Linguistic Miscellany Thread

[bradrn]

I’m trying to make a new set of sample files for my Brassica sound change applier, to show off its new and improved support for tone. I was thinking that perhaps something Sinitic might work for the purpose.

So, does anyone know of any references about this which I could look at? It could be Middle Chinese to any of the modern varieties, or Old Chinese to Middle Chinese… just as long as it’s something I can extract some sound changes and sample words from. (It doesn’t even need to be Sinitic necessarily, as long as it’s from a tonal family and reasonably comprehensive.)

[zompist]

So, does anyone know of any references about this which I could look at? It could be Middle Chinese to any of the modern varieties, or Old Chinese to Middle Chinese… just as long as it’s something I can extract some sound changes and sample words from. (It doesn’t even need to be Sinitic necessarily, as long as it’s from a tonal family and reasonably comprehensive.)

I think you have access to a university library... the info you need will be in Axel Schuessler's Etymological Dictionary of Old Chinese (for the Middle Chinese tones-- OC wasn't tonal) and Jerry Norman's Chinese (for the sound changes in the major dialects).

[bradrn]

So, does anyone know of any references about this which I could look at? It could be Middle Chinese to any of the modern varieties, or Old Chinese to Middle Chinese… just as long as it’s something I can extract some sound changes and sample words from. (It doesn’t even need to be Sinitic necessarily, as long as it’s from a tonal family and reasonably comprehensive.)

I think you have access to a university library... the info you need will be in Axel Schuessler's Etymological Dictionary of Old Chinese (for the Middle Chinese tones-- OC wasn't tonal) and Jerry Norman's Chinese (for the sound changes in the major dialects).

Thanks!

(And I do have access to a university library, yes.)

[zompist]

I think you have access to a university library... the info you need will be in Axel Schuessler's Etymological Dictionary of Old Chinese (for the Middle Chinese tones-- OC wasn't tonal) and Jerry Norman's Chinese (for the sound changes in the major dialects).

Thanks!

(And I do have access to a university library, yes.)

Great! Just to save you some time, some key bits:

1. Schuessler's entries have the modern Mandarin as a keyword, followed by the MC in parentheses.
2. MC had four tones, A B C D. The dictionary leaves A unmarked, but gives B and C. If the word ends in a stop, it's D, also unmarked.
3. The sound changes from MC to the modern dialects are given in Norman, in the Dialects chapters. They are mostly pretty simple.
4. The first stage was a division into upper and lower tones-- thus the 4 tones become 8. 1/2 = A, 3/4 = B, etc.
5. Thus Norman will refer to MC tone 1 to 8 and give their modern tone contours.
6. Mandarin tones are a mess, largely because tone D could go to any of the four modern tones. The other dialects are much more well-behaved.

Wikipedia actually has the same info as Norman (the tone correspondences)... well, I assume they're about the same, I haven't compared them. Where you'd need Schuessler is actually getting examples.

It looks like Wiktionary has MC reconstructions too... multiple ones, in fact. Just be consistent in which one you use.

[bradrn]

2. MC had four tones, A B C D. The dictionary leaves A unmarked, but gives B and C. If the word ends in a stop, it's D, also unmarked.
3. The sound changes from MC to the modern dialects are given in Norman, in the Dialects chapters. They are mostly pretty simple.
4. The first stage was a division into upper and lower tones-- thus the 4 tones become 8. 1/2 = A, 3/4 = B, etc.
5. Thus Norman will refer to MC tone 1 to 8 and give their modern tone contours.
6. Mandarin tones are a mess, largely because tone D could go to any of the four modern tones. The other dialects are much more well-behaved.

Most of this I’m already aware of, actually, but thanks for the summary!

(I’m reading through Norman now, and enjoying it a lot. I’ll find a copy of Schuessler’s dictionary if Norman doesn’t give sufficient examples.)

[bradrn]

I’ve read through most of Norman now. As far as I can tell, it doesn’t actually have many sound changes. For Min initials, he even says outright that ‘no attempt is made to indicate diachronic correspondences’ (p264). So I don’t think this will work for the purposes of a sample sound change file (which needs to be able to evolve protoforms to current forms). I’ll have a look at Schuessler next.

[Travis B.]

I have noticed some inconsistency in the application of American/Canadian Raising* to German names/loans in the dialect here. Take the following:

• [əe̯]senhower
• H[ae̯]ser (a car dealership here in the Milwaukee area)
• H[ae̯]senberg
• [ae̯]~[əe̯]genvalue, [ae̯]~[əe̯]genvector
• Schn[əe̯]der
• T[ăĕ̯]~[ə̆ĕ̯]tonia

Has anyone else noticed such inconsistency in other NAE varieties?

* The dialect here has weak allophonic raising of /aʊ/ from [ɑɔ̯] to [ʌ̆ŏ̯], which strictly speaking makes it have Canadian Raising, but in general it patterns with American dialects more than Canadian dialect with regard to raising.

[zompist]

I’ve read through most of Norman now. As far as I can tell, it doesn’t actually have many sound changes. For Min initials, he even says outright that ‘no attempt is made to indicate diachronic correspondences’ (p264). So I don’t think this will work for the purposes of a sample sound change file (which needs to be able to evolve protoforms to current forms). I’ll have a look at Schuessler next.

I thought what you wanted was the changes in tone. All the non-tone changes are going to be a taller order... Schuessler may or may not help, he's more interested in the transition from OC to MC.

[Man in Space]

Would Baxter and Sagart be useful here?

[bradrn]

Would Baxter and Sagart be useful here?

Hmm, perhaps. It looks like the relevant book is their Old Chinese: A New Reconstruction (2014).

[zompist]

Would Baxter and Sagart be useful here?

Hmm, perhaps. It looks like the relevant book is their Old Chinese: A New Reconstruction (2014).

I haven't seen their work, and I would hate to discourage an interest in OC reconstruction, which is fascinating, but this probably takes you away from your original intention: OC was not a tonal language.

Did you see Richard W's post on Thai?

[bradrn]

Would Baxter and Sagart be useful here?

Hmm, perhaps. It looks like the relevant book is their Old Chinese: A New Reconstruction (2014).

I haven't seen their work, and I would hate to discourage an interest in OC reconstruction, which is fascinating, but this probably takes you away from your original intention: OC was not a tonal language.

But MC was!

Did you see Richard W's post on Thai?

No, where was that?

[zompist]

Did you see Richard W's post on Thai?

No, where was that?

Right here

[bradrn]

Did you see Richard W's post on Thai?

No, where was that?

Right here

Ooh… I had absolutely no idea that was there. Thanks!

[bradrn]

Would Baxter and Sagart be useful here?

Hmm, perhaps. It looks like the relevant book is their Old Chinese: A New Reconstruction (2014).

Managed to get access to this book. Let’s see how I go…

[bradrn]

Further on Sinitic, I turned up something very interesting. Apparently the code here is able to predict Mandarin and Cantonese forms from Middle Chinese: https://en.wiktionary.org/wiki/Module:ltc-pron/predict. If only it had references — or, for that matter, any explanation whatsoever…

(The idea is that it’s called by this code to generate Chinese pronunciation tables for Wiktionary: https://en.wiktionary.org/wiki/Module:ltc-pron.)

[fusijui]

If you want to demo it with Chinese, I'd recommend deciding whether to work with "dialect" material or else with the literary reconstruction tradition, and exclude the other -- just to keep from having apples and oranges gumming up the works when that's not the point of the exercise (IIUC).

[bradrn]

If you want to demo it with Chinese, I'd recommend deciding whether to work with "dialect" material or else with the literary reconstruction tradition, and exclude the other -- just to keep from having apples and oranges gumming up the works when that's not the point of the exercise (IIUC).

I’m not quite sure what you’re talking about here, sorry.

[bradrn]

Further on Sinitic, I turned up something very interesting. Apparently the code here is able to predict Mandarin and Cantonese forms from Middle Chinese: https://en.wiktionary.org/wiki/Module:ltc-pron/predict. If only it had references — or, for that matter, any explanation whatsoever…

(The idea is that it’s called by this code to generate Chinese pronunciation tables for Wiktionary: https://en.wiktionary.org/wiki/Module:ltc-pron.)

Well, I managed to transcribe this into Brassica notation, and it seems to work, but the result isn’t very revealing. The code uses a big lookup table to convert all the Middle Chinese rhymes into their Mandarin equivalents, so it doesn’t really look like traditional sound changes.

But this causes me to wonder… is it even possible to adequately describe Sinitic evolution with traditional sound changes? Because the nucleus and coda are so tightly bound together, they often seem to evolve as a single unit — yielding, in effect, 160 unconditional sound changes. I made a big table to see if there’s any regularities, but if there are, they seem well-hidden. (e.g. in Standard Mandarin, {*ap,*at}→a, {*ep,*et}→ie… but {*op,*ot}→{a,e}. *jep, *jet, *jop evolve the same as *ep, *et, *op, but *jot gives ie rather than e. And don’t ask me what happens to *it.)

[zompist]

But this causes me to wonder… is it even possible to adequately describe Sinitic evolution with traditional sound changes? Because the nucleus and coda are so tightly bound together, they often seem to evolve as a single unit — yielding, in effect, 160 unconditional sound changes. I made a big table to see if there’s any regularities, but if there are, they seem well-hidden. (e.g. in Standard Mandarin, {*ap,*at}→a, {*ep,*et}→ie… but {*op,*ot}→{a,e}. *jep, *jet, *jop evolve the same as *ep, *et, *op, but *jot gives ie rather than e. And don’t ask me what happens to *it.)

I suspect that you're seeing that what comes out of the model-- the rhymes-- is what goes into it. If you recall Norman's explanations, the reconstruction is a marrying of the comparative method with the MC analysis of the Qieyun rhyme tables. And all the actual data we have (modern dialect forms, Qieyun groups) is oriented around initials and rhymes.

OTOH, there might be deeper patterns if you have the right phonemicization... and Mandarin vowels are not that. Norman offers a phonemic analysis which you might play with; FWIW he thinks Mandarin has no /o/.