Zompist Bboard Again

jal wrote: ↑Fri Mar 29, 2024 4:38 am I think the "schwa" is just the opening then closing of the channel for pronouncing the /j/.

Not really, though, because as zompist mentioned you don’t say [nəjən] — you can see in the spectrograms I shared earlier that the [nj] transition is pretty much instantaneous, whereas the [jən] takes much longer.

bradrn wrote: ↑Fri Mar 29, 2024 5:35 amNot really, though, because as zompist mentioned you don’t say [nəjən] — you can see in the spectrograms I shared earlier that the [nj] transition is pretty much instantaneous, whereas the [jən] takes much longer.

Well, I think I say something like [ɲjn]. If I try to say [njn], I do produce something like [nəjən].


JAL

bradrn wrote: ↑Fri Mar 29, 2024 5:35 am

jal wrote: ↑Fri Mar 29, 2024 4:38 am I think the "schwa" is just the opening then closing of the channel for pronouncing the /j/.

Not really, though, because as zompist mentioned you don’t say [nəjən] — you can see in the spectrograms I shared earlier that the [nj] transition is pretty much instantaneous, whereas the [jən] takes much longer.

I have [nj] here but definitely not *[jn].

zompist wrote: ↑Mon Mar 18, 2024 5:40 pm Does [ɛ] count? French -eille- is pronounced [ɛj], and produces minimal pairs like réveil [rɛvɛj] and rêvait [rɛvɛ].

(At least, that's how I was taught: [e] in rêvé vs. [ɛ] in rêvait. Maybe that distinction hasn't been maintained.)

According to Wiktionary, Italian ne/nei are [ne / nej].

Is Italian /e/ phonetically /e:/? That might work. As for French, i thought the situation there was /j/ as a complete sequence rather than part of a contour. Either way, the contour would only be in the language rather recently (from /ɛʎ/ > /ɛ.j/ > /ɛj/). However, /ɛ:/ counts just as much as /e:/. The important thing is length not height. Come to think of it, i wonder if this concept applies to non-mid vowels (/a:/ versus /aj/ /aw/, /i:/ /u:/ versus /iw/ /uj/); albeit not as strongly.

Qwynegold wrote: ↑Mon Mar 18, 2024 3:55 pmIs this rare? Swedish has /eː/ vs. /ɛj/ (the vowel is [eː] when long and [ɛ] when short), crf:

Again, the key thing is a distinction between length not height, so i would count it. However, like the French example, i thought Swedish /j/ was harder, approaching /ʝ/. Enough for it to be not quite a contour, or at least only be so recently. I keep talking about diphthongs being recent because my suspicion isn't that long mid vowels contrasting is impossible, just unstable. That said, you actually know Swedish, i'm getting info secondhand; so if i botched anything i'll take full ownership.

Starbeam wrote: ↑Mon Apr 01, 2024 12:07 pm Is Italian /e/ phonetically /e:/?

Vowel length in Standard Italian is allophonic.

Starbeam wrote: ↑Mon Apr 01, 2024 12:07 pm I keep talking about diphthongs being recent because my suspicion isn't that long mid vowels contrasting is impossible, just unstable.

Take StG for instance, which has a contrast between unrounded open-mid and close-mid long front vowels. Of course, many variations upon StG lack this very distinction.

Hallo conlangers!

I wish to share with you a thought of mine about how the High German sound shift came into being. This shift, which loosely resembles a second run of Grimm's Law but is actually quite different in detail, is generally considered to have originated in the south, in the area
roughly corresponding to today's Baden-Württemberg, Bavaria, Austria and Switzerland, in the 6th century AD. Some linguists have proposed a Celtic substratum, but that has fallen out of favour - IMHO justly, because there are two problems with it: 1. The shift resembles nothing that had happened in Celtic. 2. By the time the area became Germanic-speaking, Celtic had already been superseded by Vulgar Latin, except perhaps a few remote valleys in the Alps.

But it has occurred to me that this shift may have been caused by a Vulgar Latin substratum by a kind of hypercorrection mechanism. It has often been observed that in Romance languages, voiceless stops are never aspirated and voiced stops more strongly voiced than in Germanic languages such as English or German, to the point that the Germanic voiced stops are not recognized as voiced by many Romance speakers. This may already have been the case in 500 AD.

So what may have happened is this: When Germanic peoples had conquered what is now southern Germany, Austria and Switzerland, the Vulgar Latin speakers, in adopting the language of their new overlords, identified the Germanic voiced stops not with their own voiced stops but with their voiceless ones, and added an exaggerated aspiration to the Germanic voiceless stops which would easily tip into affricates or spirants in some environment, and thus the HIgh German sound shift took place.

This raises the question whether the Germanic sound shift ("Grimm's Law") happened in a similar way, with Celtic, an unknown IE language between Germanic and Celtic, or an unknown pre-IE language of Scandinavia as substratum language, but I think this is a different matter. Some years ago, I proposed an alternative to the PIE glottalic theory to the CONLANG mailing list. The idea is that the PIE voiceless stops once were aspirated, which would mean that the Germanic sound shift would amount to little more than the voiced unaspirated stops losing their voicing. Such a change does not need any substratum language to happen - it is just a way of filling a gap, namely the absence of a stop set with neither voicing nor aspiration, in the system. The same thing would have happened - independently - in Armenian.

WeepingElf wrote: ↑Tue Apr 02, 2024 4:00 pmSo what may have happened is this: When Germanic peoples had conquered what is now southern Germany, Austria and Switzerland, the Vulgar Latin speakers, in adopting the language of their new overlords, identified the Germanic voiced stops not with their own voiced stops but with their voiceless ones, and added an exaggerated aspiration to the Germanic voiceless stops which would easily tip into affricates or spirants in some environment, and thus the HIgh German sound shift took place.

The only flaw I could think of is that if they identified the voiced stops with their voiceless ones, I would've assumed that they'd actually lose voicing, and you'd end up with a system of voiceless stops and voiceless affricates?


JAL

jal wrote: ↑Wed Apr 03, 2024 3:31 am

WeepingElf wrote: ↑Tue Apr 02, 2024 4:00 pmSo what may have happened is this: When Germanic peoples had conquered what is now southern Germany, Austria and Switzerland, the Vulgar Latin speakers, in adopting the language of their new overlords, identified the Germanic voiced stops not with their own voiced stops but with their voiceless ones, and added an exaggerated aspiration to the Germanic voiceless stops which would easily tip into affricates or spirants in some environment, and thus the HIgh German sound shift took place.

The only flaw I could think of is that if they identified the voiced stops with their voiceless ones, I would've assumed that they'd actually lose voicing, and you'd end up with a system of voiceless stops and voiceless affricates?

Actually, this is the system in many Upper German varieties -- it just happens that traditionally the voiceless stops are written with <b d g>.

Travis B. wrote: ↑Wed Apr 03, 2024 10:04 amActually, this is the system in many Upper German varieties -- it just happens that traditionally the voiceless stops are written with <b d g>.

I thought this was a feature of the central German dialects, not the South ones?


JAL

jal wrote: ↑Wed Apr 03, 2024 10:32 am

Travis B. wrote: ↑Wed Apr 03, 2024 10:04 amActually, this is the system in many Upper German varieties -- it just happens that traditionally the voiceless stops are written with <b d g>.

I thought this was a feature of the central German dialects, not the South ones?

The pattern of traditional Central German dialects is to have fricatives instead of voiceless stops for StG /b g/ in many cases (which is reflected in StG by the standard pronunciation of -ig and in informal writing by things like writing Tag as Tach).

Travis B. wrote: ↑Wed Apr 03, 2024 10:37 amThe pattern of traditional Central German dialects is to have fricatives instead of voiceless stops for StG /b g/ in many cases (which is reflected in StG by the standard pronunciation of -ig and in informal writing by things like writing Tag as Tach).

I must've confused them with Upper Saxon. As Wikipedia says:

Wikipedia wrote:The Upper Saxon varieties outside the Ore Mountains can be easily recognized by the supposed "softening" (lenition) of the voiceless stop consonants /p/, /t/ and /k/. Speakers of other dialects hear these as if they were "b", "d" and "g" respectively. In reality, these are merely non-aspirated versions of the same /p/, /t/ and /k/, a widespread feature among Central German dialects, as opposed to strongly aspirated [pʰ], [tʰ] and [kʰ] in dominant German dialects.

JAL

I would point out that there are some phenomena which loosely resemble the High German sound shift (affrication and frication of the historically voiceless stops, and devoicing of the historically voiced ones) elsewhere in Germanic, though few of them go as far. Danish's /t/ is often an affricate, and its lenis stops are often voiceless. And some English dialects, like Scouse, are full of affricate and fricative allophones of /t/ and /k/ especially (also sometimes /d/): e.g. a Scouse better can sound quite like German besser, though the Scouse fricative is usually less sibilant.

Can I ask a completely unrelated question - but this is a miscellany thread after all. How do direct-inverse systems handle indirect objects? I get the general idea that if the agent is higher than the patient in a person-hierarchy so 2nd person is higher than 1st in Algonquian languages you have a direct marker. If not, an inverse marker and you can if you want (like Algonquian languages) further encode on the verb whether the two arguments are local, non-local or mixed. But what about. "I gave the spear to the boy" or I gave the spear to you". What we do with the indirect object? If you had a double object construction to deal with this like "I gave you the spear" is that direct (I am higher than a spear) or inverse (I am lower than you)?

In North America there are two main strategies. Most direct-inverse languages use both.

First, you can just add an additional argument to the verb and let context do its work.

Second, applicatives can elevate what would be an indirect object to a direct object.

Moose-tache wrote: ↑Fri Apr 05, 2024 5:43 pm First, you can just add an additional argument to the verb and let context do its work.

Wouldn't the typical pattern be that the indirect object would have less personhood/topicality/animacy than the subject but greater personhood/topicality/animacy than the direct object?

evmdbm wrote: ↑Fri Apr 05, 2024 5:23 am Can I ask a completely unrelated question - but this is a miscellany thread after all. How do direct-inverse systems handle indirect objects? I get the general idea that if the agent is higher than the patient in a person-hierarchy so 2nd person is higher than 1st in Algonquian languages you have a direct marker. If not, an inverse marker and you can if you want (like Algonquian languages) further encode on the verb whether the two arguments are local, non-local or mixed. But what about. "I gave the spear to the boy" or I gave the spear to you". What we do with the indirect object? If you had a double object construction to deal with this like "I gave you the spear" is that direct (I am higher than a spear) or inverse (I am lower than you)?

I checked Valentine's grammar of Nishnaabemwin (which is Algonquian), and it appears that ditransitives use the object slots on the verb for what we'd call the indirect object, e.g. Wgii-miinaan "she gave to her". The 'object' is not marked on the verb, it's simply included in the sentence. Compare ngii-daawenan "I sold it", where the inanimate object is marked on the verb.

Starbeam wrote: ↑Mon Apr 01, 2024 12:07 pm

Qwynegold wrote: ↑Mon Mar 18, 2024 3:55 pmIs this rare? Swedish has /eː/ vs. /ɛj/ (the vowel is [eː] when long and [ɛ] when short), crf:

Again, the key thing is a distinction between length not height, so i would count it. However, like the French example, i thought Swedish /j/ was harder, approaching /ʝ/. Enough for it to be not quite a contour, or at least only be so recently. I keep talking about diphthongs being recent because my suspicion isn't that long mid vowels contrasting is impossible, just unstable. That said, you actually know Swedish, i'm getting info secondhand; so if i botched anything i'll take full ownership.

It's impossible to say which feature is more important in Swedish /eː/ vs. /ɛ/, length or height. They are both important. /j/ can be [ʝ] in some dialects, but not in standard Swedish.

In Finnish you can have /eː/ vs. /ei/, but there has been a sound change in the past that turned /eː/ into /ie/, so it's hard to find minimal pairs. Except I guess there are plenty of minimal pairs of this pattern:

/ˈveneːsːæ/ (in the boat) vs. /ˈveneisːæ/ (in the boats)
/ˈsæteːsːæ/ (in the ray) vs. /ˈsæteisːæ/ (in the rays)
/ˈpurjeːsːa/ (in the sail) vs. /ˈpurjeisːa/ (in the sails)

Qwynegold wrote: ↑Sat Apr 06, 2024 6:08 am lIt's impossible to say which feature is more important in Swedish /eː/ vs. /ɛ/, length or height. They are both important.

Are there Swedish dialects which preserve the four-way distinction between /e eː ɛ ɛː/?

Qwynegold wrote: ↑Sat Apr 06, 2024 6:08 am It's impossible to say which feature is more important in Swedish /eː/ vs. /ɛ/, length or height. They are both important. /j/ can be [ʝ] in some dialects, but not in standard Swedish.

Of course it's possible. Remove length from one set of words and height distinction from the same set for different people, and ask native speakers to write what they hear.