The Library of Akinokure

In P-J, “to drink” is “nəmu” (MJ “nomu”), with an ultra-rare nasal as the final consonant of the verb stem. That means the “m” is part of the root, not a default / dummy verb suffix “-mu”. Only the “-u” is a default segment, the obligatory verb ending.

https://en.wiktionary.org/wiki/Reconstruction:Proto-Japonic/n%C9%99m-

In P-Athabaskan, it’s “naˑn”, although also reconstructed as “naˑŋ” — hinting that the coda used to be velar, way back when, but Athabaskan fronts its velars. The onset shows nothing strange, just “n”, and the “a” vowel is long.

https://en.wiktionary.org/wiki/Reconstruction:Proto-Athabaskan/na%CB%91n

In Tlingit, it’s “naa”, just like P-A but missing the coda.

In Eyak, it’s “la”, but Eyak “l” is often the correspondent of “n” in Tlingit and P-A, which it must be here.

Clearly this word goes back to P-ND, seemingly “naˑŋ”. But we can do better by bringing Wa / Japonic into the picture.

Onset “n” is the same. The nuclear “a” could have been followed by a coda “w” in the ancestor of all forms, realized as “ə” in P-J (coda “w” -> “u”, doesn’t totally replace preceding “a”, which drags it down to “ə”). In P-ND, “aw” becomes long “aˑ”, with compensatory lengthening from the lost “w”.

The coda in P-J is “m”, but this may result from earlier “ŋʷ”, where the nasal manner remains and the labial feature becomes primary place, i.e. “m”. So either the P-ND form should have coda “ŋʷ”, or it is this in the ancestor of all families and lost its labialization in P-ND.

So the common ancestor of “to drink” in Wa, Na-Dene, and (as we’ll see) Yeniseian, was “nawŋʷ”.

That takes care of the cognate status of P-J and P-ND. Now onto Yeniseian, which bans nasals in initial position, making the search a little more difficult.

There are a pair of words from Ket that suggest “to drink” in P-Ketic was “dopʰ”.

https://en.wiktionary.org/wiki/%D0%B0%D1%80%D0%BE%D0%BF

https://en.wiktionary.org/wiki/%D0%B4%D0%BE%D0%BE

Similarly, Pumpokol has a compound word “duždop”, where “dop” is cognate with the P-Ketic root above.

https://en.wiktionary.org/wiki/du%C5%BEdop

Supposedly this reflects P-Y “gʷap”, but I doubt that. Rather, it looks like the Japonic and Na-Dene forms, which begin with “n”, show Yeniseian cognates where that has hardened into “d” (still voiced, and in the same place), in order to get around the Yeniseian ban on initial nasals.

The “o” may come from earlier “aw”, as in Japonic.

The coda “p” may be the hardening of earlier “m”, since Yeniseian doesn’t care much for “m” compared to the other nasals, even when it’s allowed in coda position.

That earlier “m” may result from even earlier “ŋʷ”, as “m” is the most common (though not only) reflex of P-Y word-final “ŋʷ” in Yeniseian descendants. E.g., P-Y “tuwVŋʷ” = “black” has descendants with “m” instead of “ŋʷ” in all 4 branches of the family.

https://en.wiktionary.org/wiki/Reconstruction:Proto-Yeniseian/tuwV%C5%8B%CA%B7

So, although it’s harder to see due to the uniquely Yeniseian dislike of nasals, the “dop” root in Ketic and Pumpokol may easily derive from a more ancestral “nawŋʷ”:

“nawŋʷ” -> “nawm” -> “nom” -> “dom” -> “dop”

That clinches the cognate status for these words across all 3 families of Wa-Dene-Yeniseian!

Again, think of how fruitless my search was when I was only comparing Japonic and Yeniseian — I saw the supposed root for “to drink” as “gʷap”, and dismissed it as irrelevant to Japanese “nomu”. But since Yeniseian bans nasals in onset, while Na-Dene does not, it was Na-Dene to the rescue yet again! I’m really growing fond of that family, it’s a far more reliable place to start searching for connections to Japonic, and Yeniseian is more for secondary follow-up or triangulation.

Worth noting that Japonic and Yeniseian show the same treatment of ancestral “ŋʷ”, making it “m”, while Na-Dene simply removes the labialization and makes it “ŋ”, and/or fronts it as in Athabaskan. But not altering it to “m”, which Tlingit and Eyak both lack. Even in P-A, “m” is rare, and absent in onset.

Yeniseian shows other cases of altering earlier 2nd-ary labialization into primary labial place, so this strategy for resolving labialized consonants seems to be a shared innovation of both Yeniseian and Japonic, meaning they share a common ancestor, and that Old World super-family was sister to the Na-Dene New World family. Not too surprising, based on the geography.

Also, Yeniseian came to be heavily agglutinative just like Japonic, whereas Na-Dene descendants are still toward the polysynthetic side of the spectrum. Probably due to a common influence, namely Uralic, the main expansionist family in Northern Eurasia. Not “contact” influence, but absorbing former Uralic speakers into Yeniseian or Wa communities.

The source of the split between Yeniseian and Japonic is when the Wa people began absorbing the Emishi / Ainu speakers in Korea and Japan. That’s the source of total reduplication in Japonic, which is absent or rare in Yeniseian and Na-Dene, but present in Ainu.

By now, I don’t think Wa / Japonic ever went through a proper Altaic stage. Altaic languages are also incorporating former Uralic speakers, and both Turkic and Mongolic have Yeniseian as another parent, although Tungusic seems to have Nivkh or something further east as its other parent.

And by now, I don’t think of Japonic as a carry-over of Yeniseian, as put through an Emishi / Ainu filter. It descends from a sister to Yeniseian, called Wa. Their common ancestor is sister to Na-Dene. One great big happy Trans-Pacific Paleosiberian family. ^_^

Continuing on topic of the senses and sense organs, there’s a root “nipo” in Japanese that underlies the verb for “to smell” (“nipopu”) and the noun form of that verb, meaning “smell, scent” (“nipopi”). Later the “p” was lost between vowels, yielding MJ “niou” and “nioi”.

https://en.wiktionary.org/wiki/%E8%87%AD#Japanese

There’s a P-ND root meaning “to feel, to sense”, and in the Na-Dene family it may refer specifically to the sense of touch and using the hand. This shows up as P-A “ni̓gʸ” = “to act with the hand, to feel, to sense”, the Eyak cognate “leʼgw”, and the Tlingit cognate with a semantic shift meaning “sickness, pain” = “néekw”.

https://en.wiktionary.org/wiki/Reconstruction:Proto-Athabaskan/ni%CC%93g%CA%B8

Athabaskan fronted its velars, so the P-ND form has a velar 2nd C, with labialization, the onset was “n”, and the vowel was high-front or perhaps “e” followed by coda “j”. That is, something like P-ND “nejkʷ”.

Since the kinds of sicknesses referred to in Tlingit include respiratory ones like a chest cold, it’s possible the P-ND meaning was general, not only about touching / hands, or that it was specifically about the nose and smelling, and Athabaskan shifted the semantics to be about touching and hands.

There is also a similar root in Tlingit for “to smell, as in sensing”, which is “neexʼ”. The P-A entry above mentions that there may have been several similar yet distinct forms that were eventually fused together, both phonetically and semantically. So this could be one of them — clearly related to sensing through the nose, beginning with “n”, high-front vowel, velar 2nd C, though a fricative instead of stop, no labialization, and ejective.

Sticking with the P-ND form with widely attested descendants, “nejkʷ”, the P-J rendering of this shows all the usual sound correspondences between P-J and Dene-Yeniseian. Since Yeniseian bans nasals initially it is absent from this little party of cognates (or it’s harder for me to spot), which emphasizes the importance of looking to Na-Dene, not just Yeniseian, for the relatives of Wa / Japonic.

Onset “n” is fine, “j” -> “i” and replaces nuclear “e” (or it was nuclear “i” to begin with, or it was nuclear “e” and raised to “i” since Japonic doesn’t like “e”). Labialized velar stop becomes primary labial stop, “p”. The only open question is the final vowel to make it CVCV — it’s “o”, and that’s not the default “p” syllable by frequency in OJ. So it was likely chosen as a modifier of the preceding consonant — to remind them that it used to be back and had rounding, i.e. it used to be “kʷ”. Another case of a single C from Wa-Dene-Yeniseian being mapped onto a CV mora in P-J:

“kʷ” -> “po” (word-finally)

That gives P-J “nipo”, QED, as the root relating to “to smell, scent”.

Speaking of labialized velars turning into “p”, and speaking of the nose and smelling and sensing, I previously detailed the cognate status of P-J “pana” and P-Y “xʷaŋ”, both “nose”. Perhaps the Wa form went through an intermediate stage of being “kʷaŋ” or “kʷan”, where it lost fricatives other than “s”. Then later it lost 2nd-ary labialization, turning “kʷ” into “p”. That is, “xʷ” -> “kʷ” -> “p”.

https://en.wiktionary.org/wiki/Reconstruction:Proto-Yeniseian/x%CA%B7a%C5%8B

https://en.wiktionary.org/wiki/Reconstruction:Proto-Japonic/pana#Etymology_2

The P-Y entry above details the various Na-Dene cognates relating to “nose” and “unpleasant odor”, so P-J “pana” has Na-Dene cognates as well, not just Yeniseian ones. Some Yeniseian descendants like Ket mean “to feel, to sense”, so that shows the ancestral Wa-Dene-Yeniseian word did refer specifically to the sense of smell, which was then generalized, and only in Athabaskan (and Eyak?) did it shift to refer to feeling with the hands.

Continuing the search for Dene-Yeniseian origins of Japonic words for body parts and organs, there’s P-J “kimo” = “liver, heart, internal organ”.

https://en.wiktionary.org/wiki/Reconstruction:Proto-Japonic/kimo

This does not immediately seem to be related to the P-Y word for “liver” = “seŋʷ”, or the proposed cognates of that on the Na-Dene side, P-A “zə̓tʼ” and Eyak “sahd” (both “liver” as well). But we’ll get to those later, and they are in fact cognates.

https://en.wiktionary.org/wiki/Reconstruction:Proto-Yeniseian/se%C5%8B%CA%B7

The semantics of internal organs is that they’re all just “internal organs,” and the sounds referring to one can shift around to refer to another. As in Japonic, where “kimo” can refer to the heart or liver in the Ryukyuan branch.

However, the coda “ŋʷ” in the P-Y word for “liver” makes me think of where “kimo” could have come from. It’s less obvious that the “k” and “s” are corresponding sounds, the P-A and Eyak words have an alveolar stop as the 2nd C, and the P-A and Eyak vowels are low / back. We’ll get to them later.

Still, there is something similar about “seŋʷ” and “kimo” — the 1st vowel is high-front, and the 2nd C is a nasal, unlike the P-A and Eyak forms.

Well, what if the ancestor of “kimo” had “ŋʷ” as its 2nd C? The labialization becomes primary, and it retains nasal manner, yielding “m”. This is like the labialized velar obstruents becoming labial obstruents in P-J, namely just “p”, preserving obstruent manner as well as labial feature.

Then after the “m” is a P-J dummy vowel, “o”, the most common “m” syllable in OJ. And perhaps that’s the most common CV pair with C being “m”, since it’s trying to tell us what the origin of “m” is in P-J. There is no “m” in Na-Dene, and it’s uncommon in P-Y (and banned initially). So where could it have descended from?

By pairing “m” with a back vowel like “o”, it says the nasal used to be back — in velar position, not labial. But since it’s a labial nasal, it must have had some kind of labial feature on it somewhere, i.e. it was “ŋʷ”, which no longer exists in the Japonic inventory. This is why “m” so heavily co-occurs with back vowels in P-J words — this supposed front consonant actually comes from a back consonant ancestor.

So yet again we see a single C from Dene-Yeniseian being mapped to a CV mora in Japonic, so that the richness of the single C inventory from D-Y can be preserved in the richness of the CV mora inventory in Japonic (~50 CV pairs), instead of its sparse single C inventory (9 C’s):

“ŋʷ” -> “mo”

So let’s pursue the other member of Na-Dene that’s not represented in that list of cognates for P-Y “seŋʷ” — Tlingit! Well well well, it’s not “liver”, but “lungs”, where the word is “keigú”. Not the exact same organ, but still a critical internal / torso organ, where we know the meanings can shift around somewhat.

This is almost identical to “kimo” — initial “k”, 1st V is high-front, and 2nd V is high-back rounded. The only difference is the velar “g” — but if the ancestor to both “kimo” and “keigú” had “ŋʷ” as the 2nd C, this accounts for both descendants — it turns into “m” in P-J as already described.

As for Tlingit, it retains velar place, perhaps spreading its labialization onto the following vowel (originally not “u”) to make it “u”, or this is a dummy vowel appended solely to preserve labialization from the preceding lost “ŋʷ”.

Tlingit only has 1 nasal — “n” — so ancestral “ŋʷ” common to P-Y and P-ND (and likely P-Wa), will show up in some other way. Tlingit did not merge all nasals into “n”, as that creates a homophony problem, which would be made worse by “l” being an allophone of “n” as well. After all, it’s known that “m” was mapped to “w” rather than merged into “n”.

So instead, “ŋʷ” was mapped to “g” or “gu”, to retain velar place and avoid homophony. Maybe “ŋ” without labialization mapped to “g”, while “ŋʷ” with labialization mapped to “gu”. IDK exactly, but it’s clear enough that “g” or “gu” could easily be the descendant of ancestral “ŋʷ”, which was lost along with all other nasals besides “n”, and like the rest of them, it was re-incarnated as a non-nasal consonant.

We have a very plausible phonetic match, and a near-perfect semantic match — winner winner, chicken dinner! As usual, Yeniseian was not the most helpful family to find cognates for Japonic — and even Na-Dene as a whole didn’t turn it up, only Tlingit. Still, both forms may harmonize into a single form in the ancestor of all 3 families.

What would that single ancestor look like? Well, the 2nd C is definitely “ŋʷ”, which is impossible to get from “t / d”, whereas “t / d” could be gotten from the Eyak-Athabaskan trend of fronting velars to palatal, then to alveolar, and hardening them into stops.

It’s also easier to weaken “k” into “s / z” than vice versa. And that is even easier when the following consonant is front, like “e / i”, or at least there’s a front-like glide “j” following it. Much rarer to harden “s” into “k” before a front vowel. Also, Eyak-Athabaskan likes fronting velars.

Probably no 2nd V — it’s part of the mapping of “ŋʷ” to another C, with the following V being a modifier indicating the preceding consonant used to be back and rounded.

The 1st V is high-front sometimes, but also low-back in others. Perhaps nuclear “a” followed by coda glide “j”, as is common in Dene-Yeniseian reconstructions.

Thus, “kajŋʷ” is the ancestor of all 3 families’ word for “liver”, which P-J, P-Y, P-A, and Eyak agree on as the meaning. Later, this shifted to “lungs” in Tlingit, whose new word for “liver” is “tlʼóog̱u / tlʼóoḵ”, of origin that I don’t understand right away.

Later, the word genericized in P-Y and P-J to mean “vital internal organ”, potentially used as a new standalone word for “heart” as in Okinawan (2nd-ary meaning, primary still “liver”), or as a building block for compounds in P-Y for “stomach” etc.

Not just finding cognates that involve Japonic — but discovering a new reconstruction for the Dene and Yeniseian families, relying crucially on Japonic for triangulation! These 3 families are interconnected far more richly than anyone ever knew…

Let’s recall that P-J “ti” means both “blood” and “milk”, so there could also be a semantic connection with “milk”.

Well well well, look at the P-J word for “white”, which is the color of milk. It’s “sirao” or “sirau” = “white” (wrongly listed as “siro-” at Wiktionary, despite the vowel alternation being “a” for combining, and “o” for exposed).

https://en.wiktionary.org/wiki/Reconstruction:Proto-Japonic/siro-

“White” cannot possibly be semantically related to “blood” or “juice, sap, soup” — but it can be related to “milk”, which is another vital liquid essence than can be drained from its source, just like juice from a fruit or sap from a tree.

So there must have been a stage near P-J where “siru” = “juice, sap, liquid essence other than blood” included “milk” as one of its referents. Probably there was a slight phonetic variation on “siru”, to lexicalize the sense of “milk”. Then from it, a new word for “white” was born — “sirau”.

Recall how unusual the vowel alternation is for “white” — combining form “sira”, exposed form “siro1”. Usually, the exposed form is “e” or “i”, and the word is reconstructed with “ai” or “ui” or “o2i” as the vowel pair, with 2nd vowel always “i”. This word having “a” for combining is fine — but the 2nd vowel being “o1”-like, is odd.

But perhaps they wanted to clarify the connection between “milk” and “siru”, which ends in “u” and is not a verb unlike most words ending in “u”. So they gave it the unusual exposed vowel “o1”, hinting at “u” as the 2nd vowel in the underlying pair (“au”), not the usual “i” as in the other basic words with vowel-alternation whose 1st vowel is “a” (e.g., “mai” = “eye”, “tai” = “hand”, etc.).

The “a” after “r” is a dummy vowel, and is the most common “r” syllable in OJ. Since “white” is far more basic and frequent than “juice, sap, soup”, “sirau” received the most-common dummy vowel for “r” syllables, while as we saw, “siru” got the 2nd-most common one (and this “u” was also included in the underlying pair of “sirau”, as the 2nd vowel of the final pair).

OK, so what happened to the Wa word for “blood”, which in P-J was replaced by “ti”? It should resemble the Dene-Yeniseian words for “blood”, which we just reviewed. But it’s worth emphasizing an alternate reconstruction for P-Y “blood”, which is “suλ” or “sur” — beginning with a sibilant, ending in a liquid.

https://en.wiktionary.org/wiki/Reconstruction:Proto-Yeniseian/su%CE%BB

Considering both the Yeniseian and Na-Dene forms, the P-J form likely had a sibilant onset, a high vowel (if Na-Dene is closer, as usual, a high-front vowel), and perhaps a rare case of a 2nd C being liquid, in the root — not part of a generic suffix as in “ra” = “plural”, “ru” = “verb ending”, etc.

Voila, “siru”, where “ru” is not the default verb ending, since this is a noun, so where the “r” may belong to the root, not a suffix. It means “juice, sap, soup” — a liquid essence of something that can be drained or rendered from the source container.

https://en.wiktionary.org/wiki/Reconstruction:Proto-Japonic/siru

Given its clear phonetic connection to the D-Y words, it may have originally meant “blood” as well. But during the semantic chain reaction caused by “kokoro” shifting from “mind” to “heart”, then “ti” shifting from “heart” to “blood”, the original word for “blood” = “siru” shifted subtly to “juice, sap, or other liquid essence like blood, but not blood itself”.

The final “u” is a dummy vowel, and is tied for 2nd-most common “r” syllable in OJ, along with “i”, but “siri” was already used for “back, behind” and avoided for homophony reasons. The most common “r” syllable is “ra”, and we will see that in the next case, for a far more common word with the same “sir” root relating to liquid essences.

Consider P-J “ti”, originally concrete and meaning “blood”, then extended to mean “life force, supernatural / spiritual essence”. It also means “milk”, which will come up later.

https://en.wiktionary.org/wiki/Reconstruction:Proto-Japonic/ti

As usual, Yeniseian is less illuminating. P-Y “blood” is “çur”, and “heart” is “puj” (which would reflect in P-J “pi”).

https://en.wiktionary.org/wiki/Reconstruction:Proto-Yeniseian/%C3%A7ur

https://en.wiktionary.org/wiki/Reconstruction:Proto-Yeniseian/puj

As we saw with P-Y “çaj” = “night”, there’s a correspondence between “ç” in P-Y and “t” in P-J. Also revealing, the Pumpokol form of “blood” begins with “t”, as with “night”, whereas the other branches have a sibilant for “blood”, as with most of the other branches for “night”.

The derived color term “çurVŋʷ” = “red” (“blood-like”) also shows the Pumpokol form beginning with “t”, as well as a possible loan of that into Arinic, whereas the others have a sibilant.

https://en.wiktionary.org/wiki/Reconstruction:Proto-Yeniseian/%C3%A7urV%C5%8B%CA%B7

So P-J looks most like Pumpokol, compared to other Yeniseian branches, which is expected since the Wa people used to belong to the Xiongnu confederation, where it was not just Yeniseian, but a branch close to Pumpokol, that was the Xiongnu elite’s original language.

The coda liquid is unusual for Yeniseian, but it will be important later, just not in the P-J word for “heart”.

P-Y “heart” = “puj” has the wrong onset. Again, Yeniseian doesn’t clarify the Japonic connection as much as it should (probably due to extensive influence of Uralic on Yeniseian compared to Na-Dene or Japonic). But good ol’ Na-Dene comes to the rescue yet again, to clear up the muddled picture from Yeniseian.

In Tlingit, “heart” is “téix̱’”. Not listed at Wiktionary, but at the online Tlingit dictionary:

https://tlingitlanguage.com/dictionary/English.html

The onset is an alveolar stop, as in P-J, not a sibilant. There’s also a high-front vowel, as in P-J. The coda is a uvular fricative, although ejective rather than plain. Plain uvular ejective so far has been “k” in P-J, here it’s unclear. So let’s consult P-Athabaskan.

In P-A, “heart” is “dʒʷeˑy-əɁ”, also reconstructed as “jʳeˑ-y”.

https://en.wiktionary.org/wiki/Reconstruction:Proto-Athabaskan/-d%CA%92%CA%B7e%CB%91y-%C9%99%C9%81

Same vowel + glide sequence as in Tlingit, high-front, which parallels the “i” from P-J “ti”. The onset is a palatal rather than alveolar affricate, but it’s not a sibilant. It has either labialization or retroflex 2nd-ary feature. This is the plain version, not aspirated or ejective.

The aspirated version, written “c” for P-ND, corresponded to “s” in P-J, but that was also when the Tlingit cognate had “sh” (e.g., P-ND “caj” = “stone”), whereas here the Tlingit cognate has “t”. So perhaps the aspirated P-A consonant, at least when paired with “t” in Tlingit, corresponds to “t” in P-J.

The final syllable “əɁ” in P-A looks like it matches the “x̱’” in Tlingit — possibly a fossilized suffix from way back in the mouth, which is not in the reconstruction “jʳeˑ-y”. And so, not reflected in the P-J form either.

What about P-A “blood”? It’s “dəɬ”, more similar to the P-Y word in having a lateral coda, but unlike P-Y in having an alveolar stop instead of the sibilant from P-Y. The descendants all have high-front vowels, BTW.

https://en.wiktionary.org/wiki/Reconstruction:Proto-Athabaskan/d%C9%99%C9%AC

In Tlingit, “blood” is “shé”, which resembles P-Y in the onset being sibilant, but unlike P-Y or P-A in lacking a coda. This form is unlike P-J, with “t” onset, but like it in lacking a lateral.

https://en.wiktionary.org/wiki/sh%C3%A9#Tlingit

Overall, the Dene-Yeniesian words for “blood” are not the ones cognate with P-J “ti” = “blood”. Rather it’s the word for “heart” in Dene-Yeniseian that phonetically matches P-J “ti”. The semantic connection is clear — the heart is the blood organ.

However, as we saw earlier, the P-J word for “heart” = “kokoro” used to refer to the organ responsible for the collective senses involved in perception, emotion, etc. — i.e., the mind or brain. It was in turn derived from the Pre-PJ / Wa word for “ear”, cognate with P-ND, and related to the sense of hearing, then reduplicated to give it a collective meaning — all the senses.

By the P-J stage, “kokoro” shifted to refer to the heart rather than the mind, probably under the influence of a theory of bodily humors, including blood as one of them.

So, “kokoro” displaced some earlier word for “heart” — and that must have been the word preserved in P-J as “ti” and cognate with “heart” in Na-Dene, but now shifting slightly to refer to “blood” rather than the organ that pumps blood.

Continuing the series here, while my akinokure blog at blogspot is in limbo, we’ll keep Nonstop Nihongo New Year going. This time, the P-J word for “horn” = “tuno”. Across languages, words for “horn” and “bone” can be cognate, and that’s true within the Dene-Yeniseian families themselves. The Japonic cognate means “horn”, though.

https://en.wiktionary.org/wiki/Reconstruction:Proto-Japonic/tuno

In P-A, “bone” is “tsʼən”, in Tlingit it’s “sʼaaḵ”, and in P-Y “qawg”.

https://en.wiktionary.org/wiki/Reconstruction:Proto-Athabaskan/ts%CA%BC%C9%99n

These harmonize back to P-ND “cawnɢ”. However, I would say the onset should be ejective “c'” (usually written “ts'”), since both the Tlingit and P-A onsets are ejective, which according to the N-D obstruent correspondence table, means the P-ND form is ejective too, not plain.

That’s crucial for the Japonic connection, since we’ve seen the affricates become “s” in P-J — but those were plain, not ejective. Perhaps the ejective manner, which involves holding the airstream at the glottis and then releasing in a burst, makes an affricate sound more like the stop member than the more continuant fricative member. At least to Japonic ears, it did.

So instead of “ts” or “tl” becoming “s”, here “ts'” becomes “t”.

As usual, coda “w” -> “u”, and replaces the previous vowel (this time with no resistance from “a”). There’s still a coda cluster, and uvular “G” gets deleted in P-J anyway, so that member of the cluster is removed to simplify it, leaving “tun”.

This gets a dummy vowel, which is “o1” — a fairly rare vowel that almost only appears at the ends of words, probably indicating that it’s been added just to make the shape CVCV, and not part of the root meaning. It doesn’t seem to be a fusion vowel, like “e” resulting from the sequence “ai”. Perhaps the rarity is telling us something about what the preceding consonant used to be, like maybe it was actually “ŋ” at some point during Wa (blending “n” and “G” into a single coda), before becoming “n” in P-J. IDK, just worth noting.

So that gives P-J “tuno”, cognate with Dene and Yeniseian words for either “bone” or “horn”.

BTW, final remark on “yo” = “night”. In the P-Y word “qawGa” = “star”, the root “qaw” is probably more like “night sky” than just “night”, which already has a distinct word in P-Y (“çaj”). And it doesn’t mean “dark(ness)”, since that’s also derived from the word for “night”. What else can stars relate to, having to do with “night”? It must be a dedicated atomic word for “night sky”, contrasting with the ordinary “wes” for daytime sky.

In Japonic as well, “yo” must have originally meant “night sky”, but by P-J, it got extended to mean “night”, displacing the older Wa word that appears as “tu” in “tukui”, and cognate with all those “t” words for “night” in Na-Dene.

However, the newly coined compound “yomi” = “night sun” did not replace the older compound meaning the same thing, “tukui”.

Again, rare / obsolete words can survive in compounds — here, the “t” word for “night” cognate with Na-Dene, inside a compound meaning “night sun”, but lexicalized into “moon”.

This is the backup blog for Face to Face / Dusk in Autumn, akinokure.blogspot.com, in case Blogger locks me out or shuts down their platform. So here we are, still the cliff-dwelling sage in the ruins of the blogosphere, just publishing from a different site. Same sage, same Cliffs of Wisdom — just taking a different path up the mountain, the WordPress path rather than the Blogspot path. Same topics, same tone, same everything. Currently, going full steam ahead with Nonstop Nihongo New Year. Welcome, and thanks for coming back to make the journey up the cliffside… ^_^