Question for a phonetically simple interlang

A forum for all topics related to constructed languages
Khunjund
rupestrian
rupestrian
Posts: 22
Joined: 06 Feb 2015 03:02

Question for a phonetically simple interlang

Post by Khunjund »

I’ve very recently started working on an interlang which is designed to be easily pronounceable for (I’d like to say) ~99% of people. For this, I went through the phonetic inventories of the top 91 languages by number of native speakers (I chose native speakers instead of total speakers, because I figured not every L2 speaker might be perfectly comfortable with all the phonemes of the language in question, whereas the phonemes in his or her native language would be perfectly natural), and kept only the phonemes which were present in every single one of them, which gave me the following inventory:

/m l j/
/t k s/

/i u/
/e a/

Now I would be perfectly fine with this inventory, except that it seems to me like the lack of /p/ isn’t wholly justified. While there are several languages in the list without /p/ proper, every language which lacked it either had /b/ or /f/. I would have just used /p/, but there are footnotes in the pages of some languages (like Egyptian Arabic) which specifically say that some speakers have difficulty with this sound. Because of this, I thought I could include a phoneme /p~f/, which is voiceless but doesn’t distinguish between stop and fricative (which seemed a better option to me than /p~b/, since there are no other voiced obstruents), but up to this point all of my phonemes have one clear main pronunciation, and I can’t decide if the benefits of having this one phoneme would outweigh its unwieldiness. However, I have to make this decision before deriving substantial amounts of vocabulary.

So my question is: should I include the phoneme /p~f/, or should I keep the phonemic inventory as clean as possible? Or maybe someone has a suggestion that’s entirely different?
Fluent: :ca-qc: :eng: | Learning: :jpn: | Interested: :ara: :chn: :rus: :hkg: :ind: :por: :deu: :tha: :tur: :vie: :kor: :swe: :pol: :rou: :fin: :esp: :isr: :nga: :irl: :ita: :tib: :ell: :eth: etc.
User avatar
eldin raigmore
korean
korean
Posts: 5698
Joined: 14 Aug 2010 19:38
Location: SouthEast Michigan

Re: Question for a phonetically simple interlang

Post by eldin raigmore »

What’s your syllable structure?
Even with CV syllables you could get over 13,000 CVCVCV three-syllable words or morphemes
and could get over 300,000 CVCVCVCV four-syllable words or morphemes.

Maybe you’d like CV(C) syllable structure. With that phoneme inventory you could have 6*4*7 = 168 syllables of that shape;
and up to 28000 two-syllable morphemes or words, or more than 4.7 million three-syllable morphemes or words, way the heck more than you need.

Suppose you allow some two-consonant clusters in the onset. Maybe five of the six consonants can be first in an onset-cluster and a different five can be last in an onset-cluster and if two consonants are in an onset they have to come in a particular order. There’d be 1*1 + 1*4 + 4*1 + (4*3)/2 = 1+4+4+6 = 15 two-consonant onsets. Maybe any consonant can be a one-consonant onset.
If there are 15 legal CC onset clusters and 6 legal one-consonant onsets and onsets are mandatory there could be 21 onsets. Maybe only five of the six consonants can be in codas and codas can’t be clusters but codas are optional. Maybe you don’t allow diphthongs or longer polyphthongs in your syllable nuclei but you don’t have any consonantal syllables.
Then you could have 4*(5+1) = 24 different rimes.
And so 21*24 = 504 (C)(C)V(C) syllables and 504^2 > 254,000 two-syllable sequences, almost surely more than enough for all your morphemes.

"…………………………………………………………

How many morphemes do you need to allow for?
How many roots and how many affixes?
Among roots, how many nouns, how many adjectives, and how many verbs?

Will it be important to phonetically distinguish between prefixes and roots or between roots and suffixes?
Will it be important to phonetically distinguish between nouns and verbs? How about between adjectives and either nouns or verbs?

Will it be important to tell where syllable boundaries are?

How many syllables do you want most morphemes to have?
If you allow five-syllable morphemes then already with 24 CV syllables you have maybe up to just shy of 7.9 million.
With four-syllable morphemes instead you have 331,000, nearly twice as many as English has roots.
With 3-syllable morphemes you have around 13,800. In ordinary everyday speech about topics not requiring any expertise most people would stick to the commonest 3000 to 5000 words.
You’d have fewer than 600 2-syllable words, and I don’t think that’s enough.
If you want over 50,000 two-syllable words or morphemes you want at least 224 different syllables which would mean you’d need an average of about 2.25 consonants per syllable if there are only six consonants.
With 7 consonants, though, you could get over 30,000 2-syllable morphemes or words with an average of 1.95 consonants per syllable.
Or you could just say there are going to be a bunch of three-syllable-or-longer words used in common speech.

…………………………


I’m curious why you stop at 6 or 7 consonants. PHOIBLE and UPSID list around 19 phonemes (I think!) that occur in over half the world’s languages — or, rather, half of their samples. I would have guessed a 12-consonant inventory might be just as good for your purpose. I believe you have a good reason for just those 6 or 7 but I can’t guess what it is.

I’m also curious why you have 4 vowels instead of 5 or 3. 5 vowels is most common. But in creating auxlangs it’s often recommended to go with a 3-vowel inventory because people whose L1 has only three vowels often have trouble with a 5-vowel system.
In a conlang I once collaborated on we had a good (or “good”?) reason for our 4-vowel system; so I guess you have one as well; but what is it?
Last edited by eldin raigmore on 04 Nov 2020 00:49, edited 2 times in total.
Khunjund
rupestrian
rupestrian
Posts: 22
Joined: 06 Feb 2015 03:02

Re: Question for a phonetically simple interlang

Post by Khunjund »

My syllable structure is strictly (C)V, with a handful of additional restrictions (such as /j/ can’t occur adjacent to /i/ so there can never be a minimal pair with /i.a/ vs /ija/ for instance), which might be another reason to have one additional phoneme.
Fluent: :ca-qc: :eng: | Learning: :jpn: | Interested: :ara: :chn: :rus: :hkg: :ind: :por: :deu: :tha: :tur: :vie: :kor: :swe: :pol: :rou: :fin: :esp: :isr: :nga: :irl: :ita: :tib: :ell: :eth: etc.
User avatar
eldin raigmore
korean
korean
Posts: 5698
Joined: 14 Aug 2010 19:38
Location: SouthEast Michigan

Re: Question for a phonetically simple interlang

Post by eldin raigmore »

With 7 consonants, 4 vowels, and (C)V syllable structure, and no restrictions, there’d be 32 syllables. (Counting those that are just a vowel).
You could have 1024 two-syllable morphemes. Very small vocabulary.
But you could have 32,768 three-syllable morphemes. Pretty adequate for many purposes!
....
In that collaborative conlang I mentioned before we wanted at least 5000 roots.
If we used Semitic-style C-C-C roots then 17 consonants wouldn’t be quite enough, so we went with 18 consonants.
If we used Polynesian-style CVCV roots then 70 CV syllables wouldn’t be quite enough, so we wanted at least 71.
With 18 consonants and 4 vowels we got 72 CV syllables.
There’d have been over 93,000 CVCVC stems. And over 343,000 CVCVCV sequences. Depending on what we wanted to use.
Khunjund
rupestrian
rupestrian
Posts: 22
Joined: 06 Feb 2015 03:02

Re: Question for a phonetically simple interlang

Post by Khunjund »

eldin raigmore wrote: 04 Nov 2020 00:21 I’m curious why you stop at 6 or 7 consonants. PHOIBLE and UPSID list around 19 phonemes (I think!) that occur in over half the world’s languages — or, rather, half of their samples. I would have guessed a 12-consonant inventory might be just as good for your purpose. I believe you have a good reason for just those 6 or 7 but I can’t guess what it is.

I’m also curious why you have 4 vowels instead of 5 or 3. 5 vowels is most common. But in creating auxlangs it’s often recommended to go with a 3-vowel inventory because people whose L1 has only three vowels often have trouble with a 5-vowel system.
In a conlang I once collaborated on we had a good (or “good”?) reason for our 4-vowel system; so I guess you have one as well; but what is it?
As I said, I got my phoneme inventory by going through the Ethnologue 2019 list of the 91 languages with the most native speakers (on the Wikipedia article). These six consonants are the ones which are distinguished in all these languages. Being a little more lenient with my criteria, I could have included /p/ (the reason for my posting this in the first place) and /n/, but I have other reasons for not including /n/ as well. Also, while that 50% of languages probably includes more than 50% of the human population, I really wanted to try for something which was practically universal.

As for the four vowels, while Modern Standard Arabic has only three vowel qualities, it turns out that every variety of spoken Arabic (all the ones featured on the list, at least) distinguishes more than that: most have five (similarly to modern Hebrew), but a variety have four, which ended up being the limiting factor for my inventory. One of these (Levant Arabic I think) had something like /i e o a/, with some instances of /a/ having raised to /e/ and /u/ always lowering to /o/, whereas Algerian Arabic has merged all short vowels into /ə/, which contrasts with the three long vowels /iː uː aː/. Moreover some variety of Persian distinguished only /i u e a/. I thought about using a three-vowel system, but that didn’t feel right since all of my source languages (for phonemic inventory) distinguished at least four, so I normalized it to /i u e a/.
Fluent: :ca-qc: :eng: | Learning: :jpn: | Interested: :ara: :chn: :rus: :hkg: :ind: :por: :deu: :tha: :tur: :vie: :kor: :swe: :pol: :rou: :fin: :esp: :isr: :nga: :irl: :ita: :tib: :ell: :eth: etc.
Khunjund
rupestrian
rupestrian
Posts: 22
Joined: 06 Feb 2015 03:02

Re: Question for a phonetically simple interlang

Post by Khunjund »

Most of my roots end up being three syllables, but this is an interlang, so I think my question is more along the lines of “Is having a /p~f/ phoneme acceptable considering the initial restrictions I imposed myself, and is it worth transgressing those restrictions in favour of having some roots be slightly more recognizable?” as opposed to just “Can I make do with 19683 possible three-syllable combinations, or do I really need to go up to 29791?”

For instance, the word for “father” currently is tasa, even though the majority of my source languages (for vocabulary) had an initial labial (either /p/ or /f/), which would have lead to the word pasa in my language if it had had the /p/ phoneme. This is just one example, but is it worth “cheating” my initial rules for limiting my phoneme inventory?
Fluent: :ca-qc: :eng: | Learning: :jpn: | Interested: :ara: :chn: :rus: :hkg: :ind: :por: :deu: :tha: :tur: :vie: :kor: :swe: :pol: :rou: :fin: :esp: :isr: :nga: :irl: :ita: :tib: :ell: :eth: etc.
User avatar
eldin raigmore
korean
korean
Posts: 5698
Joined: 14 Aug 2010 19:38
Location: SouthEast Michigan

Re: Question for a phonetically simple interlang

Post by eldin raigmore »

Thanks!
Have you decided to have a whole bunch of three-syllable roots or morphemes or words?
Could one be VVV, like aei or aeo or aio or eio?

Or maybe there’s a rule that one of the first two syllables has to have an onset?
Or that one of the last two syllables has to have an onset?
Khunjund
rupestrian
rupestrian
Posts: 22
Joined: 06 Feb 2015 03:02

Re: Question for a phonetically simple interlang

Post by Khunjund »

A word that’s just VVV is possible, but as I said, I’m deriving my vocab from existing languages, so the conditions for it to happen are somewhat improbable.

At the moment, only have about fifty words or so, since I only started recently, and I’m still considering whether to add the /p/ phoneme or not, which I would like to settle before progressing further with vocab, seeing as I’ll have to review a bunch of words if I do make that change.
Fluent: :ca-qc: :eng: | Learning: :jpn: | Interested: :ara: :chn: :rus: :hkg: :ind: :por: :deu: :tha: :tur: :vie: :kor: :swe: :pol: :rou: :fin: :esp: :isr: :nga: :irl: :ita: :tib: :ell: :eth: etc.
User avatar
Vlürch
sinic
sinic
Posts: 353
Joined: 09 Mar 2016 21:19
Location: Finland
Contact:

Re: Question for a phonetically simple interlang

Post by Vlürch »

There's no reason you couldn't have /p~b/ even without any (other) voiced stops, because the bilabial stop is the one that would be crosslinguistically the most common to be the only voiced stop, while /f/ is less common. A stop inventory of /b t k/ would be perfectly naturalistic, and if you have it as [p~b] in free variation, you cover pretty much all the bases except for languages that don't have any bilabial stops, like Cherokee.
User avatar
eldin raigmore
korean
korean
Posts: 5698
Joined: 14 Aug 2010 19:38
Location: SouthEast Michigan

Re: Question for a phonetically simple interlang

Post by eldin raigmore »

Khunjund wrote: 04 Nov 2020 05:46 … as I said, I’m deriving my vocab from existing languages, …
I missed that!

Khunjund wrote: 04 Nov 2020 05:46 … At the moment, only have about fifty words or so, …
I missed that too!


Good luck! I’ll be interested in your success!
User avatar
Davush
greek
greek
Posts: 553
Joined: 10 Jan 2015 14:10

Re: Question for a phonetically simple interlang

Post by Davush »

Although I'm a fan of small inventories, such a small inventory for an interlanguage poses some interesting problems. Of course, you might just want to make an interlanguage with a such a small inventory 'just because', which is fine, but there are some pragmatic issues...

First perhaps to consider is that such small inventories are overall rare among the languages of the world. While they may be easier to pronounce, in many ways they make the language much more difficult to learn because of the large amount of words which will end up looking very similar in a strictly CV language with few phonemes. When everything starts sounding or looking like /malate tusaki jataku musule/ (and so on), words don't have much of unique 'hook' for memorisation. That is to say, some linguists (Trudgill, I think) have noted that extremely small phoneme inventories are actually on the higher end of 'complexity' in that they have a much greater possibility of confusability and increase memory load of learners.

Although I do like small inventories, I am not convinced they are particularly useful for auxlangs/interlanguages. After all, I think most people are able to learn to produce and distinguish the 20 or so most common sounds with enough exposure and practice. Languages which have a typologically average number of consonants and vowels would be expected to be easier to learn for most people, I think, as words will be more distinct. Of course, these are just my own thoughts, and a minimalist interlanguage could indeed be interesting!
DV82LECM
cuneiform
cuneiform
Posts: 171
Joined: 16 Dec 2016 03:31

Re: Question for a phonetically simple interlang

Post by DV82LECM »

Frankly, I think /m/ is a handy substitute for: 1. being the ONLY nasal, 2. being the ONLY labial, and 3. being the strangely adorable completion to your series of "approximates." Not having a nasal ISN'T odd, but noting that there is ONE left insists that there were; no labial ISN'T odd, but the same can be said about it, as noted in the previous, pertaining to the fact it's there, denoting labials at one point; and it does complete the compliment sounds to your main series of "stops" that make up your approximates (/s/ being treated as a stop in a few languages).
Khunjund
rupestrian
rupestrian
Posts: 22
Joined: 06 Feb 2015 03:02

Re: Question for a phonetically simple interlang

Post by Khunjund »

DV82LECM wrote: 05 Nov 2020 03:30 Frankly, I think /m/ is a handy substitute for: 1. being the ONLY nasal, 2. being the ONLY labial, and 3. being the strangely adorable completion to your series of "approximates." Not having a nasal ISN'T odd, but noting that there is ONE left insists that there were; no labial ISN'T odd, but the same can be said about it, as noted in the previous, pertaining to the fact it's there, denoting labials at one point; and it does complete the compliment sounds to your main series of "stops" that make up your approximates (/s/ being treated as a stop in a few languages).
So you’d be in favour of not having /p/ and using /m/ in its place?
Fluent: :ca-qc: :eng: | Learning: :jpn: | Interested: :ara: :chn: :rus: :hkg: :ind: :por: :deu: :tha: :tur: :vie: :kor: :swe: :pol: :rou: :fin: :esp: :isr: :nga: :irl: :ita: :tib: :ell: :eth: etc.
DV82LECM
cuneiform
cuneiform
Posts: 171
Joined: 16 Dec 2016 03:31

Re: Question for a phonetically simple interlang

Post by DV82LECM »

Khunjund wrote: 05 Nov 2020 03:45
DV82LECM wrote: 05 Nov 2020 03:30 Frankly, I think /m/ is a handy substitute for: 1. being the ONLY nasal, 2. being the ONLY labial, and 3. being the strangely adorable completion to your series of "approximates." Not having a nasal ISN'T odd, but noting that there is ONE left insists that there were; no labial ISN'T odd, but the same can be said about it, as noted in the previous, pertaining to the fact it's there, denoting labials at one point; and it does complete the compliment sounds to your main series of "stops" that make up your approximates (/s/ being treated as a stop in a few languages).
So you’d be in favour of not having /p/ and using /m/ in its place?
Yes. See, naturalism is what all of us insist upon, even when we want to make something "unnatural" (but is still done adjacent to something natural). The thing with naturalism, though, is that it is based upon rules which we construct, which are, themselves, based upon things that we witness in the world. Language, by its nature, is one of the MOST predictably unpredictable things, so deriving rules for naturalism in language is a funny thing. The only things that we could derive are predictions based upon a small margin of likelihood, entirely rooted within what is known and can be compared. However, these prescribed likelihoods are NOT restrictions upon what there is to be described. Could your language exist? Yes. Are there precedents set for the likelihood of your language? I don't know, but I'd imagine maybe that there ARE, though many might say NOT (although, consider Iau, Rotokas, and Piraha). At the end of the day, I figure that if you set rules for HOW, based upon our agreed prescriptions, then WHAT you describe is suitable as it falls into the framework of "naturalism yet to be attested."
User avatar
eldin raigmore
korean
korean
Posts: 5698
Joined: 14 Aug 2010 19:38
Location: SouthEast Michigan

Re: Question for a phonetically simple interlang

Post by eldin raigmore »

In my experience so far (not that there’s been a vast amount of it!),
my (in my opinion justified) working hypothesis is,
the only theory-neutral cross-linguistic predictor of naturalism
is learnability.
There are actually mathematicized and/or quantified and/or empirical studies of learnability.
I recommend them.

I could be wrong.
DV82LECM
cuneiform
cuneiform
Posts: 171
Joined: 16 Dec 2016 03:31

Re: Question for a phonetically simple interlang

Post by DV82LECM »

eldin raigmore wrote: 05 Nov 2020 14:45 In my experience so far (not that there’s been a vast amount of it!),
my (in my opinion justified) working hypothesis is,
the only theory-neutral cross-linguistic predictor of naturalism
is learnability.
There are actually mathematicized and/or quantified and/or empirical studies of learnability.
I recommend them.

I could be wrong.
You're NOT, but the essence of this discussion is about the viability of Khunjund's phonology. Of course, someone could learn it; it's tiny.

What I was saying wasn't really a theory, it was a proposal for an axiom pertaining to viability of naturalism. Nature randomly, but somewhat predictably, dictates what IS. We find languages, compare them, describe their nature, then compare them to others already attested. Once we've done all of that, we reasonably ascribe their nature to our rules of what is prescriptively acceptable. Of course, we happen upon things, though NOT many left, which challenge that notion. As conlangers, I insist not to get hung up on what is objectively, prescriptively natural, but work how to trace our steps back to how anything, within that margin I mentioned, COULD be.
Last edited by DV82LECM on 05 Nov 2020 19:44, edited 1 time in total.
Khunjund
rupestrian
rupestrian
Posts: 22
Joined: 06 Feb 2015 03:02

Re: Question for a phonetically simple interlang

Post by Khunjund »

My goal isn’t really to have a “naturalistic” phoneme inventory, though I do think it ended up being sufficiently naturalistic for my purposes.

My dilemma is more along the lines of “Am I allowed to include the /p/ phoneme, even though not all languages in my list have it as-is, when my initial constraint was ‘Use only the sounds which appear as-is in all languages on the list.’” Again, since every language without /p/ had /b/ at least, I don’t think it’s a major transgression, but it would be one nonetheless.
Fluent: :ca-qc: :eng: | Learning: :jpn: | Interested: :ara: :chn: :rus: :hkg: :ind: :por: :deu: :tha: :tur: :vie: :kor: :swe: :pol: :rou: :fin: :esp: :isr: :nga: :irl: :ita: :tib: :ell: :eth: etc.
DV82LECM
cuneiform
cuneiform
Posts: 171
Joined: 16 Dec 2016 03:31

Re: Question for a phonetically simple interlang

Post by DV82LECM »

Khunjund wrote: 05 Nov 2020 18:54 My goal isn’t really to have a “naturalistic” phoneme inventory, though I do think it ended up being sufficiently naturalistic for my purposes.

My dilemma is more along the lines of “Am I allowed to include the /p/ phoneme, even though not all languages in my list have it as-is, when my initial constraint was ‘Use only the sounds which appear as-is in all languages on the list.’” Again, since every language without /p/ had /b/ at least, I don’t think it’s a major transgression, but it would be one nonetheless.
And that is WHY allophones are your best friend.
Khunjund
rupestrian
rupestrian
Posts: 22
Joined: 06 Feb 2015 03:02

Re: Question for a phonetically simple interlang

Post by Khunjund »

DV82LECM wrote: 05 Nov 2020 19:46 And that is WHY allophones are your best friend.
Sure, but my reason for limiting myself to those precise sounds in the first place was specifically so that every phoneme would have one unambiguous pronunciation, so that a) spreakers would not need to learn how to produce any new sounds, even if they’re simple, and b) while the system would allow for a lot of allophony and still remain clear, I wouldn’t have to define any for it to work at all.

For the time being, I’ve chosen to include the phoneme /p/ with the intended pronunciation [p] and no predetermined allophony, though of course no confusion will occur if it’s pronounced [ b ] anyone know how to write this so it isn’t understood as a tag for bold text? or [f], etc. While this breaks my initial ruleset, I believe it’s minor enough to not be a problem.
User avatar
Vlürch
sinic
sinic
Posts: 353
Joined: 09 Mar 2016 21:19
Location: Finland
Contact:

Re: Question for a phonetically simple interlang

Post by Vlürch »

Khunjund wrote: 07 Nov 2020 02:04anyone know how to write this so it isn’t understood as a tag for bold text?
Use a zero width space, you can copy it from eg. this page and then paste it after the b, i, s or u: [b​] [i​] [s​] [u​]. Of course, you can also add it to your keyboard if you want to be able to type it (at least on Windows, not sure about other operating systems).
Post Reply