(Inner) Skou

[bradrn]

I'm probably over-thinking this, let's just list it as a change and note Donohue's comment about "did not apply uniformly in time to the whole linguistic community".

Yeah, good idea.

On that note, I’m wondering if it might be a good idea to allow grouping sound changes? Since at the moment the two changes *ɡʷ→*ɡ and *ɡ→*j are treated separately, yet they do belong to a single chain shift.

[Travis B.]

I'm probably over-thinking this, let's just list it as a change and note Donohue's comment about "did not apply uniformly in time to the whole linguistic community".

Yeah, good idea.

On that note, I’m wondering if it might be a good idea to allow grouping sound changes? Since at the moment the two changes *ɡʷ→*ɡ and *ɡ→*j are treated separately, yet they do belong to a single chain shift.

Clearly marking chain shifts as such is probably a good idea.

[bradrn]

I'm probably over-thinking this, let's just list it as a change and note Donohue's comment about "did not apply uniformly in time to the whole linguistic community".

Yeah, good idea.

On that note, I’m wondering if it might be a good idea to allow grouping sound changes? Since at the moment the two changes *ɡʷ→*ɡ and *ɡ→*j are treated separately, yet they do belong to a single chain shift.

Clearly marking chain shifts as such is probably a good idea.

So the question then is: what would be the best approach for doing so?

[Travis B.]

Yeah, good idea.

On that note, I’m wondering if it might be a good idea to allow grouping sound changes? Since at the moment the two changes *ɡʷ→*ɡ and *ɡ→*j are treated separately, yet they do belong to a single chain shift.

Clearly marking chain shifts as such is probably a good idea.

So the question then is: what would be the best approach for doing so?

Marking them as, say, *a → *b → *c in contrast to *a → *b and *b → *c, which would stand for two separate changes occurring one after another (equivalent to *a → *c and *b → *c from the endpoint POV).

[bradrn]

Clearly marking chain shifts as such is probably a good idea.

So the question then is: what would be the best approach for doing so?

Marking them as, say, *a → *b → *c in contrast to *a → *b and *b → *c, which would stand for two separate changes occurring one after another (equivalent to *a → *c and *b → *c from the endpoint POV).

Yeah, of course that’s the obvious approach, now that you mention it!

(Though I do wonder how flexibly it can be extended to more complex chain shifts, such as the Great Vowel Shift. Perhaps it’s just not worth giving such shifts any special representation?)

[bradrn]

Clearly marking chain shifts as such is probably a good idea.

So the question then is: what would be the best approach for doing so?

Marking them as, say, *a → *b → *c in contrast to *a → *b and *b → *c, which would stand for two separate changes occurring one after another (equivalent to *a → *c and *b → *c from the endpoint POV).

Hold on… on reflection, this seems confused. A chain shift would surely be *b→*c then *a→*b, right? *a→*b→*c is simply an ordinary sound change with an intermediate step, so not a chain shift at all.

[bradrn]

I finally got around to transcribing a few more of these sound changes (to Sangke, Wutung and Proto-Vanimo). But now I have some queries:

*ɵ → e (listed in tables, but later noted that "*ʉ merg[es] with *ɵ, which splits depending on nasality")

On p182 he specifies that ‘square brackets (“[ ]”) are used to indicate a form that is not cognate’, and indeed /e/ in Table 21 is in square brackets, so I presume this actually isn’t a real sound change (but *ʉ→*ɵ is).

*b → w, p (for reflex w only, "innovatio[n] borrowed from Vanimo")
*b → m /_V[+nasal] ("innovatio[n] borrowed from Vanimo")

I believe this can be refined: *b→w only when the following vowel is not nasalised, while *b→p competes with both pathways.

*ɵ̃ → ũ
*ɵ → i (otherwise)

I can’t find this in the paper. Rather, Table 26 suggests non-nasalised *ɵ→ɵ.

[Darren]

I finally got around to transcribing a few more of these sound changes (to Sangke, Wutung and Proto-Vanimo). But now I have some queries:

*ɵ → e (listed in tables, but later noted that "*ʉ merg[es] with *ɵ, which splits depending on nasality")

On p182 he specifies that ‘square brackets (“[ ]”) are used to indicate a form that is not cognate’, and indeed /e/ in Table 21 is in square brackets, so I presume this actually isn’t a real sound change (but *ʉ→*ɵ is).

*b → w, p (for reflex w only, "innovatio[n] borrowed from Vanimo")
*b → m /_V[+nasal] ("innovatio[n] borrowed from Vanimo")

I believe this can be refined: *b→w only when the following vowel is not nasalised, while *b→p competes with both pathways.

*ɵ̃ → ũ
*ɵ → i (otherwise)

I can’t find this in the paper. Rather, Table 26 suggests non-nasalised *ɵ→ɵ.

Thanks, you're right in all cases. I should check more carefully.

[bradrn]

After a long hiatus I tried to resume transcribing these changes, and almost immediately ran into issues with *ɡʷ in Sangke:

*gʷ → tʃ (in bound prefixes, free pronouns and nouns)
*gʷ → f (in free pronouns)

The issue for the first change here is that he actually gives quite a different chronology:

The loss of the *j triggered a move to replace the palatals, and so *gʷ shifted to become a palatal stop, **j.¹⁶ This change spread through the Proto–West Coast community, and simplified to tʃ in Proto-Border.

---

¹⁶. The other high voiced stop, *g, had already been lost by this time; it is likely that the Proto–West Coast *gʷ phoneme had first lost its rounding—see 5.1.4.

So it would seem that he’s actually suggesting *ɡʷ→*ɡ→*j in Proto-West Coast, then *j→tʃ in Proto-Border. And, in fact, we’ve already transcribed these sound changes as such! So it seems like there’s no need for *gʷ→tʃ in Sangke specifically.

EDIT: And now that I think of it, this means that your *tʃ→t in Sangke is also problematic, because it would yield *gʷ→t too. In fact, I can’t find anything at all in Donohue’s paper mentioning any *tʃ.

As for *ɡʷ→f, I can’t find any place where he suggests such a change. Where did you see it? EDIT: never mind, found it in Table 16.