Zompist Bboard Again

Here are some Mixtecan cognate sets:

Amuzgo tsʔa̤³ 'chili pepper'
Amuzgo tsʔã̤³ 'person'

Cuicatec ḭ²ja² 'chili pepper'
Cuicatec ḭ¹jã¹ 'person'

Mixtec ʃa³ʔa³⁽⁵⁾ 'chili pepper'
Mixtec ɲã³ʔã³⁽⁵⁾ 'woman'

Triqui jaʔah³ 'chili pepper'

Here are some more Mixtecan cognate sets:

Amuzgo tsʔwa⁵ 'cord'
Amuzgo tsʔwa̤³ 'market'

Cuicatec ḭ⁴βa̰¹ 'thread'
Cuicatec ḭ⁴βa̰¹ 'market'
Cuicatec ḭ²βa̰⁴ 'frost'

Mixtec ʃuʔ³βa¹⁽¹⁵⁾ 'thread'
Mixtec ʃaʔ¹βi³ 'market'
Mixtec ʃuʔ³βa¹⁽¹⁵⁾ 'frost'

Triqui juʔβeh³² 'thread'
Triqui juʔβeː³² 'market'
Triqui juʔβeʔ³ 'frost'

I hope you enjoyed these Mixtecan cognate sets.

vlad wrote: ↑Fri May 15, 2020 6:00 amI hope you enjoyed these Mixtecan cognate sets.

I did!

I forgot to explain how the correspondences work.

Amuzgo deletes vowels in root-initial syllables, and turns initial j into ts.

Cuicatec turns jVʔV into ḭjV, CVʔwV into CV̰βV̰, and combining the two, jVʔwV into ḭβV̰.

Mixtec turns j into ɲ before a nasal vowel, and j into ʒ before an oral vowel. (The dialect I'm using as an example further devoices ʒ to ʃ, but that's not widespread.) I can't explain why the vowels in the Mixtec word for 'market' are different from everything else.

The dialect of Triqui used here is fairly conservative. Though there's another dialect which deletes vowels in root-initial syllables, similar to Amuzgo:

ʔweh³e 'thread'
ra-ʔwe³ 'market'
ʔweʔ³e 'frost'

It also loses long vowels and adds an echo vowel after a glottal, as you can see.

vlad wrote: ↑Fri May 15, 2020 6:00 am ...

That's good. Thank-you.

Here are some stupid minimal pairs in Proto-Mixtec:

*ʔɨ́ɨ̨́ 'one'
*ʔɨ̀ɨ̨̀ 'nine'

*ʔònį̀ 'three'
*ʔònę̀ 'eight'

*ʔùtʃè 'seven'
*ʔùtʃì 'ten'

I enjoyed them thoroughly. We should try finding sets across otomangue, some folks doubt it is a real family.

Rad Aghast wrote: ↑Mon Jun 21, 2021 3:39 pm I enjoyed them thoroughly. We should try finding sets across otomangue, some folks doubt it is a real family.

Including myself. I can hardly find any plausible cognate sets between Mixtecan and any other branch of Oto-Manguean.

Here are some between Oto-Pame and Tlapanec-Mangue:

ma-tha “rio” (Tlap.) (first syllable is "big")
dö-the “rio” (Oto) (first syllable is "big")
tehe “water” (Oto
ndéhe “water” (Maz)
intawi “water” (Mtz.)
ndà ”water” (Ocu.)
tæ/ndæ ”water” (CP.)
(Compare also with Zapotec inda ”water” and Mixtecan ndute)


i’ya “coyote” (Tlap.)
‘yo “dog” (Oto)
šu’yo “coyote” (Ocu.)

šana “monte” (Tlap.)
šɔn-thø “monte” (Oto.) (thø = hill)

šudu “espalda” (Tlap)
tsudu ”espalda” (Azo.)
šɨthɨ “espalda” (Maz.)
šɨthɔ “espalda” (Oto)

a3hnga:23 ”palabra” (Tlap.)
hyö ”palabra” (Aca.)
nìŋhä “palabra” (NP.)
inyhää “palabra” (Mtz.)

thää “say” (Tlap.) / yä “contestar”
thä-di “contestar” (Aca.)
tømti “contestar” (Mtz.)

itha “olote” (Tlap.) /ya:3 “mazorca” (Tlap)
thö ”mazorca” (Aca.)
thä “mazorca” (Mzq.)
thu “mazorca” (Ocu.)
thwä “maiz” (CP.)


iše/iša “tree” (Tlap.) (I consider i a classifier)
tsa/za “tree” (Oto)
ntsa “tree” (Ocu.)
sá “tree” (Mtz.)


tsiakhe “fuerza” (Tlap.)
zakhi “fuerza” (Oto.)
nzaki (fuerza del cuerpo”(Mzq.)

gakhii “fuerte” (Tlap.) (see also tsiakhe “fuerza”)
zakhi “fuerte” (Aca.)
zedi “be strong” (Mzq.) < *zati/tsati
nzaki “fuerza” (Mzq.)
màgá ”strength” (Chch.)/ magä́ “strong” (Chch.)


ruwa “to rain” (Tlap.)
muwa “drip”(Azo)
‘wäi “to rain” (Oto.)

re’e “flor” (Tlap.)
dì’ì “flower” (Azo.)
di’i “flower” (Sub.)
tøni “flower” (Oto)
togn “flower” (CP.)

rambo ”tomate” (from Tl.-Subtiaba *tambo)
tempoši “tomate” (Oto. Acazulco) (note how this is xitoma:tl but backwards.... weird)
dimaši “tomate” (Mzq.)
remɔ́šɨ “tomate” (Maz.)
mpaaȼhi “tomate” (Ocu.)
nombai “tomate” (CP.)

mikhü: “bitter” (Tlap.)
nkhü “bitter” (Mzq)
khö̌’ö “bitter” (Maz.)

gon “luna” (Tlap.) (there are quite a few sets where a velar stop corresponds to a sibilant or affricate, usually with the same voicing)
zɔna “luna” (Aca.)
zä̀nä “luna” (Mez., Maz.)

go’o “woman/female” (Tlap.)
tsu/zu “woman/female” (Aca. Mzq.)

gumba “animal hembra” (Tlap.) – *zu-phani “female animal” (hypothetical Otomi form. Zu- is used as a prefix for female animals in Otomi, and zuphani would be readily intelligible to speakers of Acazulco Otomi)

šayä ”uña” (Tlap.)
saha “uña” (Mzq)
šä “uña”(Aca.)

Here's a couple of sets I think may fit your "thread"
šmida “blusa” (Tlap.) /( hmedo “skirt” (Azo.) /ahmu “pants”)
hwit’wa “pants” (Oto)
mahwi “quechquemitl” (oto)
šme3 “sew” (Tlap.)
‘wet’i “sew” (Oto.)

You appear to be new. Welcome. It seems the thing for new people is to offer them pickles and tea.

I am new to this particular manifestation of the ZBB. But thanks for the pickles.

Just found a fun one between mixtec and otomi:
vìkǒ “cloud”
vikǒ “feast”

ngo "feast"
gü "cloud"

I am guessing a protoform is *miko

Here are some possible cognate sets I found for Eastern Oto-Manguean (Mixtecan + Popolocan-Zapotecan). I want to make clear that I am personally skeptical that any of this is real. I'm trying to do a devil's-advocate-type thing here. Also I know Mixtecan a lot better than the other languages so there's more likely to be mistakes with those

Transcription is IPA.

Proto-EOM *ʔjaʔ 'chili pepper'
→ Proto-Mixtecan *jaʔaʔ (metathesis + echo vowel)
→ → Amuzgo tsʔa
→ → Cuicatec ḭja
→ → Mixtec jaʔaʔ
→ → Triqui jaʔah
→ Proto-Popo-Zapotecan *ʔɲãʔ (nasalization conditioned by final glottal stop??)
→ → Proto-Popolocan *hɲa
→ → → Mazatec hɲa
→ → → Popoloca hna (ɲ → n)
→ → Proto-Zapotecan *kii-ʔnãʔ (also ɲ → n)
→ → → Chatino kinãʔ
→ → → Zapotec kiiʔnaʔ

Proto-EOM *ʔja 'eagle'
→ Proto-Mixtecan *jaʔa (metathesis + echo vowel)
→ → Amuzgo ka-tʃi, ki-tʃi (strange development but there are a few similar examples; ka-/ki- are animal prefixes)
→ → Cuicatec ḭja
→ → Mixtec jaʔa
→ → Triqui (not found)
→ Proto-Popo-Zapotecan *ʔja (no nasalization because no final glottal stop)
→ → Proto-Popolocan *hja
→ → → Mazatec hja
→ → → Popoloca (not found)
→ → Proto-Zapotecan *-ʔja
→ → → Chatino kʷi-ʔja, kʷa-ʔja
→ → → Zapotec kʷe-ssi-ʔja (kʷe- is an animal prefix shared with Chatino, I don't know what -ssi- is)

Proto-EOM *ʔjoʔ 'vine/rope'
→ Proto-Mixtecan *joʔoʔ (metathesis + echo vowel)
→ → Amuzgo tsʔɔ
→ → Cuicatec ḭju
→ → Mixtec joʔoʔ
→ → Triqui jaʔaa
→ Proto-Popo-Zapotecan *ʔɲũʔ (again, nasalization seemingly conditioned by final glottal stop)
→ → Proto-Popolocan *ʔɲu (I can't explain why this one has *ʔ → ʔ while the above have *ʔ → h)
→ → → Mazatec na-ʔɲu
→ → → Popoloca ʔno
→ → Proto-Zapotecan (not found)

Proto-EOM *ʔjo 'flea'
→ Proto-Mixtecan *joʔo (metathesis + echo vowel)
→ → Amuzgo ka-tsʔɔ, ki-tsʔɔ (ka-/ki- are animal prefixes)
→ → Cuicatec ḭju
→ → Mixtec tɨ-joʔo (tɨ- is an animal prefix)
→ → Triqui tʃaʔa (the animal prefix coalesced with the j- of the root to form tʃ, this happens in many Mixtec varieties as well)
→ Proto-Popo-Zapotecan *ʔjo (no nasalization because no final glottal stop)
→ → Proto-Popolocan *ʔja (again we have ʔ instead of h)
→ → → Mazatec ʃi-ʔja
→ → → Popoloca (not found)
→ → Proto-Zapotecan *kʷe-ʔju (*kʷe- is an animal prefix)
→ → → Chatino kʷiʔju
→ → → Zapotec *kʷeʔju

Proto-EOM *ʔjõ "tooth"
→ Proto-Mixtecan *joʔõ "tooth", *noʔõ (← *n-joʔõ) "teeth"
→ → Amuzgo tsiʔ-nʔõ (literally "teeth bone"; -nʔõ reflects the PM plural form)
→ → Cuitatec ð-ḭjũ (ð- is a possession prefix; ḭjũ reflects the PM singular form)
→ → Mixtec noʔõ (from the PM plural form)
→ → Triqui (not found)
→ Proto-Popo-Zapotecan *ʔɲũ
→ → Proto-Popolocan *ʔɲu
→ → → Mazatec ne-ʔɲu
→ → → Popolocan nenoni (maybe? not sure what's going on here)
→ → Proto-Zapotecan *(l)le-ʔja (no nasalization??)
→ → → Chatino leʔja
→ → → Zapotec (l)leʔja

That looks pretty promising to me, I'll see if I can find Oto-Pame or Tlapanec cognates for those.

I really think the family is solid - I just think the processes are so difficult to spot because they involve so many aspects of vowels that all influence eachother (tone, nasality, quality), and then the preference for short mono-syllabic roots make it hard to find cognates that are super convincing because the vowel correspondences are so hard to figure out and you really have to rely on mostly the consonants when you spot possible cognates - with the result that they are less convincing and harder to argue for.

otomí tooth is ts'í as bone in Amuzgo.
Flea is ʔɔ
chili is i (nasal hightone)

There's a paper by Kaufman available online here that lists a bunch of Oto-Manguean etyma, but doesn't list any reflexes or sound changes to back them up. And I'm having trouble finding evidence myself.


He lists *syu7ma as EOM for "tail". The forms for "tail" in the various branches of EOM are:

Amuzgo: tsʔã (← *jaʔã)
Cuicatec: ndukuðeʔẽ (from ðeʔẽ "butt")
Mixtec: suʔwẽ
Triqui: tuneʔ
Zapotec: xu-kwa:7na ~ xxi-kwa:7na (this is Kaufman's own reconstruction)

The only one of these that looks anything like *syu7ma is Mixtec. They mostly don't even look related to one another. I can't reconstruct a Proto-Mixtecan form for "tail", let alone a Proto-EOM form.


He lists *so as an EOM form, with alleged reflexes meaning "stone" in in Mazatecan and Amuzgo, and "quern" (metate) in Mixtecan.

Mazatec for "stone" is ndjoho, which doesn't look like it comes from *so.
Popoloca for "stone" is iʂo, which I guess could come from *so.
Amuzgo for "stone" is tshɔʔ. Is he claiming this comes from *so? Because Amuzgo ts- regularly reflects *j (y), and I believe this is cognate with Mixtec juuʔ "stone".

Mixtec for "metate" is josoʔ, but I do not believe this derives from an etymon meaning "stone", but rather from an etymon meaning "flat". It is a doublet of josoʔ "plain" (distinguished from "metate" by tone).
Amuzgo has su meaning both "flat" and "plain", and the word for "metate" is tshɔʔ-su "flat stone" (the Mixtec equivalent would be *juuʔ josoʔ).
Cuicatec has juðu "metate" and juðu "plain" (distinguished by tone, as in Mixtec).
Triqui has ttaa "plain" and tto "metate". (But also note that there is a placename ta-joh ~ to-joh "plain of years", which shows a variant of "plain" with o instead of a.)


He lists the Amuzgo-Mixtecan for "chilli pepper" as *la7ah and "squash" as *lauken, and I have no idea on what basis he reconstructs *l in these words. The evidence I see points to *j (y). I also don't know why he reconstructs a diphthong *au in "squash".


He reconstructs *la as a Proto-Oto-Manguean form meaning "quicklime" in Mazatecan and Chn, and "ashes" in Amuzgo-Mixtecan and Mazatecan (again).

Like with "chilli pepper" and "squash", he's reconstructing *l for Mixtecan j (y) and I don't know why. I would reconstruct it as something like *jaha.
Popoloca for "quicklime/ashes" is nteso. Mazatec for "ashes" is ndesu, and "quicklime" is ʃuhɲu. I don't see how either of those could be from *la, unless there's some crazy sound changes I don't know about.

Yes, Kaufman's unpublished reconstruction work cannot really be trusted. It may serve as inspiration sometimes. But he gives his readers no ways to check his arguments. Someone has to sit down and do the work from the bottom up. I think Zapotecan and Mixtecan is pretty far in terms of reconstructions. Gudschinsky's Popolocan reconstruction work is well regarded as is Bartholomews Otopamean - but I think they both need an update too.

I've made a post summarizing some of what I've figured out about Mixtecan.

I'd appreciate if anyone would like to proofread it or generally check that it makes sense.

https://nahuatlahtolli.wordpress.com/20 ... -mixtecan/

I don't know much about Mixtec -- but thanks for that post. It's a great resource.

The post is very clear, and I don't have much to comment on except a few typos:

"is not always good quality. " > of good quality.
"Accurately understanding the phonology of modern Mixtecan languages is necessary for reconstructing their protolanguages." > protolanguage (singular"
"*k palatalizes to tʃ before a front vowel in both Cuicatec and Mixtec, though preceding vowel changes mean this does not always affect the same words. " (I'm not sure what you mean here)

Ares Land wrote: ↑Thu Sep 02, 2021 3:20 am "is not always good quality. " > of good quality.

I don't have a problem with the original wording.

"Accurately understanding the phonology of modern Mixtecan languages is necessary for reconstructing their protolanguages." > protolanguage (singular"

Protolanguages is intentional, because each branch has its own protolanguage.

"*k palatalizes to tʃ before a front vowel in both Cuicatec and Mixtec, though preceding vowel changes mean this does not always affect the same words. " (I'm not sure what you mean here)

Cuicatec and Mixtec both had a change of k → tʃ before front vowels, but a front vowel in Mixtec might correspond to a back vowel in Cuicatec, so you can't assume that Mixtec tʃ always corresponds to Cuicatec tʃ (or vice versa).

I know nothing about Mixtecan but you've made me want to learn more so I'd say thats a good sign. I'm always interested in learning more about the historical linguistics of indigenous languages of the Americas but I'm much more familiar with the Northwest Coast, so some of my questions here might be naive.

It might help if you had a chart showing the basic sound correspondences between languages in the "Proto-Mixtecan phonology" section before you go into the subsequent developments of each sound.

I'm curious about the velarized alveolars in Amuzgo, since I don't see many (or any?) examples of contrasting non-velarized ones?

The different reflexes of *nune "avocado" vs *nune "bean" vs *nune "water" are really interesting. I don't know what the *tˠæ- prefix does in Amuzgo, but is it possible that the "avocado" and "bean" words go back to the same root? (maybe dumb question, cause they are very different plants/foods...)

Could the *n > *r shift in those Triqui examples be due to dissilimation before the medial *n?

I think the reconstructions for *ɬ and *l are the most problematic, as you acknowledge, due to lack of cognates.

Basically, I enjoyed reading it and its very clear. This thread has inspired me to read up more on Mixtecan/Oto-Manguean, so maybe I can make some more educated comments later.

Skookum wrote: ↑Fri Sep 03, 2021 10:19 pmI'm curious about the velarized alveolars in Amuzgo, since I don't see many (or any?) examples of contrasting non-velarized ones?

Amuzgo has a contrast between velarized consonants and palatalized consonants with no corresponding plain consonants, similar to Irish.

I don't know what the *tˠæ- prefix does in Amuzgo

It means "fruit".

but is it possible that the "avocado" and "bean" words go back to the same root? (maybe dumb question, cause they are very different plants/foods...)

They also have different tones.

Could the *n > *r shift in those Triqui examples be due to dissilimation before the medial *n?

Could be.