Looking for critiques of my conlang's (tentative) phonology

[Duns Scotus]

So, I've been working rather intensely over the last several weeks on a conlang project of mine. I've really only been focusing on the morphology, with only a vague idea of what I want the phonology to sound like. Anyway, I've come to the point where in order to progress, I need to have a list of sounds so that I can put my grammar in more concrete terms.

On the subject of phonology, I really like the sound of Germanic, especially North Germanic languages. Now, I know that trying to emulate the sound of a particular language is generally a fool's errand, and pretty much destined to fail (as well as being unoriginal), but I've tried to look at the Scandinavian languages, particularly Icelandic and Faroese, for inspiration. Incidentally, it should be noted that grammatically, the language isn't really Indo-European. In fact, grammatically, it isn't really intended to be naturalistic, or at least that's not a priority. It's a language for my own personal use, designed to fit my own aesthetic criteria. I do want it to be easy for me to use, so I intend to have largely unnatural amounts of morphosyntactical regularity (though I suppose not entirely unattested, as in e.g. Quechua). Nevertheless, my dissatisfaction with most "regular" languages is that they sound, well, ugly. So while grammatical naturalness isn't important for me, phonological naturalness is (I realize these probably go somewhat hand in hand). If that means allophony at the expense of orthographic clarity, so be it. Anyhow, let me give a rough sketch of what I have so far:

Consonants:

Here is a consonant chart I made in LaTeX



As you can see, I've included the voiceless nasals and approximants found in Icelandic. However, my consonant inventory is somewhat intermediary between Icelandic and Old Norse, as I've kept the voiced stops instead of aspirated unvoiced stops. I put /z/ in parentheses, because I'm unsure about how I want to this sound to feature in the language, if at all. Also, I haven't gotten to allophones yet, so those aren't featured in this chart, but I will probably have the usual /ŋ/ and possibly /ɲ/ (in unvoiced varieties too). One thing I am curious about is to what extent allophony is necessary to make a language sound natural. I suspect at least some is needed, to break up the monotony. Also, it isn't evident in the chart, but I'm unsure if the velar fricatives will be their own phonemes or allophones of the velar stops. Anyway, on to vowels.

Vowels:

Here's the vowel chart I made



Since I wanted to imitate a germanic sort of sound, I've included a set of three front rounded vowels. The vowel system is less specifically Icelandic, than a mash up of various germanic languages (I included the ae sound, for instance, because I like it a whole lot). Also, unlike Icelandic (but like Old Norse), my language includes phonemic vowel length. Orthographically, this is indicated as in Finnish by doubling the letter.


Well, I guess that wraps it up. Phonology/phonetics is definitely the area of linguistics which I am weakest in, so I'm curious to hear whether my inventories are noobish, or respectable.

Thanks.

[StrangerCoug]

The phonological system is not bad, and I do like how the voiceless nasals and liquids make it seem less SAE. As far as orthography, I would personally switch <o> and <w> around as well as <q> and <x>, but I do like the system other than that.

[Duns Scotus]

The phonological system is not bad, and I do like how the voiceless nasals and liquids make it seem less SAE.

Thanks! That's good to hear!

As far as orthography, I would personally switch <o> and <w> around as well as <q> and <x>, but I do like the system other than that.

In hindsight, yeah, that makes a whole lot more sense.

Obviously, a phonology is more than just a list of phonemes. Do you have have any suggestions concerning allophony or what have you? I plan to include pre-aspirated stops (like Icelandic and Faroese have) as allophones of the regular stops, the difference being of course that I've retained voicing contrast from Old Norse. (Post-)aspirated stops as allophones for initials is obvious, but I don't want to make things too English-like.

[opipik]

It's more logical to write the front rounded vowels <ü ẅ ö>

[Duns Scotus]

Alright, I update my consonant and vowel charts to reflect the suggestions made to me. One other question I thought of is, how does one go about choosing a set of diphthongs, in a way that makes phonetic sense?

[Jonlang]

I like it! I'm a big fan of Icelandic, Old Norse and Old English. The only thing I'd add is /ɬ/ because I love that sound. I look forward to seeing more.

[Duns Scotus]

I like it! I'm a big fan of Icelandic, Old Norse and Old English. The only thing I'd add is /ɬ/ because I love that sound. I look forward to seeing more.

Thanks! I appreciate it! Strangely, the popularity of /ɬ/ has always confused me, since I find it to be rather unappealing. One thing I want to play with is that, apparently, Old Norse featured a series of nasalized vowels that were allophones of the unnasalized varieties, when they came before nasalized consonants. Not sure if I'd like the sound of this though. Unfortunately, as far as I know, none of the Old Norse readings on YouTube (or elsewhere) have paid attention to them.

[KathTheDragon]

ɬ could be an allophone of /l̊/

[Duns Scotus]

ɬ could be an allophone of /l̊/

Hmm, that's a good idea, thanks. I'll have to play with that.

[StrangerCoug]

Alright, I update my consonant and vowel charts to reflect the suggestions made to me. One other question I thought of is, how does one go about choosing a set of diphthongs, in a way that makes phonetic sense?

You can't go too far wrong with having a diphthong pair a high vowel (your <i ü u>) and a non-high vowel (the rest). Both ways around exist in English but I like putting the high vowel second when I come up with my phonotactics. If you want to narrow it down further, some possibilities: make both elements have the same backness, make both elements have the same rounding, and/or require enough of a height difference between the two vowels (depending on your definition of "enough"). I wouldn't fault you for something like <æi aü au> and nothing else.

[Alon]

The distinction between <a> and <æ> is a bit weird to me - it implies four degrees of vowel height in both front and back vowels. In North Germanic languages, such a system is unstable, which is why Swedish and Norwegian have had a chain shift /o: u:/ > /u: ʉ:/. In French, this is the system in oral vowels, but there are many mergers in various dialects: my understanding is that most merge /a ɑ/, many merge /ø œ/, and some merge /e ɛ/.

Is there a vowel length distinction, as in the mainland Germanic languages?

What phonotactics are you thinking of including?

[Duns Scotus]

The distinction between <a> and <æ> is a bit weird to me - it implies four degrees of vowel height in both front and back vowels. In North Germanic languages, such a system is unstable, which is why Swedish and Norwegian have had a chain shift /o: u:/ > /u: ʉ:/. In French, this is the system in oral vowels, but there are many mergers in various dialects: my understanding is that most merge /a ɑ/, many merge /ø œ/, and some merge /e ɛ/.

Okay, thanks. This is exactly the sort of thing I'm looking for. I guess I need to do more research into Swedish and Norwegian. I'm potentially interested in some sort of pitch accent system, so I will definitely have to look at how this works there.

Is there a vowel length distinction, as in the mainland Germanic languages?

Yeah, vowel length will be phonemic. While we're on the subject, is having pitch accent and phonemic vowel length too excessive?

What phonotactics are you thinking of including?

I'm not quite sure yet. I've been doing some research on Old Norse. As far as I can surmise, as of yet, it looks like Old Norse syllable structure is something like (C)(C)(j)V(C)(C).

[mèþru]

If it's for personal use, don't add sounds you do not like. As for allophony, see this thread. Ring above is used orthographically only for <u a>, so I don't think it works as a consonant diacritic (I would use dot below or <h>.) For /x ɣ/, I would use <x ŕ>, or <ch gh> for the Germanic flavour. In Icelandic, the voiceless sonorants are not indicated by diacritics. They are marked by preceding and/or succeeding voiceless consonants (contexts in which the voiced ones do not appear) or with a preceding silent <h> (when there are only preceding and/or succeeding vowels, contexts in which both types of sonorant appear). <fnd> is also used for /mt/ and <fnt> for /m̥t/.
Almost every French dialect has merged /a ɑ/ - it is the French equivalent of English's /ʍ/. /ø œ/ and /e ɛ/ are cases of partial merger - /e ɛ/ contrast in final position and /ø œ/ contrast in closed monosyllables. In fact, /ɔ o/ follow the same rule as /ø œ/. Continental North Germanic usually has its lowest back vowels around where <ɛ> is, but I think that your scheme is phonemically possible (but unlikely to remain in descendants).

While we're on the subject, is having pitch accent and phonemic vowel length too excessive?

Both Swedish and Norwegian have both of those.

[Nortaneous]

not having /z/ seems weird, i guess you could justify it by having a recent shift to... /j/? (since there's no j_0)

i like 8-bit-ascii orthographies so i'd say something like
<hm m hn n>
<p b t d k g>
<f v th dh s (z) x w h>
<hl l hr r y>
<i ü u>
<é ô o>
<e ö o>
<ä a>

and with all the consonants having voiceless-voiced pairs, you could have sino-tibetan-like initial consonant alternation, maybe for the causative, e.g.

dhäk 'be eaten' > thäk 'eat'
worm 'go' > xorm 'transport'
bü 'fly' > pü 'throw'
ymén 'fall' > shmén 'drop'
ölf 'hit' > hölf 'cause to hit'

as for allophony, low-mid and high-mid vowel contrasts are unstable; you could have /e 2 o/ be realized as [Ei 9y Ou] or [i@ y@ u@], at least in some circumstances -- maybe there's a split before different consonants but there are no phonemic contrasts yet. so e.g. [i@ y@ u@] before velars and [Ei 9y Ou] everywhere else, or [i@...] before nasals and [Ei...] everywhere else, or...

/{/ is also unstable, so you could have it become [e@] in some environments, esp. if you have [i@ y@ u@]. and then /O/ could maybe become [o@], but probably in different environments... and if you do all that, you could have marginal contrasts like in some dialects of AmEng, where, say, /O/ is usually [o@] before nasals and voiced consonants and [O] everywhere else, but one set of words has [O] where the regular rule would predict [o@]

another thing that seems likely is restricted distribution; maybe high-mid vowels can't appear with a following consonant, /E/ can't appear before a velar, etc.

if /j/ is from *z you could have /s/ palatalize to [j_0] or [C] before /e i/

[Duns Scotus]

You can't go too far wrong with having a diphthong pair a high vowel (your <i ü u>) and a non-high vowel (the rest). Both ways around exist in English but I like putting the high vowel second when I come up with my phonotactics. If you want to narrow it down further, some possibilities: make both elements have the same backness, make both elements have the same rounding, and/or require enough of a height difference between the two vowels (depending on your definition of "enough"). I wouldn't fault you for something like <æi aü au> and nothing else.

Thanks! This is good to know. I will just have to try it out, and see what I like the sound of and what flows well with the rest of the language.

As for allophony, see this thread.

Thanks, that thread looks like it has some good information.

Ring above is used orthographically only for <u a>, so I don't think it works as a consonant diacritic (I would use dot below or <h>.) For /x ɣ/, I would use <x ŕ>, or <ch gh> for the Germanic flavour. In Icelandic, the voiceless sonorants are not indicated by diacritics. They are marked by preceding and/or succeeding voiceless consonants (contexts in which the voiced ones do not appear) or with a preceding silent <h> (when there are only preceding and/or succeeding vowels, contexts in which both types of sonorant appear). <fnd> is also used for /mt/ and <fnt> for /m̥t/.

Well, I'm just following the IPA conventions, which is to use a ring, albeit below instead of on top. I think it looks better on top, and AFAIK, there's no system which uses a ring above these letters to mean something else, so I don't think it should cause too much trouble. Personally, I don't really like digraphs because of the potential ambiguity they cause. And in any case, the orthography is a romanization scheme I'm using in mean time, until I can develop a bespoke writing system (and of course, to communicate the language to others). I know in Icelandic they aren't indicated orthographically, but in my phonology, the voicing contrast is fully phonemic, whereas, per my understanding, the exact status of them in Icelandic is at least controversial. Ideally, to a native speaker, if one existed, /m̊/ and /m/ would sound as different as /k/ and /g/ do to a native English speaker.

Almost every French dialect has merged /a ɑ/ - it is the French equivalent of English's /ʍ/. /ø œ/ and /e ɛ/ are cases of partial merger - /e ɛ/ contrast in final position and /ø œ/ contrast in closed monosyllables. In fact, /ɔ o/ follow the same rule as /ø œ/. Continental North Germanic usually has its lowest back vowels around where <ɛ> is, but I think that your scheme is phonemically possible (but unlikely to remain in descendants).

Okay, it sounds like I need to look over my vowel system a bit more. Thanks for information/examples.

While we're on the subject, is having pitch accent and phonemic vowel length too excessive?

Both Swedish and Norwegian have both of those.

Good to hear!

not having /z/ seems weird, i guess you could justify it by having a recent shift to... /j/? (since there's no j_0)

Well, to be fair, it seems that Old Norse didn't have /z/ either, while having most of the same consonants in its inventory. I believe there's some evidence to suggest that it had the voiceless liquids/nasals at least as allophones.

and with all the consonants having voiceless-voiced pairs, you could have sino-tibetan-like initial consonant alternation, maybe for the causative...

While not the exact system you quoted, I did have something which plays with voicing contrasts in mind. Thanks for your other suggestions too, re: the vowel system.


Now, switching gears for a moment. I had an idea for allophony. What if /m̊/ /n̊/ /l̊/ /r̊/ all become murmured/breathy voiced intervocalically? Or is that not realistic as opposed to full voicing?

[mèþru]

Breathy voice sounds reasonable, but I heavily doubt my knowledge on the subject. Slack voice is sometimes described as half-voicing, so that might be even better.

[Alon]

Swedish and Norwegian have word tones. It's not the same as the pitch accent of Japanese or Ancient Greek, in which a word may have an accented syllable/mora, distinguished by high pitch. Systems with word tone, and two possible tones on each word, e.g. Shanghainese and Swedish, are called pitch accent as well, but are different from the Ancient Greek pitch accent. Swedish and Norwegian in fact have a Germanic stress accent (and Swedish can have two adjacent stressed syllables).

[Duns Scotus]

Swedish and Norwegian have word tones. It's not the same as the pitch accent of Japanese or Ancient Greek, in which a word may have an accented syllable/mora, distinguished by high pitch. Systems with word tone, and two possible tones on each word, e.g. Shanghainese and Swedish, are called pitch accent as well, but are different from the Ancient Greek pitch accent. Swedish and Norwegian in fact have a Germanic stress accent (and Swedish can have two adjacent stressed syllables).

Hmm, that's really interesting. I had something more in mind like Ancient Greek (which I am more familiar with).

[Duns Scotus]

Hello: I have a bit of an update on my phonology.

So I had an idea: I've been reading up on Estonian, which I learned distinguishes between three vowel lengths. I found this really attractive, so I was thinking of incorporating this into my vowel system. The way it would work is that there would be three different lengths. monophthongs would exist in all three lengths, whereas diphthongs could exist in just two, lengths 2 and 3. Furthermore, if I add a triphthong or two, then they would only exist in the third length.

Does that system make sense? Is it too excessive?

[mèþru]

Japanese has the Ancient Greek type of pitch accent and two phonemic lengths.
As for Estonian, as far as I can understand from Wikipedia, extra long vowels occur only in stressed syllables, and there are no regular long vowels in stressed syllables (the contrast is essentially between stressed and unstressed long vowel syllables in the same position). Interestingly, falling pitch occurs in Modern Estonian on these extra long vowels.

[Frislander]

Properly phonemic three-way length contrasts do occur in Dinka and Mixe.

[snappdragon]

My only critique is that I can't pronounce the trill (which makes it confusing when I'm speaking Spanish "DOES HE MEAN CAR OR EXPENSIVE? DOES HE MEAN BUT OR DOG??") so if I turn it into an approximant, will you be mad at me?