Tungusic sound changes

[bradrn]

This thread is intended as a central place to discuss, share and collate sound changes from Tungusic languages.

(I don’t actually know anything about the family, so I’ll just leave the thread here in case anyone else has something to say…)

[Tropylium]

Since this was asked, here are some outlines of consonant changes, primarily from the materials of an Intro to Tungusology course I took with Janhunen. The original source seems to be some publication (or also course materials?) by Doerfer from sometime between 1973 and 1984. I do not know if the A ~ B and A, B notation mean something different or both just mean variation in reflexes.

There would be some more detail easily available also in the EDAL preface (including in particular a lot more clusters) but this may not be entirely reliable and I might wait for the long-coming Routledge handbook of Tungusic.

Abbreviations:
M = (standard?) Manchu
S = Southern a.k.a. Nanaic (Na = Nanai, Ul = Ulch, Ok = Orok)
U = Udegeic (Ud = Udege, Oc = Oroch)
N = Northern (So = Solon = southern Evenki, Ng = Negidal, Ev = some northern dialect of Evenki, L = Even a.k.a. Lamut)

The starting Proto-Tungusic inventory:
*p *t *č *k
*b *d *ǯ *g
*s *x
*m *n *ŋ
*w *l *r *j
Presumably č ǯ = [tʃ dʒ] but I'll maintain the notation for Proto-Tungusic.
*x only occurs initially, and could have been something else like [kʰ] or [q]; it is reflected as /k/ in Alchuka, a small Jurchenic variety.

An additional PTg segment that Janhunen has elsewhere in the course materials, and which seems to be found in most modern Tungusic languages, is *ɲ. EDAL also reconstructs *š (for a correspondence Manchu ʃ, s ~ non-Manchu *č) and rejects *w (occurs mostly between vowels, replaced by *b).

Unconditional changes:
*č ǯ > ts dz in Na Ud Ev L (this info seems to be contradicted by most descriptions of Ev L I've seen, also by EDAL)
*ǯ > dz in Ng (ditto)
*č > s in So (EDAL has ʃ)
*č *ǯ > t d in Ok
*č > tʃ ~ tś in M
*ǯ > dz ~ d in M
*k > k ~ h in M
*s > s ~ ʃ in M
*s > h in L

Palatalizations:
*p / _i > s in Ud (presumably first > *x)
*p / _i > h (zero, s) in Ev (ditto)
*t d / _i > tś dź in M (presumably [tɕ dʑ])
*t d / _i > ts dz in Na
*s / _i > s, h in M
*č / _i > tʃ in Ok
*č / _i > ʃ in So
*x / _i > s in S
*x / _i > zero (s) in Ud
*x / _i > zero (n, s) in M
*ŋ / _i > n in Ng

Initial changes:
*p > f in M (not yet Proto-Jurchenic: some smaller varieties have /pʰ/ or /pf/)
*p > x in U Ng
*p > h in Ev L
*p > zero in So
*k > k ~ x in So
*x > x in Ul Ok
*x > h in Na
*x > zero in M U N (not yet Proto-Jurchenic: Jurchen has /h/)
*ŋ > g, w, zero in M
*ŋ > ŋ, m, g, w, zero in Na (jeez that's a mess)
*ŋ > ŋ, n, g, w, zero in Ul
*ŋ > ŋ, w, zero in Ud

Medial mostly-lenitions:
*-p- > f ~ zero in M
*-p- > p ~ zero in S Oc
*-p- > f ~ zero in Ud
*-p- > w ~ g in So (is this a typo for w ~ ɣ?)
*-p- > w in Ng Ev L
*-b- > zero in S
*-b- > w ~ zero in U
*-b- > w ~ ɣ in So
*-b- > w in Ng Ev L
*-č- > -s- in Ud
*-ǯ- > dz~zero in M
*-g- > g ~ ɣ in So Ng
*-g- > zero in M S U
*-k- > x in So Ng
*-k- > k ~ x in Or
*-k- > ʔ ~ x ~ g in Ud
*-k- > zero in S
*-ŋ- > ŋg in M
*-ŋ- > g ~ zero in Na Ul
*-ŋ- > ŋ ~ zero in Ok Oc
*-r- > r (j) in Na Ul
*-r- > zero (j) in Oc
*-r- > zero in Ud
*-r- > j in Ng

Final *r
*-r > zero in M Ok U Ng
*-r > l ~ zero in Ul

Sonorant + *s clusters:
*ms > ŋg in M (presumably first > *ns > *nh)
*ms > ps in Ok
*ms > mVh in Ud
*ms > nd in So
*ms > mn in Ng Ev
*ms > mr in L
*ns / _i > ŋgʲ in M (presumably is phonemically /ŋg/)
*ns / _a > h in M
*ns > (n)t in Na Ul
*ns > (t)t in Ok
*ns > s (~ h) in Oc
*ns > h in Ud
*ns > nd in So
*ns > nn, nd, nr in "some" Ev L dialects
*ns > n in Ng
*ls > lh in M
*ls > lt in S
*ls > kt in Oc (sic! compare *rC > kC/gC in Ud)
*ls > lVh in Ud
*ls > l in Ng
*ls > lr in L
(*ms > ms in Na Ul Oc; *ns and *ls do not seem to remain intact anywhere)

*r + stop clusters:
*rp > p ~ rf in M
*rp > rp ~ p in So
*rp > lp in Ok
*rp > tp in Ng
*rb > b in M
*rb > lb in Ul Ok Ng
*rp *rb > kp gb in Ud
*rk *rg > ts dz in M
*rk *rg > jk (j)g in Na
*rk *rg > tʃ dʒ in Ul
*rk *rg > t d in Ok
*rk *rg > k g in Ud
(presumably general *rk *rg > *jk *jg in pre-M and Proto-S)
*rk > kk in So
*rk > jk ~ śk, tk in Ng (no idea what is ś, if not typo for š = ʃ)
*rg > rg ~ gg in So
*rg > jg ~ dg in Ng
*rp *rb *rk *rg > pp bb kk gg in Oc

[bradrn]

Since this was asked, here are some outlines of consonant changes, primarily from the materials of an Intro to Tungusology course I took with Janhunen. The original source seems to be some publication (or also course materials?) by Doerfer from sometime between 1973 and 1984. I do not know if the A ~ B and A, B notation mean something different or both just mean variation in reflexes.

I appreciate your transcription of these changes, but how reliable is this source? It would be good if you could track down the original source where these sound correspondences were published.

[Moose-tache]

Having combed through plenty of tungusic dictionaries for a conlang, this data matches everything I've recorded. For example, the languages that derive x/h from p versus the ones that derive it from k are listed accurately. The Manchu changes are accurate, from what I remember. It's at the very least Not Utter Bullshit, which is more than I can say for some sound changes I've seen copy-pasted around the internet. The only obvious flaw is some missing changes, especially for Nanai. For example, the medial b > w change.

[bradrn]

Having combed through plenty of tungusic dictionaries for a conlang, this data matches everything I've recorded. For example, the languages that derive x/h from p versus the ones that derive it from k are listed accurately. The Manchu changes are accurate, from what I remember. It's at the very least Not Utter Bullshit, which is more than I can say for some sound changes I've seen copy-pasted around the internet. The only obvious flaw is some missing changes, especially for Nanai. For example, the medial b > w change.

OK, that’s really great to hear! I’d just like some better sourcing for these changes before we add them.

[Moose-tache]

After consulting my Janhunen and Vovin, I'm beginning to suspect there is no monograph where a Tungusic scholar lays out "here are all the sound changes" with any kind of scientific rigor. The Altaic nerds tried to do that in their landmark piece of tree-genocide the EDAL, but honestly if Your Honor will admit the EDAL as admissible evidence, my vibes-based approach should be more than sufficient.

Maybe if I have time, I'll dig up my Tungusic cognate list. Most of the sound changes, for the consonants at least, are pretty clear from that.

[bradrn]

After consulting my Janhunen and Vovin, I'm beginning to suspect there is no monograph where a Tungusic scholar lays out "here are all the sound changes" with any kind of scientific rigor. The Altaic nerds tried to do that in their landmark piece of tree-genocide the EDAL, but honestly if Your Honor will admit the EDAL as admissible evidence, my vibes-based approach should be more than sufficient.

I don’t know what EDAL is. At least my aim is to include all scholarly sources with sound changes, though it sounds like this one would deserve a score of 0/3 on quite a lot of our criteria.

On the other hand… the ‘vibes-based approach’ was more or less the approach behind the existing ID, and it’s exactly the approach I’d like to avoid with this one. As I’ve said before, restricting the project to transcribing published works may have its own problems, but overall it’s easier to validate (and is a necessary component of the research anyway).

[Man in Space]

I’ve been meaning to ask: So if we have multiple sources about sound changes in a given language, we’d just have multiple sections on the page?

[bradrn]

I’ve been meaning to ask: So if we have multiple sources about sound changes in a given language, we’d just have multiple sections on the page?

That’s what I’ve been doing so far. On the website you can see it for e.g. Takuu (Austronesian).

[TomHChappell]

….
I don’t know what EDAL is.
….

Etymological Dictionary of Altaic Languages, I think.

[fusijui]

Moose-tache is right I think, in that there's no single monograph on Proto-Tungusic and its sound changes. I do have the Routledge volume, and it still doesn't really provide that, in satisfactory detail.

(Well, IMO, at least. In any case it's not as blatantly B.S. as EDAL's Tungusic circus.)

I still have a lot of issues with how even a good, careful set of sound changes (like Tropylium's here), it only "works right" by only looking at carefully picked vocabulary. And I come across exceptions/counter-examples that, if they're intra-Tungusic loans, can't be produced by any other language's sound changes. Etc. It just feels very sloppy and often circular, after all these years. Tungusic historical linguistics has always been the garbage dump and whipping-boy of "Altaic"/North Asian philology TBH.

[Moose-tache]

If we're not allowed any original scholarship (i.e. looking at Tungusic and saying "yup, that's a sound change") then we just have to skip Tungusic. But this approach raises a bigger question:

Why is this project not just a bibliography?

I have Charles Julian's work on Proto-Iroquoian on my hard drive. So do you. So does anyone who's going to use the Index Diachronica. If there's no room for authorial addition, if the only thing the ID does is tell me to go read my Julian, I'm not sure what the point is. What are we advancing?

[bradrn]

If we're not allowed any original scholarship (i.e. looking at Tungusic and saying "yup, that's a sound change") then we just have to skip Tungusic. But this approach raises a bigger question:

Why is this project not just a bibliography?

I have Charles Julian's work on Proto-Iroquoian on my hard drive. So do you. So does anyone who's going to use the Index Diachronica. If there's no room for authorial addition, if the only thing the ID does is tell me to go read my Julian, I'm not sure what the point is. What are we advancing?

The point is that this is a database, not just a bibliography. It’s comparable to other databases, like WALS and PHOIBLE.

And the key thing about being a database is that it combines all the data from all these different sources, in a consistent and searchable form. That makes it possible to answer questions which cannot otherwise be answered: for instance, ‘what are the attested sources of /ŋ/?’. Or, ‘under what conditions has /t/→/s/ occurred?’. Or, ‘which languages in this family have undergone sound change X→Y?’. Or… anyway, you get the idea, hopefully.

Moose-tache is right I think, in that there's no single monograph on Proto-Tungusic and its sound changes. I do have the Routledge volume, and it still doesn't really provide that, in satisfactory detail.

Even if there’s no single monograph, surely there’s at least individual papers which set out the sound changes for individual languages or subgroups? That’s what we’ve mostly been relying on for the other families.

[Man in Space]

The problem is that there are a lot of papers that focus on or report only a single sound change (or a small handful). (Pama-Nyungan seems particularly prone to this.)

[bradrn]

The problem is that there are a lot of papers that focus on or report only a single sound change (or a small handful). (Pama-Nyungan seems particularly prone to this.)

I’m not sure that’s a problem per se. The website already lists a couple of papers like that for Polynesian.

[Moose-tache]

I don't think WALS is a good example, because there is a huge amount of authorial input and decision making in WALS, but I see what you mean about making it searchable.

As for individual languages within Tungusic... You'd be surprised. I know for sure some of these languages don't have that, even though some of the sound changes are pretty obvious. This is partly because, thanks to the Altaicists, serious scholars are reluctant to offer any firm reconstruction of Proto-Tungusic. There is an unofficial list of uncontroversial sound changes, the ones that are obvious from looking at a Swadesh list. And there are plenty of discussions about diagnostic changes used to differentiate sub-branches. That's about as good as it gets. If you want an exhaustive list of sound changes, even for just one language, you're going to need to visit tinfoil hat country.

[bradrn]

I don't think WALS is a good example, because there is a huge amount of authorial input and decision making in WALS

Hmm, how so?

As for individual languages within Tungusic... You'd be surprised. I know for sure some of these languages don't have that, even though some of the sound changes are pretty obvious. This is partly because, thanks to the Altaicists, serious scholars are reluctant to offer any firm reconstruction of Proto-Tungusic. There is an unofficial list of uncontroversial sound changes, the ones that are obvious from looking at a Swadesh list. And there are plenty of discussions about diagnostic changes used to differentiate sub-branches. That's about as good as it gets. If you want an exhaustive list of sound changes, even for just one language, you're going to need to visit tinfoil hat country.

I suspect this is true for quite a lot of languages and families, barring only the most heavily-studied ones (e.g. English). Probably very few languages will get an exhaustive list of changes, and I’m fine with that. It’s a sign that we’re sticking to stuff that can be reliably validated.

(If you don’t want a reliable source, well, we already have a version of the ID which doesn’t worry about it!)

[Moose-tache]

I don't think WALS is a good example, because there is a huge amount of authorial input and decision making in WALS

Hmm, how so?

Please, I am begging you. Tell me you're joking. Even if you don't mean it.

[Zju]

I don't think WALS is a good example, because there is a huge amount of authorial input and decision making in WALS

Hmm, how so?

Please, I am begging you. Tell me you're joking. Even if you don't mean it.

I thought WALS was pretty uncontroversial? Of course there has to be decision making if you only want to sum up linguistic diversity.

[bradrn]

Hmm, how so?

Please, I am begging you. Tell me you're joking. Even if you don't mean it.

I thought WALS was pretty uncontroversial? Of course there has to be decision making if you only want to sum up linguistic diversity.

Ah, I hadn’t noticed that Moose-tache said ‘authorial input and decision making’. Of course, there is plenty of decision making in WALS, though I think ‘huge authorial input’ is stretching it a bit.

The thing about WALS is that its individual data points tend not to be very reliable, at least when I’ve checked. It’s often hard to track down the source of a claim in WALS, even when they give a reference. Occasionally, a datum is outright wrong. This is why I’ve placed such an emphasis on reliability in this ID.