Austronesian etymology and loanwords

[hansioux]

I got one more. The Sinitic word 竹. 竹 is tik in Taigi, trjuwk in Middle Chinese, *truk in Old Chinese. It is a loanword from Kra-Dai’s *b.twuːkᴰ, originally meaning a strip of bamboo, cognate with ตอก (dtɔ̀ɔk, “bamboo strip”) in Thai. In Proto-Kam-Sui the word is *thruk7 ‘bamboo strip’.

Since Sagart claimed Kra-Dai is a daughter language of Proto-Austronesian, and a sister language of Malayo-Polynesian, it would ultimately mean the Chinese word bamboo came from Taiwan.

So I went looking for what that word is supposed to be in Austronesian languages.

In Paiwan, the word for a strip of split and flattened bamboo is taqtaq, and adding the m would make t-m-aqtaq into a verb meaning splitting bamboo.

It ultimately came from the root -tak/-taqak, meaning the sound of splitting or cracking something. Related words are *betak meaning to split or crack, which is cognate with biták in Tagalog.

It would make sense that this word is a loan into Old Chinese, at a time when they still created new characters for new things. There would have no bamboo around the bend of the yellow river, and bamboo wear that made its way to that region probably were weaved with bamboo strips.

	OC	PAN	Proto-Kra-Dai	Proto-Kam-Sui	Thai	Comments
bamboo strip	*truk	*taqtaq	*b.twuːkᴰ	*thruk7	dtɔ̀ɔk	bamboo in Sinitic

[hansioux]

Eh… just realized the moon goddess 嫦娥 (*dangŋˤajʔ) sounds extremely similar to the Pangcah Amis moon goddess Dongi.

[hansioux]

Something I just thought of blew my mind…

The famous sword making warring state country Yue (越), whose swords used to be mere legends, but with the unearthing of the Sword of Goujian, which remained sharp after at least 2500 years have been proven to be historical facts. The locations of Yue (越) and culturally similar Wu (吳) used to be Kra-Dai speaking. So I was trying to see if there are any hints in the names of the states or kings.

Turns out, the tribal name (氏) of the kings of Yue is 諸稽, which in Old Chinese would have be *ta-kˤij.

I think it came from the Austronesian word *takis, which meant… sword.

[arcticsquid]

So 7 in some Formosan languages and rest of Austronesian languages should be 5+2 then?

[hansioux]

@arcticsquid

So 7 in some Formosan languages and rest of Austronesian languages should be 5+2 then?

Yes, pitu is abbreviated from RaCep+pi+tuSa, 5 and 2.

It is obscured because RaCep was replaced by lima.

[slawa]

I came across this twitter post and it seems there could be a great leap with AI towards rare complicated languages translation and interpretation.

https://twitter.com/hahahahohohe/status/1765088860592394250

[hansioux]

@slawa

I came across this twitter post and it seems there could be a great leap with AI towards rare complicated languages translation and interpretation.

I threw my short possible OC-PIE cognate list from this thread at Claude, with only the Hanji, OC, PIE, and English columns. I didn’t provide the column containing my comments, or the OE, ME columns.

I asked it for more possible cognates, and it hallucinated. However, when I asked it to provide possible cognates for the characters with an empty English cell… well, it got interesting.

For the empty cells in the English column, here are some potential English cognate words based on the Proto-Indo-European (PIE) forms provided:

先 *sˤər
The PIE root is not given, but this could be related to the PIE root *ser- ‘to line up, arrange’. A possible English cognate would be “series”.

昭 *taw
照 *taw-s

朝 *taw
These Old Chinese forms correspond to the PIE root *dyeu ‘to shine’. Possible English cognates derived from this PIE root include “day”, “dawn”, and “jovial” (from the name of the Roman god Jupiter < *dyeu-pater ‘sky father’).

So potential English cognate entries could be:

先 *sˤər | *ser- | series

昭 *taw | *dyeu | day, dawn
照 *taw-s | *dyeu | day, dawn

朝 *taw | *dyeu | day, dawn

Note that these are speculative suggestions based on the limited data provided. Establishing robust etymological links requires much more comprehensive evidence and analysis by historical linguists.

Of course, it could be just referencing the English etymology training data that it saw before and not really doing anything with the OC data I provided.

So I asked about the cells which I didn’t provide any possible PIE cognates, mostly because I couldn’t think of one at the time of compiling that list.

Here are the possible English cognates for 路, 客, and 旅 based on the information provided:

路 (*Cə.rˤak-s in Old Chinese):
There is no Proto-Indo-European reconstruction given, so an English cognate is not clear from this data. The Old Chinese form may be related to words meaning “road” or “path”.

客 (*kʰˁrak in Old Chinese):

Again, no Proto-Indo-European form is listed, but the Old Chinese reminds of the English “caravan” from the Proto-Indo-European root *ker- “to go”.

旅 (*[r]aʔ in Old Chinese):
No Proto-Indo-European form is provided for this Old Chinese reconstruction either. However, it may be related to the root *lē- “to go, travel”, which gave rise to English words like “lore” (knowledge from traveling) and “less” (one who travels slower).

So in summary, while direct English cognates are not obvious from the data provided, the Old Chinese forms suggest possible connections to words related to traveling, roads, and going on journeys in English, but firmer conclusions cannot be made without the corresponding Proto-Indo-European reconstructions.

Ok, now I’m scared. Those actually make sense, except for the one for 客, that PIE -ker for caravan is a hallucination.

Obviously I didn’t pay for Claude OPUS, and just used the free Claude 3. So maybe the professional version could do even more.

[slawa]

@hansioux

I asked it for more possible cognates, and it hallucinated.

Here a reply from Opus

@Claude Opus

Here are some additional possible Old Chinese and Indo-European cognates to consider:

Hanji	OC	MC	PIE	English	Comment
門	*mˤə[n]	mun	*manu-	manu	Meaning human, mankind
母	*mˤaʔ	muwX	*máh₂tēr	mother
父	*[b]aʔ	bjuwX	*ph₂tḗr	father
是	*[d]e	dzyeX	*h₁es-	is	3rd person singular of “to be”
吾	*ŋˤa	ngu	*éǵh₂	I, me	1st person pronoun
斧	*p(r)aʔ	pjuX	*péleḱu-	axe
牛	*[ŋ][w]ə	ngjuw	*gʷṓws	cow
狗	*Cə.[k]ˤro	kæwX	*ḱwṓ	hound
蜂	*s.[b]roŋ	phjowng	*bhi-	bee
冬	*tˤoŋ	towng	*ǵheim-	winter
日	*C.nik	nyit	*di-	day
死	*sijʔ	sijX	*stAi-	stay	OC meaning “to die”

Some of these, like the words for mother, father, and the pronouns, are very basic vocabulary items that could suggest a deeper relationship. Others like 牛 (*[ŋ][w]ə) for cow and 門 (*mˤə[n]) for human/mankind are quite striking in their phonetic and semantic similarity to the PIE roots.

However, it’s important to be cautious about drawing conclusions from a list of potential cognates. Chance resemblances do occur between unrelated languages. To argue convincingly for a genetic relationship or creole origin scenario, linguists would want to see a large number of reliable cognates that follow systematic sound correspondences. They would also look for similarities in morphology and grammar, not just vocabulary.

Additionally, proposing an Indo-European superstrate would raise questions about the historical and archeological evidence for such extensive contact and influence. The mainstream view is that Old Chinese developed from earlier Sino-Tibetan languages without major outside influence.

So while the list of possible cognates is intriguing, much more evidence would be needed to overturn the standard theory of Old Chinese origins. The similarities could potentially reflect very old areal roots or chance resemblances rather than direct linguistic ancestry. More research comparing Old Chinese to Sino-Tibetan and Indo-European languages in a systematic way would help shed light on the issue. But overall, an Indo-European superstratum seems unlikely based on our current state of knowledge.

Here are the remaining rows from the original table, with comments on potential connections or lack thereof to Proto-Indo-European roots:

Hanji	OC	MC	Taigi	PIE	OE	English	Comment
腹	*p(r)uk	pjuwk	pak				No clear PIE cognate
夾	*kˤ<r>ep	keap	kiap				No clear PIE cognate
窟	*[kʰ]ˤut	khwot	khut				No clear PIE cognate
歿	*mˤut	mwot	bu̍t				Possible connection to PIE *mer- “to die” but phonology is not a clear match
沸	*Nə.p[u][t]-s	pj+jH	hut				Possible connection to PIE *bhreu- “to boil” but phonology is not a clear match
先	*sˤər	sen	sing				Possible connection to PIE *sen- “old” but semantics and phonology are uncertain
輻	*pәk	pjuwk	hok				No clear PIE cognate

For these words, either no plausible Proto-Indo-European cognate is known, or the phonological and semantic match to proposed cognates is uncertain.

In historical linguistics, some proposed cognates will inevitably be less secure than others. Chance similarities can occur, especially with simple syllables. Sound changes over time can also obscure ancestral relationships.

To demonstrate relatedness, linguists look for a sizeable number of reliable cognate sets that show systematic sound correspondences. Isolated resemblances are not as convincing, as they may be coincidental or result from separate onomatopoeia or sound symbolism processes rather than shared inheritance.

The words above lacking strong Indo-European cognates do not negate the other potential cognates on the list. However, they underscore the need for caution in proposing new long-distance linguistic relationships. Any hypothesis of Old Chinese being a creole with Indo-European influence would need to be based on a preponderance of evidence, with the understanding that not every lexical item will have a clear cognate. Uncertain cases like these are to be expected even in languages known to be related.

Here is the complete list of potential Old Chinese and Indo-European cognates using a markdown table:

Hanji	OC	MC	Taigi	PIE	OE	English	Comment
腹	*p(r)uk	pjuwk	pak			belly, bucket
夾	*kˤ<r>ep	keap	kiap	*ghrebh	græpsan	grasp
窟	*[kʰ]ˤut	khwot	khut	*kutą	cot	cot
歿	*mˤut	mwot	bu̍t	*mer	mort	mort
沸	*Nə.p[u][t]-s	pj+jH	hut	*bhreu-	broþ	broth	liquid that’s boiled
雁	*C.[ŋ]ˤrar-s	ngaenH	gān	*ghans	gos	goose
長	*Cə-[N]-traŋ	drjang	tn̂g	*dlonghos	lang	long
屯	*[d]ˤun	dwon	tun	*dhu-no-	tun	town
株	*tro	trju	tsu	*drew-o-	treo	tree
球	*[g](r)[u]	gjuw	kiû	*gleu-	cliewen	clew
號	*[C.g]ˤaw	haw	hō	*gal-		call	Old Norse: kalla. Also 吼、哮、嘷
何	*[g]ˤaj	ha	hô	*kwo-	hwæt	what	also why
磨	*mˤaj	ma	buâ	*mele-	mylen	mill
飛	*Cə.pə[r]	pj+j	pue	*pleu-	fleoge	fly
浮	*m.b(r)u	bjuw	phû	*pleu-	flowan	flow
先	*sˤər	sen	sing	*sen-			Meaning old. Root for senator, senior.
孫	*[s]ˤu[n]	swon	sun	*su(e)-nu-	son	son
帝	*tˤek-s	tejH	tè	*deiwos	tig	deity	OE name of a god. *deiwos also evolved to be Zues, Jupiter
馬	*mˤraʔ	maeX	bé	*márkos	meare	mare	Old Norse: merr
車	*C.q(r)a	kjo	tshia	*krsos	carre	car	Anglo-French
輻	*pәk	pjuwk	hok	*spei-	spaca	spoke	PIE meaning sharp point, i.e. spike
羆	*praj	pje	phî	*bher-	bera	bear	OC meaning brown bear, PIE meaning brown. Bears seems to be sacred to IE speakers, and Huangdi is supernaturally related to bears. One of his tribal names is 有熊
牘	*lˤok	duwk	to̍k	*dek-		doc	Also 劄. As in document, doctrine, doctor. OC meaning instruction, official announcement. PIE meaning instruction, teach.
門	*mˤə[n]	mun		*manu-	manu		Meaning human, mankind
母	*mˤaʔ	muwX		*máh₂tēr	mother
父	*[b]aʔ	bjuwX		*ph₂tḗr	father
是	*[d]e	dzyeX		*h₁es-	is		3rd person singular of “to be”
吾	*ŋˤa	ngu		*éǵh₂	I, me		1st person pronoun
斧	*p(r)aʔ	pjuX		*péleḱu-	axe
牛	*[ŋ][w]ə	ngjuw		*gʷṓws	cow
狗	*Cə.[k]ˤro	kæwX		*ḱwṓ	hound
蜂	*s.[b]roŋ	phjowng		*bhi-	bee
冬	*tˤoŋ	towng		*ǵheim-	winter
日	*C.nik	nyit		*di-	day
死	*sijʔ	sijX		*stAi-	stay		OC meaning “to die”
焚	*[b]u[n]	bjun	hûn		brennen	burn	Proto-Germanic *brennanan
割	*Cə-kˤat	kat	kat			cut	North Germanic *kut-, replacing OE ceorfan (carve)
曲	*kʰ(r)ok	khjowk	khiok	*keg-	hoc	hook	PIE meaning hook, tooth
壯	*[ts]aŋ-s	*tsrjangH*	tsòng	*strenk-	strang	strong	PIE meaning tight, narrow
背	*pˤək-s	*pwojH*	puè		bæc	back	Proto-Germanic *bakam
給	*[k](r)[ə]p	kip	kip	*ghabh-	giefan	give	PIE meaning to give or receive
葉	*l[a]p	yep	ia̍p	*leub(h)-	leaf	leaf	PIE meaning peel off
頁	*l[a]p	yep	ia̍h	*leub(h)-	leaf	leaf	leaf like thin material
約	*[q](r)ewk	ʔjak	iok	*yeug-	geoc	yoke	Derived from to join, to connect, to attach, to constrain
岬	*[k]ˤr[a]p	kˠap̚	kap	*kaput-		cape	English word from Latin caput, French cap. PIE meaning head
蓋	*[k]ˤap-s	kajH	khap	*kaput-	cæppe	cap	PIE meaning head

I can’t really comment on those replies. Hope those are helpful @hansioux

[hansioux]

Most of them are hallucinations, like sure 日’s OC is *C.nik, but day ultimately came from the PIE root *dhegh for to burn or heat. Even if OPUS got the etymology for day right, these two don’t look they are related. The same for 吾, they just don’t look related.

Although, Claude OPUS definitely has read books in PIE and OC before. It can recall most of these reconstructions, even if it doesn’t consistently stay within one system.

[hansioux]

Just learned why tomatoes are called kam-á-bi̍t (柑仔蜜) in Taigi. Apparently it was originally called kam-á-ti̍t (柑仔得), and the etymology is as such:

tomātl (Nahuatl) → tomate (Spanish) → tamatis (Cebuano) →kamatis (Tagalog) → kam-á-ti̍t (original loan in Taigi) → kam-á-bi̍t (current popular Taigi)

So not exactly an Austronesian loan into Taigi, but not exactly not either.

[hansioux]

@hansioux

‘twice-three’ with repeating 3 twice, *Nem-Nem > *emnem (Enemish)

So I actually tried to look for evidence of this when I first posted this. I think I’ve found evidence of people using *Nem as 3 in the languages not in the Limaish family, such as Pazeh, and in some Sirayaic languages, which is supposedly in the Enemish family.

Taivoan and Makatao, which are Sirayaic languages, *Nem went through some sound changes, *Nem > nom > lom > rom. So we get ra-ruma for 3 and lumta for 4 (lum 3 + ta 1) in Makatao.

There is also records of pinum for 3 in Basay. Although plenty of them also use the *telu form of 3.

So in terms of numeral innovations, replacing 6 with *NemNem was probably the first one to happen, followed up with replacing 3 *Nem with *telu, followed up by Pituish and Walu-Siwaish, since we still need *RaCep for 5 to form those abbreviations, then we get Limaish and Puluqish.

[hansioux]

The Evolution of Maori in 22 Words

Great to see a video laying the evolution out.

[hansioux]

The Austronesian Discovery

Great video about the history of Austronesian linguistic study.

[hansioux]

@hansioux

Kaxabu’s 7 is xasepbidusa, which retains *RaCepituSa perfectly.

This Kaxabu audio dictionary by Phuann Kim-gio̍k and published online by iTuan is amazing.

Kaxabu numbers:

PAZIH LANGUAGE, PEOPLE, & CULTURE

Video on Pazih or Pazeh from ILoveLanguages!, which is a youtube channel ran by a Filipino linguist.

The number system is almost identical to Kaxabu.

[hansioux]

My thought on the word Māori’s connection with Taiwan’s Indigenous languages.

Word	Language	Meaning
*qulid	PAN	life, alive
'orip	Pangcah	life
q-m-uzi-quzip	Paiwan	living beings (animals)
hidup	Malay	life
urip	Javanese	life
*ma-qudip	PMP	living, soul, spirit
m-urip	Kayan	alive, to sprout, reside at
*maqoli	POC	true, genuine
m-ori	Bimanese	living, alive
mauli	Suau	life
mo-ʔui	Tongan	to live, be living or alive
ma-uli	Samoan	seat of the emotions
Ma-uli	Hawaiian	life, heart, seat of life, spirit
Ma-uri	Maori	normal

[Dravanvaru]

What about Cook Islands Māori?

[hansioux]

@hansioux

I was looking up native words for the Formosan leopard cat (also called mountain cats or 石虎). The Saisiyat calls them balbalan. What’s interesting is the Truku word for these rare felines is Kramu.

I think it’s a cognate for the Malay word for Malaysian tigers, harimau.

Kramu → harimau
K > ha’ ra → ri mu → mau

@hansioux

The qali- initial is reserved for things that are holy, or connected to the spiritual world. The examples are rainbow *qali-waRwaR, butterfly *qali-baŋbaŋ, and whirlwind *qali-puspus. This “prefix for words with a sensitive reference to the spirit world” often evolves into kali- or ali-.

So I just saw somewhere that claimed harimau came from *qari-maquŋ, which came from wiktionary. Not sure if the *qari- prefix is the same as *qali-, also don’t know hwere the maquŋ came from.

The word for Formosan clouded leopard is *lukeNaw and remains so in most Formosan language.

[hansioux]

@hansioux

Sagart claimed Kra-Dai is a daughter language of Proto-Austronesian, and a sister language of Malayo-Polynesian, it would ultimately mean the Chinese word bamboo came from Taiwan.

Saw this chart of how “eat chicken” evolved from PAN to modern Kra-Dai and Austronesian languages.

[hansioux]

More Kra-Dai and Austronesian cognates:

Austronesian Languages (Cognation & Comparing) | Some Proto-Kra (a branch...

Gloss	Proto-Kra-Dai	Proto-Kra	Proto-Austronesian	Comment
1	*(C)itsɤː	*tʂəm	*isa
2	*saː	*sa	*duSa
3		*tu	*telu
4		*pə	*Sepat
5		*r-ma	*lima
6		*x-nəm	*enem
7		*t-ru	*pitu
8		m-ru	*walu
9		*s-wa	*Siwa
10		*pwlot	*sa-puluq
back		*lon	*biliN	walk behind
boat	*Cuɖá		*aluja	oar
bear	*kəˀmˠúj	*C-me	*Cumay
before		*kun	*(q)uNah
behind		*lon	*biliN	walk behind
bird		*ɳok	*manuk
black		*hl/dəm	*tidem
blood	*pɤlaːc	*plat	*daRaq
blood		*kɣa	*daRaq
boil		*plaɯ	*baʀəq
bone		*dək	*CuqelaN
borrow	*C-ɟáːm		*Sezam
chest		*tək	*daSdaS
child		*lak	*aNak
crab		*d-rat	*gaRaŋ
cut		*te	*tebteb
day		*(h)wən	*waRi
die	*maTaːj		*ma-aCay
dry in sun		*m-ʈak	*betak	crak, split open (as dry ground).
ear		*k-ra	*Caliŋa
eat		*kan	*kaen
excerment		*kai	*Caqi
eye	*maTaː	*m-ʈa	maCa
fall		*tok	*taktak
fart		*C-tot	*qetut
fire	*(C)apuj	*pui	*sapuy
flea		*x-mət	*qatimela
flower		*hŋa	*buŋa
front		*kun	*(q)uNah
fruit		*C-mak	*buaq
grandmother		*ja	*aya
hair		*mot	*gumuN	body hair
hair		*mot	*kumiS	public hair
hair		*mot	*bukəS	head hair
hand	*(C)imɤː	*mja	*(qa)lima
head		*krai	*qulu
head louse	*KuTuː	*C-ʈu	*kuCux
horn		*C-ku	*sequŋ
I (fst p singular)	*akuː	*ku	*aku
intestines		*C-si	*Cinaqi
leg		*C-ka	*qaqay
month		*x-ŋo	*qaNeb
moon		*m-ɖjan	*bulaN
mouth		*ŋuŋ	*ŋusu
navel		*m-ɖaɯ	*puja
navel		*m-ɖaɯ	*pujek
nose		*teŋ	*mujiŋ
pig		*x-mu	*babuy
pluck		*bet	*putput
rain		*jəl	*quzaN
rat		*hlai	*labaw
raw		*(k-)ɖep	*qudip
sand		*hŋai	*qenay
sand		*hŋai	*bunaj
see		*kai	*kita
see		*ti	*kita
sesame		*l-ŋa	*leŋa
shellfish		*tʃui	*ku(S)ul
shoulder		*m-ba	*qabaRa
skin		*boŋ	*baŋeS
snake		*ŋa	*Sulaʀ
sun		*(l-)wən	*waRi
taro	*pəˀrˠáːk	*p-ɣak	*biRaq
that		*ʔ-ɲa	*-na
this		*ʔ-ni	*-ni
tongue	*(C)əmaː	*I-ma	*Sema
tooth	*lipan	*I-pən	*nipen
water	*(C)aNam		*daNum
weep		*ɲit	*Caŋis
who		*ʔ-nau	*na-nu	what
you (snd p singular)	*isuː; amɤː	*mə	*k-amu

[Dravanvaru]

@hansioux

@hansioux

I was looking up native words for the Formosan leopard cat (also called mountain cats or 石虎). The Saisiyat calls them balbalan. What’s interesting is the Truku word for these rare felines is Kramu.

I think it’s a cognate for the Malay word for Malaysian tigers, harimau.

Kramu → harimau
K > ha’ ra → ri mu → mau

@hansioux

The qali- initial is reserved for things that are holy, or connected to the spiritual world. The examples are rainbow *qali-waRwaR, butterfly *qali-baŋbaŋ, and whirlwind *qali-puspus. This “prefix for words with a sensitive reference to the spirit world” often evolves into kali- or ali-.

So I just saw somewhere that claimed harimau came from *qari-maquŋ, which came from wiktionary. Not sure if the *qari- prefix is the same as *qali-, also don’t know hwere the maquŋ came from.

The word for Formosan clouded leopard is *lukeNaw and remains so in most Formosan language.

So I followed your lead and looked up harimau on wiktionary. Then followed the link to *qari-maquŋ. The data source was listed so I followed that link and found a very nice resource. ACD combined with openstreetmap to show cognate sets on a map. You probably already know of it, but I thought I’d share it here because it really tied everything together for me visually.

ACD - Austronesian Comparative Dictionary Online - Cognateset...

Here’s a screenshot:

The map for *pitu₁ (seven, or 5+2) is really cool (below is just a screenshot, hover over the dots on the actual map and it shows the language/dialect and the cognate word):