Morneau phonetic alphabet

A forum for discussing linguistics or just languages in general.
Post Reply
User avatar
Coveny
rupestrian
rupestrian
Posts: 12
Joined: 17 Nov 2017 20:38
Contact:

Morneau phonetic alphabet

Post by Coveny »

Hello.

I posted this in the you thread but maybe this is a better place to post this.

I'm looking to get http://www.rickmor.x10.mx/segmental_phonemes.png translated into the current form of the IPA. http://www.rickmor.x10.mx/phonology.html is a link to the article.

Also has anyone else attempted to replicate Rick Morneau's work to find common phonetic sounds across major languages?

Thanks,
Mark
ForDebating.com is coming in February. A debating website that I hope to populate with intelligent critical debator. You can follow updates at https://www.facebook.com/fordebating

Feature comparison chart
User avatar
Creyeditor
MVP
MVP
Posts: 5091
Joined: 14 Aug 2012 19:32

Re: Morneau phonetic alphabet

Post by Creyeditor »

What exactly would be the criteria for identity of segments across languages? Identical phonetics (probably impossible), identically oppositions with other sounds (pretty language specific) or maybe some comparative definition of each segment? I think this is really not a trivial question.
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
Coveny
rupestrian
rupestrian
Posts: 12
Joined: 17 Nov 2017 20:38
Contact:

Re: Morneau phonetic alphabet

Post by Coveny »

Creyeditor wrote: 17 Nov 2017 23:23What exactly would be the criteria for identity of segments across languages?
Well Ricks criteria was "Choose phonemes that most people in the world already know or can learn to pronounce easily", and that is exactly what I would very much like to find.

Creyeditor wrote: 17 Nov 2017 23:23 Identical phonetics (probably impossible), identically oppositions with other sounds (pretty language specific) or maybe some comparative definition of each segment? I think this is really not a trivial question.
Well to be fair I didn't ask for anyone to reproduce Rick's work. I just can't find the old IPA symbols he used for the chart, and for someone who knows the old symbols and the new it would be trivial to convert the symbols over to the new IPA. For someone who's new to IPA, and conlangs though. If you aren't familiar with anyone else who's attempted to replicate the work then so be it, but if anyone with better knowledge knows of a second opinion on the matter I'd like to hear it. Rick's chart is referenced in several articles I've read.

Now on the topic of what how close the phonemes are to each other in real life between native speakers. I'm sure Rick allowed for a fair amount of leeway. He gives this disclaimer "Basically, what I did was eliminate phonemic feature distinctions that were made by only a small minority of languages. Aspiration went first, then nasalization, then labialization, etc. " in the article. I haven't gotten to the point where I understand that stuff, but hopefully Rick's words answered your question on his chart. :)
ForDebating.com is coming in February. A debating website that I hope to populate with intelligent critical debator. You can follow updates at https://www.facebook.com/fordebating

Feature comparison chart
User avatar
Creyeditor
MVP
MVP
Posts: 5091
Joined: 14 Aug 2012 19:32

Re: Morneau phonetic alphabet

Post by Creyeditor »

Oh, sorry, I misunderstood you there. Good luck for your search, I might not be of help [:S]
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
Sectori
hieroglyphic
hieroglyphic
Posts: 32
Joined: 07 Oct 2010 18:12
Location: Tkaronto

Re: Morneau phonetic alphabet

Post by Sectori »

I think these are almost all normal IPA, except the alveolar/palatal-alveolar fricatives; they're just hard to read because the image is grainy. in order through the consonants and then vowels, based on the languages involved, I'm pretty sure these are:

Consonants: /p/, /b/, /t/, /d/, /k/, /g/, /q/, /ʔ/ (i.e., glottal stop), /f/, /v/, /θ/, /ð/, /s/ or /ʂ/ (i.e., voiceless alveolar fricative or voiceless retroflex fricative), /z/ or /ʐ/ (i.e., voiced alveolar fricative or voiced retroflex fricative), /ʃ/ or /ɕ/ (i.e., voicelss palato-alveolar fricative or voiceless alveolo-palatal fricative), /ʒ/ or /ʑ/ (i.e., voiced palato-alveolar fricative or voiced alveolo-palatal fricative), /x/, /ɣ/, /ç/, /ħ/ (i.e., voiceless pharyngeal fricative), /ʕ/ (i.e., voiced pharyngeal fricative), /ts/ or /tʂ/ (presumably /ʈʂ/), /dz/ or /dʐ/ (presumably /ɖʐ/), /tʃ/ or /tɕ/, /dʒ/ or /dʑ/, /l/, /ʎ/, /w/, /j/, /m/, /n/, /ŋ/, /ɲ/, /r/, ??? or /ɻ/, /ʁ/ or /ʀ/, ????, /ɽ/??, /ŋ͡m/, /k͡p̚/, /ɡ͡b/

Notes
the ??? for the second rhotic sound is because Vietnamese doesn't seem to have any rhotics and Arabic and Chinese only have one each (/r/ and /ɻ/, respectively), so I don't know what the alternative sound is. it could potentially be /ɹ/, but then wouldn't English be there?

the ???? for the fifth rhotic sound is because Hausa does have two rhotic sounds (one of which is the retroflex flap, as below), but the other is /r/, which the chart marks it as not having. (also speaking of /r/, afaik the default rhotic in Japanese is /ɾ/, not /r/.)

/ɽ/?? is because Hindi-Urdu also has this sound.

Vowels: /i/, /y/, /ɨ/, /ɯ/, /u/, /ɪ/, /ʊ/, /e/, /ø/, /o/, /ə/, /ɛ/, /œ/, /ɔ/, /æ/, /ɐ/, /a/, /ɑ/
inida elish, er·jīse pan.
sheb olnezī, on zūl kaid
nyer maudem? māzeye gejegura,
ib·zhiyorī aur mādaresh; kep panī weram.
e pel zherokareshī, onyek ne rād:
izholen tekab. yerogim nyer.

(semar pel, i.1-6)
User avatar
Coveny
rupestrian
rupestrian
Posts: 12
Joined: 17 Nov 2017 20:38
Contact:

Re: Morneau phonetic alphabet

Post by Coveny »

Sectori wrote: 18 Nov 2017 20:04 I think these are almost all normal IPA, except the alveolar/palatal-alveolar fricatives; they're just hard to read because the image is grainy. in order through the consonants and then vowels, based on the languages involved, I'm pretty sure these are:

Consonants: /p/, /b/, /t/, /d/, /k/, /g/, /q/, /ʔ/ (i.e., glottal stop), /f/, /v/, /θ/, /ð/, /s/ or /ʂ/ (i.e., voiceless alveolar fricative or voiceless retroflex fricative), /z/ or /ʐ/ (i.e., voiced alveolar fricative or voiced retroflex fricative), /ʃ/ or /ɕ/ (i.e., voicelss palato-alveolar fricative or voiceless alveolo-palatal fricative), /ʒ/ or /ʑ/ (i.e., voiced palato-alveolar fricative or voiced alveolo-palatal fricative), /x/, /ɣ/, /ç/, /ħ/ (i.e., voiceless pharyngeal fricative), /ʕ/ (i.e., voiced pharyngeal fricative), /ts/ or /tʂ/ (presumably /ʈʂ/), /dz/ or /dʐ/ (presumably /ɖʐ/), /tʃ/ or /tɕ/, /dʒ/ or /dʑ/, /l/, /ʎ/, /w/, /j/, /m/, /n/, /ŋ/, /ɲ/, /r/, ??? or /ɻ/, /ʁ/ or /ʀ/, ????, /ɽ/??, /ŋ͡m/, /k͡p̚/, /ɡ͡b/

Notes
the ??? for the second rhotic sound is because Vietnamese doesn't seem to have any rhotics and Arabic and Chinese only have one each (/r/ and /ɻ/, respectively), so I don't know what the alternative sound is. it could potentially be /ɹ/, but then wouldn't English be there?

the ???? for the fifth rhotic sound is because Hausa does have two rhotic sounds (one of which is the retroflex flap, as below), but the other is /r/, which the chart marks it as not having. (also speaking of /r/, afaik the default rhotic in Japanese is /ɾ/, not /r/.)

/ɽ/?? is because Hindi-Urdu also has this sound.

Vowels: /i/, /y/, /ɨ/, /ɯ/, /u/, /ɪ/, /ʊ/, /e/, /ø/, /o/, /ə/, /ɛ/, /œ/, /ɔ/, /æ/, /ɐ/, /a/, /ɑ/
THANKS!!!!

So this would be the symbols per row.

Consonants:
Row 1 = /p/, /b/, /t/, /d/, /k/, /g/, /q/, /ʔ/
Row 2 = /f/, /v/, /θ/, /ð/, /s/ or /ʂ/, /z/ or /ʐ/, /ʃ/ or /ɕ/, /ʒ/ or /ʑ/, /x/, /ɣ/
Row 3 = /ç/, /ħ/, /ʕ/, /ts/ or /tʂ/, /dz/ or /dʐ/, /tʃ/ or /tɕ/, /dʒ/ or /dʑ/, /l/, /ʎ/
Row 4 = /w/, /j/, /m/, /n/, /ŋ/, /ɲ/, /r/, ??? or /ɻ/, /ʁ/ or /ʀ/, ????
Row 5 = /ɽ/??, /ŋ͡m/, /k͡p̚/, /ɡ͡b/

Vowels
Row 6 = /i/, /y/, /ɨ/, /ɯ/, /u/, /ɪ/, /ʊ/, /e/
Row 7 = /ø/, /o/, /ə/, /ɛ/, /œ/, /ɔ/, /æ/, /ɐ/, /a/, /ɑ/

Also when it uses the "or" does that mean that the two sounds are close enough that they sound the same to non-native speakers? Or are the sounds distinctive and he's just grouping them together for some reason? How can two different sounds overlap like that?
ForDebating.com is coming in February. A debating website that I hope to populate with intelligent critical debator. You can follow updates at https://www.facebook.com/fordebating

Feature comparison chart
User avatar
Sectori
hieroglyphic
hieroglyphic
Posts: 32
Joined: 07 Oct 2010 18:12
Location: Tkaronto

Re: Morneau phonetic alphabet

Post by Sectori »

Coveny wrote: 18 Nov 2017 21:15THANKS!!!!
np
Coveny wrote:Also when it uses the "or" does that mean that the two sounds are close enough that they sound the same to non-native speakers? Or are the sounds distinctive and he's just grouping them together for some reason? How can two different sounds overlap like that?
/ʃ ɕ/, /ʒ ʑ/, /tʃ tɕ/, and /dʒ dʑ/ I would say are probably close enough together that someone whose native language doesn't distinguish them might not hear the difference right off, and /ʁ ʀ/ are in variation with one another in various dialects of French, Portuguese, German, and I think Swedish, so that makes conceptual sense to me, but I have no idea why he'd group the retroflex sounds with the alveolar sounds — they sound totally different to me; as someone whose native language has no retroflex sounds except /ɻ/, I'd be way more likely to conflate /ʂ/ (e.g.) with /ʃ ɕ/ than with /s/.
inida elish, er·jīse pan.
sheb olnezī, on zūl kaid
nyer maudem? māzeye gejegura,
ib·zhiyorī aur mādaresh; kep panī weram.
e pel zherokareshī, onyek ne rād:
izholen tekab. yerogim nyer.

(semar pel, i.1-6)
User avatar
Coveny
rupestrian
rupestrian
Posts: 12
Joined: 17 Nov 2017 20:38
Contact:

Re: Morneau phonetic alphabet

Post by Coveny »

In listening to them, from the IPA chart I can't tell a difference in the "or" ones so that makes sense. (or at least as much sense as I can make of this stuff at my low level of understanding)

The upside down R sounds super strange to me...

From listening to the /ʂ/, /ʃ /, and /ɕ/ to my inexperienced ear they do sound the same. (although the /ʂ/ sounds a bit more breathy on the front)

Ok and /ŋ͡m/, /k͡p̚/, /ɡ͡b/ are two IPA blended together a sound?

So why do they name the place of articulation for vowel and consonants differently? It seems like the same location to me, and that it should be the same. Could you help a beginner understand the difference?

Location Vowels Consonats
Forward Labial Front
Middle Coronal Central
Back Dorsal Back

I understand that Consonants breaks location down further (Bilabial, Labio­dental, Dental, Alveolar, Postalveolar, Retroflex, Palatal, Velar, Uvular, Pharyngeal, Glottal), but all vowels do is define how wide your mouth is open. (Close, Close-mid, Open-mid, Open) Could you help a beginner understand the difference?

Do languages other than English have something like Anglophones?

Dude can I buy you a case of beer or something? You have been super helpful. I can Paypal you some money or something...
ForDebating.com is coming in February. A debating website that I hope to populate with intelligent critical debator. You can follow updates at https://www.facebook.com/fordebating

Feature comparison chart
User avatar
eldin raigmore
korean
korean
Posts: 6352
Joined: 14 Aug 2010 19:38
Location: SouthEast Michigan

Re: Morneau phonetic alphabet

Post by eldin raigmore »

Coveny wrote: 17 Nov 2017 23:13 .... has anyone else attempted .... to find common phonetic sounds across major languages?
You could look at UPSID Sound Selection and find all the phonemes that occur in more than, or at least, some fraction (say 50%) of their sample's languages;
for a more up-to-date notation and database with a bigger sample (between twice as big and five times as big, I think) you could look at PHOIBLE.
The 12 phonemes that UPSID says are in >=50% of their languages are among the 19 phonemes that PHOIBLE says are in >=50% of their (obviative 3rd person) languages.

UPSID's collection includes these nine consonants and three vowels:
Sound Description Sound occurs in % of UPSID's languages
m voiced bilabial nasal 94.24%
k voiceless velar plosive 89.36%
i high front unrounded vowel 87.14%
a low central unrounded vowel 86.92%
j voiced palatal approximant 83.81%
p voiceless bilabial plosive 83.15%
u high back rounded vowel 81.82%
w voiced labial-velar approximant 73.61%
b voiced bilabial plosive 63.64%
h voiceless glottal fricative 61.86%
g voiced velar plosive 56.10%
N voiced velar nasal 52.55% (UPSID doesn't use the eng ŋ IPA uses)

PHOIBLE's collection includes these fourteen consonants and five vowels:

Name Representation Description Segment class Combined class
m 2053/2155 (95%) LATIN SMALL LETTER M consonant c
k 2016/2155 (94%) LATIN SMALL LETTER K consonant c
i 1998/2155 (93%) LATIN SMALL LETTER I vowel v
a 1961/2155 (91%) LATIN SMALL LETTER A vowel v
j 1901/2155 (88%) LATIN SMALL LETTER J consonant c
u 1873/2155 (87%) LATIN SMALL LETTER U vowel v
p 1873/2155 (87%) LATIN SMALL LETTER P consonant c
w 1812/2155 (84%) LATIN SMALL LETTER W consonant c
n 1742/2155 (81%) LATIN SMALL LETTER N consonant c
s 1663/2155 (77%) LATIN SMALL LETTER S consonant c
t 1604/2155 (74%) LATIN SMALL LETTER T consonant c
b 1533/2155 (71%) LATIN SMALL LETTER B consonant c
o 1459/2155 (68%) LATIN SMALL LETTER O vowel v
e 1458/2155 (68%) LATIN SMALL LETTER E vowel v
l 1421/2155 (66%) LATIN SMALL LETTER L consonant c
h 1405/2155 (65%) LATIN SMALL LETTER H consonant c
ɡ 1380/2155 (64%) LATIN SMALL LETTER SCRIPT G consonant c
ŋ 1217/2155 (56%) LATIN SMALL LETTER ENG consonant c
d 1161/2155 (54%) LATIN SMALL LETTER D consonant c

/n s t o e l d/ are in half or more of PHOIBLE's sampled languages but in less than half of UPSID's sampled languages.

In any case these are a good guess to the phonemes that more than half of the world will "recognize".

Don't know if that helps?

I once posted a consonant-phoneme inventory created by taking every place-of-articulation at which half or more of UPSID's languages had a pulmonic egressive consonant, and every manner-of-articulation in which half or more of UPSID's languages had a pulmonic egressive consonant, and looking at the Cartesian product.
The thing is, most of the rows and most of the columns had at most one or two consonants in them that actually occurred in 50% or more of UPSID's languages.
Most of the consonants in my post occurred in around 1% or fewer of UPSID's languages.
One place-of-articulation was the doubly-articulated labial-velar / g͡b / ; one manner-of-articulation was affricate.
I just decided it was impossible to have a labial-velar voiced affricate. I could have been wrong; but even if it can be pronounced, I couldn't find any language that has it.
User avatar
Coveny
rupestrian
rupestrian
Posts: 12
Joined: 17 Nov 2017 20:38
Contact:

Re: Morneau phonetic alphabet

Post by Coveny »

eldin raigmore wrote: 19 Nov 2017 03:25
Coveny wrote: 17 Nov 2017 23:13 .... has anyone else attempted .... to find common phonetic sounds across major languages?
You could look at UPSID Sound Selection and find all the phonemes that occur in more than, or at least, some fraction (say 50%) of their sample's languages;
for a more up-to-date notation and database with a bigger sample (between twice as big and five times as big, I think) you could look at PHOIBLE.
The 12 phonemes that UPSID says are in >=50% of their languages are among the 19 phonemes that PHOIBLE says are in >=50% of their (obviative 3rd person) languages.

UPSID's collection includes these nine consonants and three vowels:
Sound Description Sound occurs in % of UPSID's languages
m voiced bilabial nasal 94.24%
k voiceless velar plosive 89.36%
i high front unrounded vowel 87.14%
a low central unrounded vowel 86.92%
j voiced palatal approximant 83.81%
p voiceless bilabial plosive 83.15%
u high back rounded vowel 81.82%
w voiced labial-velar approximant 73.61%
b voiced bilabial plosive 63.64%
h voiceless glottal fricative 61.86%
g voiced velar plosive 56.10%
N voiced velar nasal 52.55% (UPSID doesn't use the eng ŋ IPA uses)

PHOIBLE's collection includes these fourteen consonants and five vowels:

Name Representation Description Segment class Combined class
m 2053/2155 (95%) LATIN SMALL LETTER M consonant c
k 2016/2155 (94%) LATIN SMALL LETTER K consonant c
i 1998/2155 (93%) LATIN SMALL LETTER I vowel v
a 1961/2155 (91%) LATIN SMALL LETTER A vowel v
j 1901/2155 (88%) LATIN SMALL LETTER J consonant c
u 1873/2155 (87%) LATIN SMALL LETTER U vowel v
p 1873/2155 (87%) LATIN SMALL LETTER P consonant c
w 1812/2155 (84%) LATIN SMALL LETTER W consonant c
n 1742/2155 (81%) LATIN SMALL LETTER N consonant c
s 1663/2155 (77%) LATIN SMALL LETTER S consonant c
t 1604/2155 (74%) LATIN SMALL LETTER T consonant c
b 1533/2155 (71%) LATIN SMALL LETTER B consonant c
o 1459/2155 (68%) LATIN SMALL LETTER O vowel v
e 1458/2155 (68%) LATIN SMALL LETTER E vowel v
l 1421/2155 (66%) LATIN SMALL LETTER L consonant c
h 1405/2155 (65%) LATIN SMALL LETTER H consonant c
ɡ 1380/2155 (64%) LATIN SMALL LETTER SCRIPT G consonant c
ŋ 1217/2155 (56%) LATIN SMALL LETTER ENG consonant c
d 1161/2155 (54%) LATIN SMALL LETTER D consonant c

/n s t o e l d/ are in half or more of PHOIBLE's sampled languages but in less than half of UPSID's sampled languages.

In any case these are a good guess to the phonemes that more than half of the world will "recognize".

Don't know if that helps?
Yes thank you. The Phoible's database is VERY cool...
ForDebating.com is coming in February. A debating website that I hope to populate with intelligent critical debator. You can follow updates at https://www.facebook.com/fordebating

Feature comparison chart
User avatar
Aevas
admin
admin
Posts: 1445
Joined: 11 May 2010 05:46
Location: ꜱᴇ

Re: Morneau phonetic alphabet

Post by Aevas »

I'd put the image in a spoiler, but since they don't work at the moment, I'll just link it here.
User avatar
Coveny
rupestrian
rupestrian
Posts: 12
Joined: 17 Nov 2017 20:38
Contact:

Re: Morneau phonetic alphabet

Post by Coveny »

Aszev wrote: 19 Nov 2017 12:45 I'd put the image in a spoiler, but since they don't work at the moment, I'll just link it here.
That's much more in line with what I'm looking for, and the image is pretty cool as well. Thanks. I'm trying to go through it and look for the total number of native speakers rather than just "a distinctive language". I decided to go with Chinese, Spanish, English, Hindi, Arabic, Portuguese, Russian, and Japanese. Now I want to find not only phonemes that are common between them but also removing ones that are difficult for non-native speakers to learn. (so like the Mandarin tones, and the English /ð/, /θ/ sounds are gone)

Do you guys know of any listing where they show segments that are difficult for non-native speakers?

I want my phonemes to be both common to most people on the planet AND easy to learn. I'd like a segment of around 30 constants and 10 vowels (which I know is more than is recommended) for my conlang. From what I've read getting the phonemes first is the prefered way to start, then I can move on to words, grammar, and writing. But this whole process is turning out to be more difficult than I expected. The bar of entry to being a conlanger is much higher than it looked like from the onset. I research stuff all the time, but this stuff is HARD to research...
ForDebating.com is coming in February. A debating website that I hope to populate with intelligent critical debator. You can follow updates at https://www.facebook.com/fordebating

Feature comparison chart
User avatar
eldin raigmore
korean
korean
Posts: 6352
Joined: 14 Aug 2010 19:38
Location: SouthEast Michigan

Re: Morneau phonetic alphabet

Post by eldin raigmore »

For any two phonemes each of which occur in 75% or more of the world's languages, the two must both occur together in 50% or more of the world's languages.
So each pair of /m k i a j u p w n s/ must occur together in 50% or more of PHOIBLE's languages.

For any three phonemes each of which occur in 83.34% or more of the world's languages, the three must all occur together in 50% or more of the world's languages.
So each three of /m k i a j u p w/ must occur together in 50% or more of PHOIBLE's languages.

For any four phonemes each of which occur in 87.5% or more of the world's languages, the four must all occur together in 50% or more of the world's languages.
So each four of /m k i a j/ must occur together in 50% or more of PHOIBLE's languages.

Does any of that help any?

Normally, for something to be a phoneme in a language, the language needs a minimal pair of words that differ in meaning, whose difference in pronunciation lies only in that that sound occurs in one of them where no sound occurs in the other. For instance,
/draft/ vs /raft/ is a minimal pair for /d/ (vs. nothing),
/draft/ vs /daft/ is a minimal pair for /r/ (vs. nothing),
/draft/ vs /drat/ is a minimal pair for /f/ (vs. nothing).


Normally, for two things to be different phonemes in a language, the language needs a minimal pair of words that differ in meaning, whose only difference in pronunciation lies in that one sound occurs in one of the words where the other sound occurs in the other word.

e.g.
/lamp/ vs /ramp/ distinguishes /l/ from /r/,
/lamp/ vs /limp/ distinguishis /a/ from /i/,
/lamp/ vs <lump> (/l^mp/ or /l@mp/) distinguishes /a/ from /^/ or /@/,
/lamp/ vs /larp/ distinguishes /m/ from /r/,
etc.

Difficulties non-native speakers have with phonemes, tend more to be about distinguishing (in hearing and/or in speech) one phoneme from another in a particular spot, rather than in telling whether a phoneme is present or absent.
Or at least that's my impression.
An exception might be the clicks in the "click-belt"; a person whose native language(s) do not have any consonants except pulmonic egressive consonants, who tries as an adult to learn a language with clicks, frequently (or so "they" say) has trouble at first hearing the clicks as speech sounds; instead s/he at first hears the clicks as a separate percussive accompaniment to the "speech" consisting of all the other sounds.

Another difficulties non-native speakers often have, is recognizing two sounds as allophones of the same phoneme in the language they are learning, when they are always separate phonemes in the learner's native language. E.g. maybe bilabial and labiodental sounds are distinct in the learner's L1, but are allophones of the same phonemes in the language being learned.

Does any of that help any?

How many consonants and how many vowels do you want?
For my conlang Adpihi:
I found that if I included all the consonants in 25% or more of UPSID's languages, but none of those in fewer than 20% of UPSID's languages, I had 32 consonants -- which are plenty. I did need to check to make sure each pair of them were distinguishable from each other; I did this by trying to pronounce them and to hear the difference, which is pretty subjective, but worked pretty well.
For vowels, I just crowded vowel-space as much as I could and still be able to tell each two vowels apart. That gave me nine vowels -- plenty. Trying to use all the vowels that occur in, say, 25% or more of UPSID's languages caused problems because of distinguishing mid vowels from close-mid vowels and/or from open-mid vowels.

Does that help any?

How many roots do you need?

If you have 18 consonants and 4 vowels you'll have potentially 18*4=72 CV syllables and so potentially 72^2=5184 CVCV roots.
Or, if you use a triconsonantal-root system, 18^3=5832 C-C-C roots, with 4^2=16 -V-V- transfixes.

With 20 consonants and 5 vowels, you could have up to 100 CV syllables, and up to 1,000,000 three-syllable CVCVCV words.

Have you thought about any of that?

Does thinking about it help any?
User avatar
Creyeditor
MVP
MVP
Posts: 5091
Joined: 14 Aug 2012 19:32

Re: Morneau phonetic alphabet

Post by Creyeditor »

eldin raigmore wrote: 19 Nov 2017 22:24 [...]

Normally, for something to be a phoneme in a language, the language needs a minimal pair of words that differ in meaning, whose difference in pronunciation lies only in that that sound occurs in one of them where no sound occurs in the other. For instance,
/draft/ vs /raft/ is a minimal pair for /d/ (vs. nothing),
/draft/ vs /daft/ is a minimal pair for /r/ (vs. nothing),
/draft/ vs /drat/ is a minimal pair for /f/ (vs. nothing).


Normally, for two things to be different phonemes in a language, the language needs a minimal pair of words that differ in meaning, whose only difference in pronunciation lies in that one sound occurs in one of the words where the other sound occurs in the other word.

e.g.
/lamp/ vs /ramp/ distinguishes /l/ from /r/,
/lamp/ vs /limp/ distinguishis /a/ from /i/,
/lamp/ vs <lump> (/l^mp/ or /l@mp/) distinguishes /a/ from /^/ or /@/,
/lamp/ vs /larp/ distinguishes /m/ from /r/,
etc.

Difficulties non-native speakers have with phonemes, tend more to be about distinguishing (in hearing and/or in speech) one phoneme from another in a particular spot, rather than in telling whether a phoneme is present or absent.
[...]
I just wanted to stress the point about constrast, because that's the defining property of a phoneme. One point that I want to add about is that phonemic analysis of the same languages by different phonologists might lead to different resulting phoneme inventories. That's very evident if you look at PHOIBLE.
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
Coveny
rupestrian
rupestrian
Posts: 12
Joined: 17 Nov 2017 20:38
Contact:

Re: Morneau phonetic alphabet

Post by Coveny »

eldin raigmore wrote: 19 Nov 2017 22:24 Difficulties non-native speakers have with phonemes, tend more to be about distinguishing (in hearing and/or in speech) one phoneme from another in a particular spot, rather than in telling whether a phoneme is present or absent.
Or at least that's my impression..

Another difficulties non-native speakers often have, is recognizing two sounds as allophones of the same phoneme in the language they are learning, when they are always separate phonemes in the learner's native language. E.g. maybe bilabial and labiodental sounds are distinct in the learner's L1, but are allophones of the same phonemes in the language being learned.

How many consonants and how many vowels do you want?
For my conlang Adpihi:
I found that if I included all the consonants in 25% or more of UPSID's languages, but none of those in fewer than 20% of UPSID's languages, I had 32 consonants -- which are plenty. I did need to check to make sure each pair of them were distinguishable from each other; I did this by trying to pronounce them and to hear the difference, which is pretty subjective, but worked pretty well.
For vowels, I just crowded vowel-space as much as I could and still be able to tell each two vowels apart. That gave me nine vowels -- plenty. Trying to use all the vowels that occur in, say, 25% or more of UPSID's languages caused problems because of distinguishing mid vowels from close-mid vowels and/or from open-mid vowels.


How many roots do you need?

If you have 18 consonants and 4 vowels you'll have potentially 18*4=72 CV syllables and so potentially 72^2=5184 CVCV roots.
Or, if you use a triconsonantal-root system, 18^3=5832 C-C-C roots, with 4^2=16 -V-V- transfixes.

With 20 consonants and 5 vowels, you could have up to 100 CV syllables, and up to 1,000,000 three-syllable CVCVCV words.

Have you thought about any of that?
I'm not so concerned about the number of places a consonant or vowel exists. As how difficult they are to pronounce to non-native speakers.

Is it possible to stay away from allophones and still get to 30 constants and 10 vowels?

So you got 9 vowels that were distinctive, and easy to pronounce for non-native speakers? 9 is about what I was shooting for. English has 44 phonemes so that where I was shooting, but to be fair if the phonemes were easy to pronounce for non-native speakers I would go all the way up to 52 total.

I haven't looked at words yet, not sure that I will do roots with affixes but I want to be able to create TON of small words. If I could have a fully functioning language where none of the words were longer than 6 letters that would be great. I want the language to be able to pack a lot of information in a short sentence, and have the speech distinctive enough to understand as well as easy to learn to hear and pronounce the words.

I also haven't started on grammar either so I'm not sure, but I'll likely go SVO, or OVS as they are 80% of the languages.

And I have not started on the writing either, so although I have some ideas about what I want from the glyph conceptually I have nothing on paper or on the computer about what they look like.

Full disclosure I'm a beginner, and this is my first Conlang. The general theme of my Conlang is that is full of sounds that would be easy for a child to learn. This is the stage that I'm at right now, getting the sounds for my language. The concept of the language that it's built for efficiency where speakers can convey a lot of information quickly. I don't intend to have articles. (a, an, the, ...) I'm looking at a posteriori language that takes the phonemes from the languages listed above and use them in my language, but I don't want to create a "Euroclone". I'm also interested in creating new phonemes that don't currently exist, so long as they are easy to learn for non-native speakers. (From the IP there looks like there could be some more constants added, they would just need to be easy and distinctive)

I assume that after I get it created I'll tinker with it, as several of the articles that I read indicated that's common for a conlang.
ForDebating.com is coming in February. A debating website that I hope to populate with intelligent critical debator. You can follow updates at https://www.facebook.com/fordebating

Feature comparison chart
User avatar
eldin raigmore
korean
korean
Posts: 6352
Joined: 14 Aug 2010 19:38
Location: SouthEast Michigan

Re: Morneau phonetic alphabet

Post by eldin raigmore »

Coveny wrote: 20 Nov 2017 01:12 Is it possible to stay away from allophones and still get to 30 constants and 10 vowels?
I'd say so. Look at WALS.info to find out what the maximum number of consonants and vowels are in their sample database:
http://wals.info/chapter/1 says their languages have from a minimum of 6 to a maximum of 122 different consonant phonemes;
http://wals.info/chapter/2 says their languages have from a minimum of 2 to a maximum of 14 different "vowel-qualities";
http://wals.info/chapter/3 says their languages have consonant-to-vowel ratios from a minimum of 1.11 to a maximum of 29;

There's also information about which phonemes, and/or which contrasts between phonemes, are common and which are uncommon.
http://wals.info/chapter/4 says how many languages have voicing contrasts (viz. voiced vs unvoiced) in their plosives, how many have voicing contrasts in their fricatives, how many have voicing contrasts in both, and how many have voicing contrasts in neither;
http://wals.info/chapter/5 is about gaps in the bilabial, alveolar, and velar voiced and unvoiced plosives /b p d t g k/; some languages have all six, some have all but /p/, some have all but /g/, some have all but /p/ and /g/, and some are "other";
http://wals.info/chapter/6 is about uvular consonants, which many languages don't have;
http://wals.info/chapter/7 is about glottalized consonants, which again many languages don't have;
http://wals.info/chapter/8 is about languages which do or don't have a lateral liquid (most do), and/or do or don't have lateral obstruents (most don't); there is some discussion in the chapter also about the fact that some languages do (most do), and some don't, contrast a lateral liquid with a rhotic liquid. (If a speaker's L1 has only one liquid, they may not be able to distinguish between an L2's rhotic liquid(s) and its lateral liquid(s). Also, some languages have two lateral liquids; if a non-native speaker's L1 has only one lateral liquid, they may not be able to distinguish between the L2's lateral liquids. Most languages have at least two liquids; most languages have at most two liquids; most languages have at least one rhotic; most languages have at most one rhotic; most languages have at least one lateral liquid; most languages have at most one lateral liquid.);
http://wals.info/chapter/9 talks about which languages don't have a velar nasal /ŋ/ (most don't, in this chapter; though more than 50% of PHOIBLE's or UPSID's languages do); which do have /ŋ/ but never at the beginning of a syllable (most of the rest); and which have /ŋ/ also at the beginnings of syllables (fewest of all).
http://wals.info/chapter/18 talks about languages which lack certain common sounds; they don't have any bilabials (a missing place-of-articulation), and/or don't have any fricatives (a missing manner-of-articulation), and/or don't have any nasals (another missing MoA).
http://wals.info/chapter/19 talks about languages which have certain uncommon sounds; clicks (velaric ingressive airstream), and/or labial-velar plosives (/w/ is a common labial-velar approximate, but other labial-velar consonants are uncommon), and/or pharyngeals, and/or apico-dental non-sibilant fricatives (aka "thibilants", like unvoiced <th> in <thorn> or voiced <th> in <than>).
Coveny wrote: 20 Nov 2017 01:12So you got 9 vowels that were distinctive, and easy to pronounce for non-native speakers? 9 is about what I was shooting for.
I am not a native speaker of my conlang.
In "portato" speech -- slow and careful speech -- I was able to enunciate them and to tell which one was being pronounced. So at least one non-native speaker could handle them in slow, careful speech.
In "allegro" speech I don't know whether I could do that. But I found out from another CBBean that my conlang's vowel-inventory was the same as some dialect of Somali's. So some group of native speakers of a natlang with that same vowel-inventory does just fine with them in allegro speech.
An average number of vowel-qualities would be 5 or 6.
An average number of consonant phonemes would be 19 to 25.

Coveny wrote: 20 Nov 2017 01:12English has 44 phonemes so that where I was shooting, but to be fair if the phonemes were easy to pronounce for non-native speakers I would go all the way up to 52 total.
That's a lot.
An average (i.e. median) language has around 3.5 consonants per vowel.
So with 44 phonemes you'd expect about 34 consonants and about 10 vowels, or maybe about 35 consonants and about 9 vowels.
With 52 phonemes you'd expect about 40 consonants and about 12 vowels, or maybe about 41 consonants and about 11 vowels.
(English has 32 consonants and 12 vowels, but some of those "vowels" are diphthongs; WALS.info wouldn't count them as "vowel qualities", but rather as tautosyllabic vowel-clusters, pairs of vowels together in the nucleus of the same one syllable.)

I do not have any cross-linguistic theory-neutral way to decide which phonemes are easy to pronounce (or otherwise learn) for non-native speakers.
Coveny wrote: 20 Nov 2017 01:12I haven't looked at words yet, not sure that I will do roots with affixes but I want to be able to create TON of small words. If I could have a fully functioning language where none of the words were longer than 6 letters that would be great.
Is 30,000 to 50,000 roots enough to be "fully functioning"?
English has over 177,000 roots, according to the latest OED; more than half are noun-roots, about a third are adjective-roots (up from a quarter in the 2nd edition of the OED), about a seventh are verb-roots, and the rest (around 1/42 in the current edition, around 3/28 in the 2nd edition) are roots of other parts-of-speech.
Would 177,000 "+ change" be enough roots to be "fully functioning"?

And, btw, I'm sure you meant "6 phonemic segments" rather than "6 letters".
If your character-set is an abugida or alphasyllabary you could probably get over a million words writable with 3 characters or fewer.
If it's a logography like Chinese hanzi or kanji or like Egyptian hieroglyphics most words would be writable with 1 character.
If it's an abjad a "6 letter word" could easily contain 13 phonemes because you wouldn't write the vowels.
... ... ...
Anyway, with 15 consonants and 10 vowels you could have 225,000 VCVCV words or roots or morphemes;
with 10 consonants and 5 vowels you could have 625,000 VCVCVCV words or roots or morphemes (seven phonemes, I know);
and with 20 consonants and 5 vowels if your allowable root-shapes or word-shapes or morpheme-shapes were:
VCVCV
CVCVC
CCVCV
VCCVC
CVCCV
VCVCC
you could have 5,000 + 5*(200,000) = 1,005,000 roots or words or morphemes, all of them 5 phonemes long, having 2 or 3 consonants and 2 or 3 vowels each, no vowel clusters, at most one consonant cluster, no consonant cluster longer than two consonants.
Coveny wrote: 20 Nov 2017 01:12I want the language to be able to pack a lot of information in a short sentence,
An admirable design goal that not every conlang nor every natlang meets (though it's not a goal for some of the conlangs, and natlangs aren't designed so the idea of "design goals" for natlangs is problematic).
Coveny wrote: 20 Nov 2017 01:12 and have the speech distinctive enough to understand as well as easy to learn to hear and pronounce the words.
Every natlang must satisfy this goal in order to be acquired as a first language by the children in its speech-community.
So every conlang which wants to be naturalistic and/or realistic also has to be "learnable"; not only that, but have a certain minimal degree of "learnability".
However that doesn't mean it has to be easy to acquire as an L2 by an adult who only speaks their L1 and doesn't know any languages similar to the L2 they're attempting to learn.
Coveny wrote: 20 Nov 2017 01:12I also haven't started on grammar either so I'm not sure, but I'll likely go SVO, or OVS as they are 80% of the languages.
You mean SVO or SOV. Around 82% (give or take a few, I imagine) of natlangs are either SVO or SOV ("subject-initial"). Another 17% (maybe give or take a percent) are either VOS or VSO ("verb-initial"). IIRC the OVS and OSV languages ("object-initial") together constitute fewer than 1% of natlangs. I suppose I could be wrong in detail; but SVO and SOV are definitely the biggest and OVS and OSV are definitely the smallest.
Coveny wrote: 20 Nov 2017 01:12And I have not started on the writing either, so although I have some ideas about what I want from the glyph conceptually I have nothing on paper or on the computer about what they look like.
You might not be able to talk about "how many letters a word has" yet. (Or, maybe, you are ready to do that.)
Coveny wrote: 20 Nov 2017 01:12Full disclosure I'm a beginner, and this is my first Conlang.
Brave soul!
I wish you success.
Odds are any success you have will surpass mine so far!
Coveny wrote: 20 Nov 2017 01:12 The general theme of my Conlang is that is full of sounds that would be easy for a child to learn.
There is literature on "learnability" of languages; and it includes some ideas about how to rank a language's learnability.
But I don't have much idea about how learnable a set of sounds would be.
Maybe somebody has done work on that.
The best I can point you to is WALS.info chapters 1-9 and 18-19; look up their references, and/or write to their authors. Most of them will answer you -- once. A few of them will answer a follow-up question too.
Coveny wrote: 20 Nov 2017 01:12This is the stage that I'm at right now, getting the sounds for my language.
The next thing to do, then, is phonotactics.
That means (in the strict sense): "what pairs of sounds can occur consecutively in a particular order in words, and what pairs can't?"
It's actually got a broad sense, too.
What's the longest consonant-cluster that can happen in a word?
What's the longest vowel-cluster that can happen in a word?
Which phonemes can be whole words all by themselves?
Which phonemes can be the first phoneme of a word with more than one phoneme?
Which phonemes can be the last phoneme of a word with more than one phoneme?
Same questions for phoneme-pairs; Which can be whole words by themselves? Which can be the first two phonemes of a word with three or more phonemes? Which can be the last two phonemes of a word with three or more phonemes?
Ask and answer all the same questions replacing "word" by "morpheme".
Ask again, replacing "word" by "syllable".
Ask again, replacing "syllable" by "syllable onset"; then by "syllable nucleus"; then by "syllable coda".
For the phoneme-pairs, especially if one is a consonant and the other is a vowel, you may want to also look at "syllable rime" (the part after the onset) and/or "syllable body" (the part before the coda).
Coveny wrote: 20 Nov 2017 01:12The concept of the language that it's built for efficiency where speakers can convey a lot of information quickly.
It's possible that your goal of "efficiency" and your goal of "learnability" will fight each other.
Coveny wrote: 20 Nov 2017 01:12I'm also interested in creating new phonemes that don't currently exist, so long as they are easy to learn for non-native speakers. (From the IP there looks like there could be some more constants added, they would just need to be easy and distinctive)
Yes, it has happened in the past that there were empty spots on the IPA chart that any linguist could pronounce, but no-one had found a natlang using them. Then, after some years, someone did find one or a few such natlangs.

.... .... .... ....

Good luck!
PM me if you want me to say more about some topic or other.
(Or just post it in your thread if you don't care who answers!)
BTW you don't have to quote the whole post to which you are responding; you can quote just the relevant bits if you want.
I wish the spoilers worked already; some of the above I would put in spoilers, in case the whole post would otherwise be too long, and not be read.
User avatar
Coveny
rupestrian
rupestrian
Posts: 12
Joined: 17 Nov 2017 20:38
Contact:

Re: Morneau phonetic alphabet

Post by Coveny »

eldin raigmore wrote: 20 Nov 2017 03:36
Coveny wrote: 20 Nov 2017 01:12The concept of the language that it's built for efficiency where speakers can convey a lot of information quickly.
It's possible that your goal of "efficiency" and your goal of "learnability" will fight each other.
Thank you, that's some really good stuff there that's going to take me some time to go through.

I understand there may need to be compromise, but ya gotta have goals right?

Any thanks again, it'll take me days to digest all that, and it looks like it may take me in a different direction for my phonemes, but the depth and breath of knowledge required for the undertaking is insane. Had I known how difficult this was going to be I might not have decided I wanted to do it...
ForDebating.com is coming in February. A debating website that I hope to populate with intelligent critical debator. You can follow updates at https://www.facebook.com/fordebating

Feature comparison chart
User avatar
eldin raigmore
korean
korean
Posts: 6352
Joined: 14 Aug 2010 19:38
Location: SouthEast Michigan

Re: Morneau phonetic alphabet

Post by eldin raigmore »

It might be profitable to read the thread this post is in.
Like Creyeditor said here, you might find the World Phonotactics Database useful.
Post Reply