Sound Change Quickie Thread

[Xwtek]

1. Can the glottal stop do anything other than disappearing or turning into /h/?

Yes, being treated as voiceless consonant, so if you devoice vowel between voiceless consonant, you can do that. Another way is to combine it with another stop to form an ejective. Or to combine with vowel to form a creaky voice or tone.

4. In the same language, is it plausible to have [+short] {w,y,r} [+short] → [+long] / when vowels are the same? e.g. tæjæs → tæːs, ʔewe → ʔeː

No, you must either unconditionally delete the intervocalic semivowels or only /j/ for front vowel or /w/ for unrounded vowel.

[bradrn]

1. Can the glottal stop do anything other than disappearing or turning into /h/?

Yes, being treated as voiceless consonant, so if you devoice vowel between voiceless consonant, you can do that. Another way is to combine it with another stop to form an ejective. Or to combine with vowel to form a creaky voice or tone.

I’m aware of those possibilities already. I suppose the question I really meant to ask was: can the glottal stop, without combining with anything else, change into a phoneme other than /ʔ/? (Sorry for being imprecise in the first place though.)

4. In the same language, is it plausible to have [+short] {w,y,r} [+short] → [+long] / when vowels are the same? e.g. tæjæs → tæːs, ʔewe → ʔeː

No, you must either unconditionally delete the intervocalic semivowels or only /j/ for front vowel or /w/ for unrounded vowel.

I’m not quite sure what to do now… Whimemsz has already said almost exactly the opposite. I think I’ll wait for now until a couple more people reply with their opinions.

(Also, did you mean ‘rounded’ at the end there, or was the reference to ‘unrounded’ intentional?)

[Whimemsz]

.

[Kuchigakatai]

I’m aware of those possibilities already. I suppose the question I really meant to ask was: can the glottal stop, without combining with anything else, change into a phoneme other than /ʔ/? (Sorry for being imprecise in the first place though.)

I honestly don't think so.

You can probably convert it into a [j] or [w] or [ŋ] by proxy though, by dropping it from syllable-initial position and then installing a rule that says syllables must start with a consonant. I think this is basically what happened in Cantonese at some point in its history, since in its early and mid-20th century form (when it started to be more carefully described) syllables that previously probably started with [ʔ] on the surface followed by a vowel now started with /j w ŋ/, i.e. /i/ now was /ji/, /y/ was /jy/ [ɥy] and /u/ was /wu/ [wuː] or /ŋu/ [ŋʊ]. (Words could begin with /o/ [ʔɔː] or /a/ [ʔa:] though.)

I’m not quite sure what to do now… Whimemsz has already said almost exactly the opposite. I think I’ll wait for now until a couple more people reply with their opinions.

Personally I have a hunch that sound change might be okay.

For what it's worth, Arabic tends to do some vowel coalescence when /w j/ are surrounded by short vowels, while generally maintaining /w j/ if one of the vowels is long. This shows that semivowel-dropping and subsequent vowel coalescence can be sensitive to the surrounding vowels; you don't have to drop all intervocalic /w j/ to make this kind of thing happen. Arabic doesn't only do it for identical vowels, but I think that might be okay in your conlang.

Examples of vowel coalescence: *zawara > zaara 'he visited', *zuwira > ziira 'he was visited', *daaʕiwun > daaʕin 'caller', *banaya > banaa 'he built', *baaniyun > baanin 'builder', *yansayu > yansaa 'he forgets', *buyiʕa > biiʕa 'it was sold' (but buniya 'it was built', *duʕiwa > duʕiya 'he was called'). Note also that "awa" and "awi" are allowed in the result of the more complex *munTawiyun > munTawin 'introverted', *munTawayan > munTawan 'enclosed, folded in'.

Examples of /w j/ maintenance: Haawala 'he tried', axawaatun 'sisters', musaawaatun 'equality', tanaawubun 'alternation', mutanaawibun 'alternating (adj.)', futuuwatun 'youth', numuuwun 'growth (nominative)', numuuwin 'growth (genitive)', ʕanaawiinu 'titles', zawaayaa 'corners', šatawiiyun 'of the winter', uurubbiiyuuna 'Europeans', as-siyaasiiyu 'the politician (nominative)', as-siyaasiiyi 'the politician (genitive)', tazaayudun 'increase (noun)', mutazaayidun 'increasing (adj.)'.

[Raholeun]

I’m not quite sure what to do now… Whimemsz has already said almost exactly the opposite. I think I’ll wait for now until a couple more people reply with their opinions.

There is really no reason to be insecure about these things and ask for validation on matters of sound change. You really are the best judge in this case. What works for me is to pronounce the sequence of sounds, in this case *tæjæs and *ʔewe, a bunch of times without proper articulation. Slur them a bit. For me personally, they become [tæ:s ~ tɛːs] and [ʔeːʊ̯] quite rapidly (guess my L1).

In my conlang Sataw, I have a sound change *ʁ → sː which I am not sure is attested. Yet, it arose quite naturally through my own sloppy articulation (through *x → *ç → *s). So does this sound change have the proper pedigree? Maybe not, but I judge her realistic enough as it developed unconsciously and it feels instinctive.

[mèþru]

I think /ʔ/ can turn into /ŋ ɴ/ through rhinoglottophilia.

[Kuchigakatai]

I think /ʔ/ can turn into /ŋ ɴ/ through rhinoglottophilia.

I remember Nortaneous, long time ago, once mentioning some minority language that had nasal mutation (as part of a morphological process), and the language used /ŋ/ as the nasal counterpart of /ʔ/. I can't remember what the language was at all though.

[Xwtek]

No, you must either unconditionally delete the intervocalic semivowels or only /j/ for front vowel or /w/ for unrounded vowel.

I’m not quite sure what to do now… Whimemsz has already said almost exactly the opposite. I think I’ll wait for now until a couple more people reply with their opinions.

(Also, did you mean ‘rounded’ at the end there, or was the reference to ‘unrounded’ intentional?)

What to do is ignore him because that answer is ridiculous and makes no sense. (In general my advice to people in this thread would be to ignore Akangka's pronouncements, which frequently are stated very confidently but seem to have no basis in actual attested sound changes or known patterns of changes)

Well, sorry. I'm still learning.

[Xwtek]

I think /ʔ/ can turn into /ŋ ɴ/ through rhinoglottophilia.

Apparently similiar phenomena occurs in Standard Tibetian before velar consonant.

(Here, h̩ is actually h with underdot and it's originated from Proto Tibeto-Burma *ʔ)

h̩č'am-pa > ñč'ampa
h̩t'ag-pa > nt'agpa
h̩gul-ba > ŋgul-ba

However, this is probably because it's originally ʔə̃.

Source: http://stedt.berkeley.edu/pdf/JAM/Matis ... philia.pdf

I think /ʔ/ can turn into /ŋ ɴ/ through rhinoglottophilia.

I remember Nortaneous, long time ago, once mentioning some minority language that had nasal mutation (as part of a morphological process), and the language used /ŋ/ as the nasal counterpart of /ʔ/. I can't remember what the language was at all though.

Are you sure that it's not retention of /ŋ/ where it's lost in other places, leaving nasal mutation?

[mae]

-

[bradrn]

So I just checked this thread and my question got a lot of replies — thanks so much everyone! I’d better go through them one by one…

What to do is ignore him because that answer is ridiculous and makes no sense. (In general my advice to people in this thread would be to ignore Akangka's pronouncements, which frequently are stated very confidently but seem to have no basis in actual attested sound changes or known patterns of changes)

I think that’s a bit unfair… Akangka must get it right some of the time at least. Besides, this is incredibly unhelpful for me as someone trying to learn which sound changes are realistic, since you don’t give any reasoning — you’re simply stating that Akangka is usually wrong. Please next time could you at least go into a bit more detail as to why his reasoning is incorrect, so I can actually learn something?

I’m aware of those possibilities already. I suppose the question I really meant to ask was: can the glottal stop, without combining with anything else, change into a phoneme other than /ʔ/? (Sorry for being imprecise in the first place though.)

I honestly don't think so.

You can probably convert it into a [j] or [w] or [ŋ] by proxy though, by dropping it from syllable-initial position and then installing a rule that says syllables must start with a consonant. I think this is basically what happened in Cantonese at some point in its history, since in its early and mid-20th century form (when it started to be more carefully described) syllables that previously probably started with [ʔ] on the surface followed by a vowel now started with /j w ŋ/, i.e. /i/ now was /ji/, /y/ was /jy/ [ɥy] and /u/ was /wu/ [wuː] or /ŋu/ [ŋʊ]. (Words could begin with /o/ [ʔɔː] or /a/ [ʔa:] though.)

Good idea! I might use this, actually.

I’m not quite sure what to do now… Whimemsz has already said almost exactly the opposite. I think I’ll wait for now until a couple more people reply with their opinions.

Personally I have a hunch that sound change might be okay.

For what it's worth, Arabic tends to do some vowel coalescence when /w j/ are surrounded by short vowels, while generally maintaining /w j/ if one of the vowels is long. This shows that semivowel-dropping and subsequent vowel coalescence can be sensitive to the surrounding vowels; you don't have to drop all intervocalic /w j/ to make this kind of thing happen. Arabic doesn't only do it for identical vowels, but I think that might be okay in your conlang.

Examples of vowel coalescence: *zawara > zaara 'he visited', *zuwira > ziira 'he was visited', *daaʕiwun > daaʕin 'caller', *banaya > banaa 'he built', *baaniyun > baanin 'builder', *yansayu > yansaa 'he forgets', *buyiʕa > biiʕa 'it was sold' (but buniya 'it was built', *duʕiwa > duʕiya 'he was called'). Note also that "awa" and "awi" are allowed in the result of the more complex *munTawiyun > munTawin 'introverted', *munTawayan > munTawan 'enclosed, folded in'.

Examples of /w j/ maintenance: Haawala 'he tried', axawaatun 'sisters', musaawaatun 'equality', tanaawubun 'alternation', mutanaawibun 'alternating (adj.)', futuuwatun 'youth', numuuwun 'growth (nominative)', numuuwin 'growth (genitive)', ʕanaawiinu 'titles', zawaayaa 'corners', šatawiiyun 'of the winter', uurubbiiyuuna 'Europeans', as-siyaasiiyu 'the politician (nominative)', as-siyaasiiyi 'the politician (genitive)', tazaayudun 'increase (noun)', mutazaayidun 'increasing (adj.)'.

I didn’t know Arabic already did this — that does help me a lot.

I’m not quite sure what to do now… Whimemsz has already said almost exactly the opposite. I think I’ll wait for now until a couple more people reply with their opinions.

There is really no reason to be insecure about these things and ask for validation on matters of sound change. You really are the best judge in this case. What works for me is to pronounce the sequence of sounds, in this case *tæjæs and *ʔewe, a bunch of times without proper articulation. Slur them a bit. For me personally, they become [tæ:s ~ tɛːs] and [ʔeːʊ̯] quite rapidly (guess my L1).

I do try to do this, but I have two problems:

1. If I just try to guess where sounds go based on phonetics only, I get a bunch of very regular, very boring sound changes. In and of itself that isn’t so much of a problem, but often this makes the derived words sound very similar to the base words. Plus, if you look through (say) the Index Diachronica, almost every set of changes listed there has at least one non-boring change.
2. If I do come up with an interesting change, I do use it, but I’m often not quite sure if it's realistic… so I post it on here. (It frequently isn’t, confirming my theory that I’m just not very good at figuring out interesting yet realistic changes.)

I think /ʔ/ can turn into /ŋ ɴ/ through rhinoglottophilia.

I forgot about rhinoglottophilia! This could certainly be useful for me — I was looking for a way to add more /ŋ/.

[akam chinjir]

Is it plausible for tj to be stable for (say) 100 years? 200? (I ultimately want it to turn into c, but would prefer for that not to happen right away. I'm assuming that if it became tʃ first, which would otherwise be fine, it would be very unlikely to end up as c.)

[Nortaneous]

I think /ʔ/ can turn into /ŋ ɴ/ through rhinoglottophilia.

I remember Nortaneous, long time ago, once mentioning some minority language that had nasal mutation (as part of a morphological process), and the language used /ŋ/ as the nasal counterpart of /ʔ/. I can't remember what the language was at all though.

That might've been Nias, where /n/ is one possible nasal alternant of /ʔ/ (the other is /g/), but words where /ʔ/ mutates to /n/ originally had initial *V or *ʔ, whereas words with /ʔ/~/g/ originally had initial... velars or uvulars that debuccalized; a quick search turned up an overview paper that doesn't specify the precise proto-segments.

You can also have ʔ > h (some dialects of Nahuatl?), so in principle ʔ > anything you can get to from /h/.

I don't know of direct ʔ > ŋ, but h > ŋ is attested from Nyole - which lenited most of its stops, developing *h > ŋ from *p, and then lost contrastive prenasalization (reintroducing plain stops)... so there's a synchronic rule of n + ŋ > p.

Is it plausible for tj to be stable for (say) 100 years? 200? (I ultimately want it to turn into c, but would prefer for that not to happen right away. I'm assuming that if it became tʃ first, which would otherwise be fine, it would be very unlikely to end up as c.)

Icelandic

[akam chinjir]

Is it plausible for tj to be stable for (say) 100 years? 200? (I ultimately want it to turn into c, but would prefer for that not to happen right away. I'm assuming that if it became tʃ first, which would otherwise be fine, it would be very unlikely to end up as c.)

Icelandic

Ah, thanks.

[Kuchigakatai]

Ah, thanks.

Yeah, Icelandic has instances of [tj] that go back to Proto-Germanic (roughly 2000 years), e.g. *hwatjanan > hvetja 'to sharpen' (the actual English cognate is "to whet").

[Knit Tie]

To continue dicking around with future Englishes, would you guys say that merging the fortis and the lenis series together into lenis word-initially and word-finally, but not intervocalically, and then voicing the lenis for an overabundance of voiced consonants is a good idea?

[Nortaneous]

Would it be realistic to lengthen all nasal vowels if they have a pre-existing length contrast? (and then lose nasality, so *a *a: *a~ *a~: > a a: a: a:)

[mae]

-

[bbbosborne]

and then voicing the lenis for an overabundance of voiced consonants

what would trigger this tho?

[Xwtek]

and then voicing the lenis for an overabundance of voiced consonants

what would trigger this tho?

Assuming the lenis stop comes from English voiced stops, it's possible that they're not voiceless to begin with. In RP, voiced consonant is partially/optionally devoiced both word initially and word finally.

https://notendur.hi.is/peturk/KENNSLA/0 ... rtlen.html