/
/p͡s p͡ʃ t̪͡s̪ d̪͡z̪ ʈ͡ʂ ɖ͡ʐ t͡ɕ d͡ʑ k͡s k͡ʃ/
/p͡s̪ʰ p͡ʂʰ t̪͡s̪ʰ ʈ͡ʂʰ t͡ɕʰ/
/p͡ɸ b̪͡v d̪͡ð k͡x q͡χ/
/p̪͡fʼ k͡xʼ q͡χʼ/
/s z ʃ ʒ/
/sː ʃː/
/s̪ʰ ʂʰ/
/s̪ʰː ʂʰː/
/ʰʃː/
/sʼ/
/f v θ̼ θ ɻ j x ɣ h/
/fː ð̼ː/
/ɾ ʀ̆/
/r̪ ɽ͡r̥ ʀ̥ ʢ/
/r̼ː r̪̊ː ʀː/
/b͡ʙ d͡r ɢ͡ʀ/
/t͡r̥ʼ q͡ʀ̥ʼ ʡ͡ʜʼ/
/p͡ɬ t̪͡ɬ̪ ʈ͡ɭ̝̊ c͡ʎ̝̊ k͡ʟ̝̊/
/p͡ɬʰ t̪͡ɬ̪ʰ ʈ͡ɭ̝̊ʰ c͡ʎ̝̊ʰ k͡ʟ̝̊ʰ/
/p͡ɬʼ t͡ɬʼ k͡ʟ̝̊ʼ/
/ɬ̪ l ɭ ɭ̝̊/
/ɬ̪ː ɭ̝̊ː/
/ɬʼː/

/à ɛ̀ ɔ̀/
/á í ú/
/āː ēː ōː/
Locked