Random phonology/phonemic inventory thread

[ɶʙ ɞʛ]

I'm not sure about this, but I seem to remember that no languages distinguishing a front and a central /a/ is precisely the reason why no separate symbols for the two exist in IPA. However, a language that does distinguish them may have been discovered since the invention of the IPA.

The Hamont dialect of Limburgish has previously been cited on this forum as a language that distinguishes front, central and back low vowels.

Even if it does, you can just use /æ a ɑ/.

[Vlürch]

Just a random phonology I came up with and thought I'd post since I doubt I'll do anything with it:

/m n ŋ/
/p t k/
/pˀ tˀ kˀ/
/t͡s/
/s z/
/ʪ ʫ/
/ɸ β j x ɣ/
/h̪͆ ɦ̪͆/
/r/

/ə ɨ/
/á é í ó ú/
/à è ì ò ù/
/áː éː íː óː úː/
/àː èː ìː òː ùː/

With short vowels, tone is only contrastive on the penultimate syllable in polysyllabic words, the final syllable in bisyllabic words, and in monosyllabic words; elsewhere, pitch is allophonic. Long vowels can only occur in the first syllable of a word, but with them tone is always contrastive. /ə ɨ/ can only occur in the initial syllabe of a word (or the second syllable if the initial syllable has a long vowel) and their pitch is always purely allophonic.

/p t k/ are [pʰ tʰ kʰ] word-initially and when geminated; only homorganic stop clusters are allowed intervocalically.

/pˀ tˀ kˀ/ are [pʼ tʼ kʼ] word-initially and after voiceless consonants, [ɓ ɗ ɠ] after voiced consonants and [ʔ̚ɓ ʔ̚ɗ ʔ̚ɠ] intervocalically.

/ɸ β/, /x ɣ/ and /h̪͆ ɦ̪͆/ are only contrastive word-initially; only /β ɣ ɦ̪͆/ occur intervocalically and after voiced consonants while only /ɸ x h̪͆/ occur after voiceless consonants.

/r/ is [ɾ] intervocalically.

Intervocalically, in addition to homorganic nasal-plosive clusters, /mt ŋp ŋt/ occur and are pronounced [n͡mtʰ~mp̚tʰ ŋ͡mpʰ~ŋk̚pʰ ŋ͡ntʰ~ŋk̚tʰ]. Also, /mt͡s ŋt͡s/ occur and are [n͡mt͡sʰ~mp̚t͡sʰ ŋ͡nt͡sʰ~ŋk̚t͡sʰ].

The intervocalic clusters /rp rt rk/ are [rb~r̝̊p̚pʰ rd~r̝̊t̚tʰ rg~r̝̊k̚kʰ].

Only /m n ŋ p t k r/ are allowed in coda, but the clusters /mp nt ŋk rp rt rk/ do occur word-finally; they tend to be [mbᵊ ndᵊ ŋgᵊ rbᵊ rdᵊ rgᵊ], the very short vocalic release's pitch being variably whatever creates the smoothest transition to the next word.

A few random meaningless words:
/sóːʪì/ [só̞ːʪ̠ʲì]
/tˀənáŋ/ [tʼə̀ná̠ŋ]
/ɸúːŋzúmi/ [ɸúːŋzúmʲí]
/kˀakˀú/ [kʼà̠ʔ̚ɠú]
/kɨpˀarèɦ̪͆e/ [kʰɯ̽ʔ̚ɓá̠ɾè̞ɦ̪͆è̞]
/áːsərámp/ [á̠ːsə̀ɾámbᵊ]
/h̪͆imízi/ [h̪͆ìmʲíʑí]
/ɨmt͡sòrko/ [ɨ̀mp̚t͡sʰór̝̊k̚kʰó]
/ùːppˀəɣupˀùʫuŋ/ [ùːp̚pʼə̀ɣùʔ̚ɓùʫùŋ]
/ìːməβéna/ [ìːmə̀βé̞ná̠]
/tùːɦ̪͆ú/ [tʰùːɦ̪͆ú]

...so yeah, yet another kitchen sink phonology. Obviously it's not naturalistic, since bidental fricatives and lateral sibilants are very rare sounds and implosives aren't all that common either so all three existing in the same language would be insanely unlikely, but I could imagine it coming about naturally if /h̪͆ ɦ̪͆/ developed from /*f *v/ to make them more distinct from /ɸ β/ and /ʪ ʫ/ developed from /*sl *zl/ clusters or something while /*l/ otherwise merged into /r/ or whatever. More natural sound changes from that situation would probably be /*ɸ *β/ -> /p b/ and /*sl *zl/ -> /ɬ ɮ/, but eh.

[Birdlang]

/m n ŋ ŋʷ/ m n ŋ ŋʷ
/p b t d k g kʷ gʷ/ p b t d k g q g̊
/f v s z ʃ ʒ x ɣ xʷ ɣʷ h hʷ/ f v s z ś ź x ğ xʷ ğʷ h hʷ
/ʦ ʣ ʧ ʤ/ c j ć ʒ
/l ɹ j ʎ ɥ w ɰ/ l ż j ĺ ŷ w ŵ
/r/ <r>

/i ɨ u ɯ̈ e ə o ɛ ʌ ɔ ɐ a/ i ï u ü e ë o è ö ò a

[Creyeditor]

Just a random phonology I came up with and thought I'd post since I doubt I'll do anything with it:

/m n ŋ/
/p t k/
/pˀ tˀ kˀ/
/t͡s/
/s z/
/ʪ ʫ/
/ɸ β j x ɣ/
/h̪͆ ɦ̪͆/
/r/

/ə ɨ/
/á é í ó ú/
/à è ì ò ù/
/áː éː íː óː úː/
/àː èː ìː òː ùː/

With short vowels, tone is only contrastive on the penultimate syllable in polysyllabic words, the final syllable in bisyllabic words, and in monosyllabic words; elsewhere, pitch is allophonic. Long vowels can only occur in the first syllable of a word, but with them tone is always contrastive. /ə ɨ/ can only occur in the initial syllabe of a word (or the second syllable if the initial syllable has a long vowel) and their pitch is always purely allophonic.

/p t k/ are [pʰ tʰ kʰ] word-initially and when geminated; only homorganic stop clusters are allowed intervocalically.

/pˀ tˀ kˀ/ are [pʼ tʼ kʼ] word-initially and after voiceless consonants, [ɓ ɗ ɠ] after voiced consonants and [ʔ̚ɓ ʔ̚ɗ ʔ̚ɠ] intervocalically.

/ɸ β/, /x ɣ/ and /h̪͆ ɦ̪͆/ are only contrastive word-initially; only /β ɣ ɦ̪͆/ occur intervocalically and after voiced consonants while only /ɸ x h̪͆/ occur after voiceless consonants.

/r/ is [ɾ] intervocalically.

Intervocalically, in addition to homorganic nasal-plosive clusters, /mt ŋp ŋt/ occur and are pronounced [n͡mtʰ~mp̚tʰ ŋ͡mpʰ~ŋk̚pʰ ŋ͡ntʰ~ŋk̚tʰ]. Also, /mt͡s ŋt͡s/ occur and are [n͡mt͡sʰ~mp̚t͡sʰ ŋ͡nt͡sʰ~ŋk̚t͡sʰ].

The intervocalic clusters /rp rt rk/ are [rb~r̝̊p̚pʰ rd~r̝̊t̚tʰ rg~r̝̊k̚kʰ].

Only /m n ŋ p t k r/ are allowed in coda, but the clusters /mp nt ŋk rp rt rk/ do occur word-finally; they tend to be [mbᵊ ndᵊ ŋgᵊ rbᵊ rdᵊ rgᵊ], the very short vocalic release's pitch being variably whatever creates the smoothest transition to the next word.

A few random meaningless words:
/sóːʪì/ [só̞ːʪ̠ʲì]
/tˀənáŋ/ [tʼə̀ná̠ŋ]
/ɸúːŋzúmi/ [ɸúːŋzúmʲí]
/kˀakˀú/ [kʼà̠ʔ̚ɠú]
/kɨpˀarèɦ̪͆e/ [kʰɯ̽ʔ̚ɓá̠ɾè̞ɦ̪͆è̞]
/áːsərámp/ [á̠ːsə̀ɾámbᵊ]
/h̪͆imízi/ [h̪͆ìmʲíʑí]
/ɨmt͡sòrko/ [ɨ̀mp̚t͡sʰór̝̊k̚kʰó]
/ùːppˀəɣupˀùʫuŋ/ [ùːp̚pʼə̀ɣùʔ̚ɓùʫùŋ]
/ìːməβéna/ [ìːmə̀βé̞ná̠]
/tùːɦ̪͆ú/ [tʰùːɦ̪͆ú]

...so yeah, yet another kitchen sink phonology. Obviously it's not naturalistic, since bidental fricatives and lateral sibilants are very rare sounds and implosives aren't all that common either so all three existing in the same language would be insanely unlikely, but I could imagine it coming about naturally if /h̪͆ ɦ̪͆/ developed from /*f *v/ to make them more distinct from /ɸ β/ and /ʪ ʫ/ developed from /*sl *zl/ clusters or something while /*l/ otherwise merged into /r/ or whatever. More natural sound changes from that situation would probably be /*ɸ *β/ -> /p b/ and /*sl *zl/ -> /ɬ ɮ/, but eh.

I just wanted to stress again how much I appreciate random phonologies that are more than a phoneme inventory.

[DesEsseintes]

/pˀ tˀ kˀ/ are [pʼ tʼ kʼ] word-initially and after voiceless consonants, [ɓ ɗ ɠ] after voiced consonants and [ʔ̚ɓ ʔ̚ɗ ʔ̚ɠ] intervocalically.

Love this.

[Birdlang]

/m n/ m n
/p b t d k g/ p b t d k g
/f v s z ʃ ʒ x h/ f v s z š ž x h
/ʦ ʣ ʧ ʤ/ c j č ǰ
/l j w/ l y w
/r/ r

/i iː u uː e eː o oː a aː/ I ī u ū e ē o ō a ā

[DesEsseintes]

Oniihek phonology: smol language I came up with because I can’t sleep

/n/ n
/t t͡ʃ k ʔ/ t c k ’
/s h/ s h
/r w/ r w

/a e i o/ a e i o
/ai/ ai

/a i o/ have long counterparts, written doubled
/ai/ is the long counterpart to /e/
/e/ is also the epenthetic vowel

CC clusters are common and largely unrestricted
’CC clusters are also permissible
→ nʔC rʔC wʔC clusters are formed when first C is a voiced continuant

Aesthetic:

on’weriirek
hwan’seriirek
o’htewer
aire’haat
iicrawek → iicrawkew
ckoorwak

[Birdlang]

/m n ɲ ŋ/ m n ņ/ñ/ny~ni~n ŋ/ŋ/nh
/p b t d k g ʔ/ p b t d k g ʼ/h/q
/ɸ β s z ʃ ʒ x~h ɣ/ f v s z ç/š/sy~si~s ʒ/ž/zy~zi~z ķ/x/h ģ/ď/gh~dd~g
/ʦ ʣ ʧ ʤ/ c/c/c ż/ź/dz ŧ/č/ty~ti~t đ/x/dy~di~d
/l ʎ~ɟ~ʝ j~ʝ~ʑ ɰ~w~ɣʷ/ l y/j/y~ll~i j/y/j~ź~y w/w/gu~gw~w~u
/ɾ r/ r ŗ/ŕ/rr~r
/ɮ/ ź/ŀ/zl

/i ɯ u ʊ e ɤ o ə ɛ ʌ ɔ æ a~ɑ~ɐ ɒ/ i ï/û/ú u ö/õ/ù e/é/é ü/ã/ò o/ó/ó è/â/è é/e/e ä/ŭ/à ó/o/o æ a ò/ă/á
/Vː/ Vh/V̄/VV

This is Birdish updated. The first letter is Modern Standard, second is Central, and third is Coastal.

[DesEsseintes]

I’m wreaking havoc with Híí diachronics again, and here is my latest sketch for Middle Híí, the last common ancestor of Híí Proper and Húú.

Code: Select all

*m  *n
    *t  *ts *tł *tʃ *k  *kw *ʔ
        *s  *ł  *ʃ  *x  *xw *h
    *ð      *l  *j  *ɣ  *w

*a  *e  *i

*ð is back, and though it largely merges with *j, there are also reflexes in dd in Húú.

I’ll post the latest version of Húú if it becomes anything more than a scribbled mess.

[Birdlang]

/m n ŋ/ m n ŋ
/p b t̟ d̟ t̪ d̪ c ɟ k g ʔ/ p b ṭ ḍ t d ḱ ǵ k g '
/f v θ̟ ð̟ s̪ z̪ ʃ ʒ ʕ h/ f v ṣ ẓ s z ś ź ğ h
/l ʎ j ɥ ʟ ɰ w/ l ĺ y ẃ ł ý w
/r/ r

/i y ɨ u e ø ə o ɛ œ ɜ ʌ ɔ ɑ/ i ü ï u e ö ë o è ȍ ä ȅ ò a
/i u e o ɛ ɔ ɑ/ > /ɨ y ə ø ʌ œ ɜ/

[DesEsseintes]

/m n/ m n
/t t͡ʃ d͡ʒ k g ʔ/ t c j k g ’
/s/ s
/v r/ v r
/a e o/ a e o

Permissible clusters are nC vC rC ’C.
Permissible geminates are thus mm nn vv rr but all are underlyingly clusters.

Vowels can occur in hiatus ad nauseam.

A voicing contrast in dorsal occlusives!? Feels so wrong and yet so right...

t c k only occur in clusters word-medially.

[Nortaneous]

I'm not sure about this, but I seem to remember that no languages distinguishing a front and a central /a/ is precisely the reason why no separate symbols for the two exist in IPA. However, a language that does distinguish them may have been discovered since the invention of the IPA.

The Hamont dialect of Limburgish has previously been cited on this forum as a language that distinguishes front, central and back low vowels.

Also various dialects of English - some AmE has PRICE monophthongization but doesn't merge this with FATHER, giving [æ a ɑ] for TRAP, PRICE, FATHER; other AmE apparently has the TRAP-BATH split with [a] for BATH, so [æ a ɑ] for TRAP, BATH, FATHER.

A voicing contrast in dorsal occlusives!? Feels so wrong and yet so right...

/p t k g/ is attested - mostly I think in Papuan where /b d/ are more prone to lenition, although some cases of this in PHOIBLE have mistranscribed /ɣ/.

The Rotokas consonant inventory is sometimes given as /p t k β ɾ g/. But /g/ is pretty rare in Rotokas. Buin is a clearer case. It's also been claimed for Mekeo, but many things have been claimed for Mekeo.

There's also /b t k g/, attested in Bwaidoka, but there are other stops - the full inventory is /b bʷ t k g kʷ/.

Speaking of Papuan languages, another thing you can do is lenite all non-glottal plosives except in ʔ_ N_, analyze the lenited forms as basic, and get a plosive inventory of /ʔ/. The consonants of Ontena Gadsup are /ʔ ɸ β s x m n r j/.

[Ser]

Ontena Gadsup

I... I-... I don't even.

I HAVE BEEN LIVING A LIE

[opipik]

There's also /b t k g/, attested in Bwaidoka, but there are other stops - the full inventory is /b bʷ t k g kʷ/.

The tables in the Organized Phonology Data papers aren't accurate - they often omit prenasalized stops and sometimes they contain even more glaring errors. Elsewhere in that paper it's stated that Bwaidoka has a /d/ as well (which isn't surprising, seeing as all other Austronesian languages in Milne Bay have /b t d k g/ or /p b t d k g/.
Onobasulu, on the other hand, has /b d k g/.
Buin seems to have /p t k g/, but I recall reading (in another source) that /g/ can be pronounced [ɣ].

[VaptuantaDoi]

Some ideas for the smallest possible inventory for a Romlang while still looking and sounding normal. A couple of possible routes:

1. Dalmatian (15)
/p t k b d g m n f s l r/
/i u a/

Spoiler:

Dalmatian is already overloaded with /i u a/. The vowel breaking results were /ẹ ọ ę ǫ/ → /ai̯ au̯ i̯a u̯a/; unstressed vowels were either raised or lowered. The numbers one to ten: /i̯un dau̯ tri kutru sanki si̯as sipti u̯aptu nu di/.

2. Iberian (17)
/p t k β~b d~ð g m n s h l ɾ/
/i u e o a/

Spoiler:

The results or palatalisation were /s/ for most voiceless consonants (dz z dʒ → ts s ts → s) and /h/ for most voiced consonants (ts s tʃ → tʃ ʃ tʃ → x → h). Numbers one to ten are /uno duo tɾes ku̯atɾo hinke sei̯s sete oho nu̯eβe di̯ehe/.

3. Gallo-Romance (14)
/m n p t d k s v l ʁ/
/i ɛ o a/

Spoiler:

When /u/ fronted, /o/ was not raised even after /y/ merged with /i/. The first paragraph of the Wikipedia page on Dalmatian:

Dalmaçian oe Dalmatic (Dalmaçian: langa dalmata, dalmato; Italian: lingua dalmatica, dalmatico) es oen Lenva Romanç obsolett qui af se favlar eon ve reioena de Dalmaçia, eon Croaçia modern, e toetment austral eon Cattar eon Monteneir. Ve noemn indicatt oen trib de ves formaçioen linguistic, ves Dalmates. Ve diaveitt Rausan, ve diaveitt cu ploes de estudi, af esdr va lenva offiçial de ve Repuvlic de Rausa par mult de va histoeri medieval, estac el lentment af se sopplantar de lenvas altres.

/daːmaˈsan oː daːmaˈti (daːmaˈsan: laŋga dalmata, dalmatə; italiˈan: lingu̯a dalmatika dalmatikɔ) ɛː oː ˈlɛnva ʁoˈmans osoˈlɛt ki aː sɛ favˈlaː ɛː vɛ ʁiˈoːna dɛ daːˈmasa ɛː kʁoːˈasa moˈdɛː, ɛ toːˈmɛnt oːˈtʁaː ɛː kaˈtaː ɛː moːtɛˈnai̯. vɛ noːm indiˈkat oː tʁiː dɛ vɛː foːmaˈsoː lingisˈti, vɛː daːmaˈtɛː. vɛ diaˈvai̯t ʁai̯ˈsan, vɛ diaˈvai̯t ki ploː dɛ ɛːtiˈdi, aː ɛːd va ˈlɛnva ofiˈsaː dɛ vɛ ʁɛpiˈvli dɛ ʁai̯sa paː miːt dɛ va iːtoːˈʁi mɛdiɛˈvaː, ɛːˈta ɛː lɛntˈmɛnt aː sɛ soplanˈtaː dɛ lɛnˈvaː aːˈtʁɛː/

4. Southern Romance (17)
/m n p t k v d g f s l ɾ/
/i u e o a/

Spoiler:

Sample: /unu ɾuo tɾis battɾu kimbi sis seti ottu novi deki/