The origin of "thing" in Chinese
« previous post | next post »
I recall that when I began learning Mandarin, one of the things (!) that troubled me greatly was why the word for "thing" was written with the characters for "east" and "west": dōngxi 東西. My classmates came up with all sorts of outlandish, speculative explanations for the supposed etymology. All along, I suspected that the meaning "thing" for the disyllabic word dōngxi 東西 was not derived from the characters used to write it but was the phonetic reflection of a borrowing or the representation of some colloquial, topolectal term.
From Mok Ling:
A friend of mine, Lucy, is in a Mandarin learning group. She told me about the bizarre etymology she was taught for the word dōngxi 東西. Apparently, 東西 being used to mean "thing, item" is based on the conception of the Five Phases (wǔxíng 五行 [VHM: formerly translated as Five Agents or Five Elements, which brings out the correspondences with the Four Elements of Western classical thought, also in the metaphysics of Indian, Tibetan, and other cultures]): East is represented by the element of Wood (木) and West is represented by the element of Metal (金). Objects are made of metal and wood, therefore "east-west" became a shorthand "thing" — obviously pretty ridiculous.
I had no real competing theories, so I couldn't make a convincing counter-explanation. Looking it up on Google I found some other… "interesting" explanations on sites like Zhihu, such as "merchants selling things come from the east to the west". Everybody's so creative.
My next instinct was to look at parallel texts, so I cracked open Takekoshi's Contrast Text version of the 老乞大. 東西 seems to have been pretty commonly used as far back as the Yuan era, though some lines in the older editions notably have the expected wù 物 [VHM: see separate note on this term below] to mean "thing":
【舊本老乞大 (~14C)】早來喫了乾物事，有些乾渴。(48/17b10-18a1)
【重刊老乞大 (1795)】吃了些乾東西，有些渴。(48/19b1
(I know the 老乞大* isn't the most reliable yardstick for dating vocabulary, but I thought I'd at least start there)
[*VHM: This is actually one of my favorite Korean books. Here's a note about it and a companion volume from Wikipedia:
Bak Tongsa (Chinese: 朴通事; lit. 'Pak the interpreter') was a textbook of colloquial northern Chinese published by the Bureau of Interpreters in Korea in various editions between the 14th and 18th centuries. Like the contemporaneous Nogeoldae ('Old Cathayan'), it is an important source on both Late Middle Korean and the history of Mandarin Chinese. Whereas the Nogeoldae consists of dialogues and focusses on travelling merchants, Bak Tongsa is a narrative text covering society and culture.]
I also found Zhengzhang Shangfang's paper titled 《Dōngxī tàn yuán sān tí 东西探源三题》 ("Three notes on the origin of 'dōngxi 東西'"). As I'm writing this I'm having a friend try and help me access the article on Wanfang Data because I don't have a Chinese phone number. Do you happen to know of any articles / papers written in English about this? Google's been completely overrun with SEO slop and other search engines aren't particularly useful for finding academic works.
VHM: A note on the pre-dōngxi 東西 word for thing, wù 物 ("thing; object; matter; substance"):
Compare Proto-Sino-Tibetan *(m/b)rəw (“grain; seed; lineage”) (STEDT), whence Tibetan འབྲུ ('bru, “seed; grain”), Jingpho amyu (ə myú, “kind; sort”), Burmese မျိုး (myui:, “seed”). Within Sinitic, compare 苗 (OC *mrew, “seedling, sprout”), 麰 (OC *mu, “barley”).
This proposed etymology of wù 物 does not match well the apparent structural origin of the glyph used to write it on the oracle bones, which is supposedly composed of 牛 ("bovine") as its semantaphore and 勿 ("do not; don't") as its phonophore. (Wiktionary) The latter in turn allegedly depicts drops of blood on the blade of a knife; the original character of 刎 (OC *mɯnʔ, “to cut”) (phonetically borrowed for the negative particle since the time of the oracle bone script). (Wiktionary)
To wrap up this post on "thingness", in the early stages of my Mandarin education, I was repeatedly instructed that one of the most withering insults one could utter was "Nǐ bùshì gè dōngxī 你不是個東西 ("You are not a thing"), the idea being that "you [the person being insulted] are not even as worthy as a mere thing / object")! Tant pis!
Selected readings
- "When did 'a thing' become a thing?" (4/18/16)
- "Not a gerund, not a thing" (3/5/10)
Chris Button said,
May 10, 2025 @ 6:46 pm
Any chance you could share a copy?
I wonder what it could be?
Incidentally, I'm glad to see that Zhengzhang Shangfang does not perpetuate the notion that 西 had some kind of Old Chinese nasal cluster sn- as its onset. The paleographic record shows that not to have been the case, and a simple Old Chinese s- will suffice. Having said that, Old Chinese is probably not relevant at the time depth for 東西, and I don't think anyone would dispute a simple s- onset in Middle Chinese in any case.
The reference seems rather to be to the coloring of the ox.
Victor Mair said,
May 11, 2025 @ 8:16 am
"Any chance you could share a copy?"
Mok Ling was telling ue about the article by Zhengzhang Shangfang that he found and was asking if anyone could help him obtain a copy.
Jonathan Smith said,
May 10, 2025 @ 9:11 pm
Mand. dōngxi 'thing' is clearly historically related to structures "VERB east VERB west" 'VERB this and that.' So at some point 'this and that' >> 'thing(s)'. Phrases of this type occur, among other places, in Amoy/Taiwanese from the earliest POJ literature down to the present (e.g. chò tang chò sai 'do this and that'; I see there is a wiktionary page). This could easily be a case in which southern languages influenced the standard Mandarin emerging around Nanjing during the Ming — lots of examples of such influence are unrecognized.
Chas Belov said,
May 10, 2025 @ 9:14 pm
東西 for "thing" has puzzled me ever since I saw it on a multilingual sign at the library not to throw things into the restroom toilet. I assumed it was a translation error, only to learn when I got home and looked it up that 東西 meaning "thing" was indeed a thing. Good to know even experts in Chinese (which I am pretty much the opposite of) are puzzled as well.
Elizabeth Okada said,
May 10, 2025 @ 9:38 pm
This is funny from a Japanese perspective because 東西 (touzai) is an expression used to get the audience's attention at a Kabuki theater etc. (Something like "ladies and gentlemen!")
Jonathan Smith said,
May 11, 2025 @ 12:25 am
Just coincidentally saw that in discussing early Guanhua syntax, Coblin ("A brief history of Mandarin" /2000:546) cites three example sentences from Francisco Varo's Arte de la Lengua Mandarina (1703) that happen all to contain a word 'thing' (context = Catholic Confession):
1. 你偷了人的物件麽 [Did you steal something of someone else's?]
2. 這個物件值多少銀子 [How much is the thing which you stole worth?]
3. 你告解後頭的東西共值多少銀子 [How much is the total value of the things you have stolen since your last confession?]
Coblin doesn't discuss 'thing' in this paper (I don't think), but note 1 and 2 use 物件 while 3 contrasts in using 東西 — it appears to retain a collective sense, "all the stuff/things this that and the third"
Yves Rehbein said,
May 11, 2025 @ 12:40 am
The mainstream theory states that people in ancient times (Some theories might claim that the vocab origins from the Han or Tang Dynasty) would go to East and West markets […] https://en.wiktionary.org/wiki/東西 w.f.r. in Zh.
This is compatible with "do this and that", @ Jonathan Smith.
I was rather thinking of Italian cosa, che cosa, which presumably underlies the mathematical notation of X since Descartes in lieu of Arabic šayʔ "something; little" (says Grimm's Wörterbuch), compare Persian čīz, cognate to Italian che, Portuguese que, Latin quid "what", though already in Egyptian (j)ḫt "thing", ḫt "wood". Note that šayʔ in Arabic also connotes "small", like xiǎo 小 MC sjewX, which I find difficult to separate from the noun forming suffix 子 MC tsiX, Jp. し shi す su つ tsu (Phoshime anyone?) ず zu こ ko ("child") み mi. This is important because when I say hot sh*t if you know what I mean I do not mean droppings – 氣 not 屎, n.b.: "The 米 component was originally three (, representing 小, as seen in ), four (, representing 少) or five dots (as seen in ) […]" https://en.wiktionary.org/wiki/屎 with embedded images
Unfortunately, the only way I see to explain dōng in this case is with Greek τον, eventually cognate to this and that. So why don't I shut up about it?
On a separate yet intriguing note, xiàng 象 ("elephant", "idea"), here specifically dàxiàng, has a vexing folk etymology that is reminicent of previous discussions about Griffins and bixie vel sim. (note Cyclopes as well). I believe as much as ivory was used to carve little thingies, frequently subject to folketymology (e.g. German Elfenbein from elephant "ivory" – now "elf" + "bone"), it has to go way back to the neolithic. Our esteemed host Victor has once made a point of a burial found north of China with hundreds of ivory beads! Whereas Semitic supports šin "tooth" and Sumerian supports zi "horn", loanwords are virtually guaranteed because the elephant is regionally limited, the hippopotamus more so; Egyptian supports db.
In conclusion I suppose that 東西 was originally reduplicative and underwent drastic changes. You figure it out :-)
Chris Button said,
May 11, 2025 @ 8:54 am
動使 > 東西 according to this article: http://www.guoxue.com/?p=3764
Chris Button said,
May 11, 2025 @ 9:02 am
動使 > 東西 according to this article.
Melanesian priest said,
May 11, 2025 @ 9:15 am
is the etymology of "bird" related to Chinese and Vietnamese? What do you think: https://www.reddit.com/r/sinotibetan/comments/1kayi45/recent_crossfamily_borrowing_or_a_really_ancient/
Jerry Packard said,
May 11, 2025 @ 1:34 pm
I always thought it was “… this and that…” @Jonathan Smith. In any event it does not seem cognate to S Min mi gia.
Chris Button said,
May 11, 2025 @ 2:22 pm
@ Jerry Packard
Is 東西 used to mean "thing" in S.Min? If yes, I think the 東 … 西 … etymology might have legs. If not, then why would the concept be borrowed (exclusively?) for an entirely novel use elsewhere in Chinese?
Jonathan Smith said,
May 11, 2025 @ 3:14 pm
To be clear, Mand. dōngxi would not have derived from the above Taiwanese etc. formula per se. Rather, dōngxi is in origin a "dvanda"-type-compound — "[things] over here + [things] over there" — which shares a motivation with (among others) Tw. VERB tang VERB sai (also cf. Mand. dōng VERB xī VERB 'VERB this way and that). The best place to look for the phrasal applications most directly anticipating the modern word might be Wu / NEastern Min languages…
The above late 17th cent. Guanhua sentences cited by Coblin, where "東西" is collective and 'thing' is instead simply "物件" — to which cf. Tw. mi̍h-kiāⁿ noted by Jerry Packard — is really quite instructive… sorry I am too lazy to type out Varo's complicated phonetic renditions.
And re: mi̍h-kiāⁿ and many others, it is often hard to say where Nanjing Guanhua has influenced e.g. Wu/Min languages and where the effect is the reverse. Note e.g. ubiquitous Mandarin shénme 什麽: in the phonetically similar Taiwanese (etc.) word siáⁿ-mi̍h, the second syllable actually means 'thing', whereas "me" is obscure inside Mandarin… influence must be from southern languages > Guanhua.
Jonathan Smith said,
May 11, 2025 @ 3:35 pm
Decisively supporting the collective angle, in Tâi-Ji̍t Tōa Sû-tián 臺日大辭典 (1932) the second definition of tang-sai is "só͘-ū ê mi̍h-kiāⁿ" 'all the things / everything', usage example being tang-sai bû só͘ put glossed as 'siánn-mih mi̍h-kiāⁿ to beh the̍h' i.e. 'all the things are taken / included'.
Note the example is markedly acrolectal and then glossed basilectally, leaving open the question of exactly where the development 'this thing and that thing' > 'all the things' > 'thing' (first) took place. Sure, it could have been in early Mandarin itself. But this was pretty clearly the development.
Chris Button said,
May 11, 2025 @ 5:31 pm
@ Jonathan Smith
什麽 has been shown to be a survival in the colloquial strata of 什物. Compare the similar relationship between 的 and 之.
As for 東西 in S.Min, a dictionary entry should be treated with caution–not to mention an "acrolectal" (!) and "basilectal" (!) entry.
David Marjanović said,
May 12, 2025 @ 8:21 am
Oh. That practically clinches it IMHO.
*nw > mm like in southern German? I like that.
Jonathan Smith said,
May 12, 2025 @ 12:04 pm
Re: 什麽~什物 — no, the fact that Mand. 'what?' historically involves 'thing' is not "shown", it's simply "known" if you… read a Qing novel or something. My above is one possible more specific/concrete formulation of why (historical) m- and not expected (medieval northern development) w- in the modern Mand. item, as m- is retained in 'thing' only in some southern languages.
Re origins of 東西, IMO Early Guanhua sentences like Varo's clinch it if it needed clinching — i.e. clench it i.e. make it cling TIL. Native scholars tend not to ever read such materials or these days to ever learn non-Mandarin languages, so we get "動使" and the like from the philologists.
Chris Button said,
May 12, 2025 @ 5:03 pm
@ Jonathan Smith
Respectfully, I think you're overthinking it and then needlessly speculating as a result. It is simply a perfectly natural retention of a more archaic sound in the colloquial language.
@ David Marjanović
Actually, the point is that the earlier m- onset has persisted to the present.
I think it's a good idea that is quite possibly correct. But i don't see any solid evidence yet. To my earlier comment, I'm not seeing 東西 and … 東 … 西 clearly coexisting.
Jonathan Smith said,
May 12, 2025 @ 6:46 pm
Returning to "dōng VERB xī VERB" — this has been common for centuries, e.g. (18th c.) Hongloumeng and modern Mandarin dōng qiáo xī wàng 東瞧西望 'gaze this way & that', dōng lā xī chě 東拉西扯 'blather on about this & that', etc., etc. HOWEVER the notion that this might be (indirectly) related to Mand. dōngxi 東西 'thing' is (it would seem) not at all obvious.
Differently, "VERB east VERB west" places "east" and "west" (even if still "adverbial" in force) in "object" position such that it is clear how development towards 'this & that, various things' could occur. And it turns out that this template, very common in e.g. modern Taiwanese, is not totally absent in stuff like Hongloumeng: e.g. I find shuō dōng tán xī 說東談西 'chat about this & that'.
Combining this with sentences from e.g. Varo's Guanhua grammar — just found qiú dōng xī 求東西 'ask for this and that' (NOT = 'ask for a single literal thing' as this text uses a different item for 'thing') renders this account a complete posterization.
Question of shenme is OTOH not a posterization, but no @Chris Button handwavy "colloquial retention" is woefully insufficient; read some of the texts above for why.
Chris Button said,
May 12, 2025 @ 7:57 pm
@ Jonathan Smith
It's not called "handwavy", it's called "real world". What I am describing is a common phenomenon among very commonly used words in languages around the world.
I noted the genitive 的 ~ 之 above
You have not addressed that example. So let me try another. Have you never wondered why the negative 不 has an irregular b- [p-] onset going back to its earlier form instead of the expected f- ?
If it helps you at all, Pulleyblank (1999) "Morphology of demonstrative pronouns" has a nice little bit on the etymology of 的 and 不. And Zhang (1982) “Shi ‘shenme’” is apparently the classic article on 什麽, although I haven't read that one myself so can't speak to the details.
Jonathan Smith said,
May 12, 2025 @ 9:39 pm
^ above is 東拉西扯 'blather about unrelated stuff'. No one said such retentions aren't a thing. Long story short this won't work in a straightforward manner for shenme — e.g., one must first explain the first syllable. /thread
Chris Button said,
May 12, 2025 @ 10:32 pm
Think about how the word actually surfaces in its various guises.
The first syllable represents assimilation of -p to -m, which then bumps up with a phonotactic constraint calling for -n even if the -n doesn't necessarily surface as such.
From a purely orthographic perspective with substitutions like 甚, the latter stage is a little like Burmese orthographic mramma > mranma for "Myanmar".
David Marjanović said,
May 18, 2025 @ 6:32 am
Do you mean it's a borrowing from another dialect? Sound changes don't just randomly fail to apply in individual words.
However, the preceding nasal blocking [m] > [w] makes perfect sense.
Chris Button said,
May 18, 2025 @ 1:43 pm
Resistance to regular sound change sometimes happens in very high frequency words.
Also, it was a preceding Middle Chinese -p coda in the first syllable of the case at hand.
Chris Button said,
May 18, 2025 @ 2:49 pm
And the regular evolution of Middle Chinese m- to Mandarin w- in 物 would be via ʋ- on account of the following u–a change that a preceding -p coda on the syllable before would not have prevented.
What actually seems to have happened is that the reduction of the -u- to schwa in the second syllable (we're dealing with a high frequency word after all) prevented the change. I noted the negative 不 above as another high frequency word that did not evolve regularly. The situation seems to be similar with a reduced Early Middle Chinese form of pət versus put, which results in a failure to produce expected Mandarin f-.
Chris Button said,
May 18, 2025 @ 2:53 pm
So to your point, yes there is regularity in the irregularity :)
Chris Button said,
May 18, 2025 @ 3:50 pm
And just to close the loop here…
As for "zhī" 之 EMC ʨɨ versus "de" 的, the crux of Pulleyblank's discussion (1999) is that the genitive marker failed to palatalize "when unaccented as an enclitic particle in colloquial speech."
So, here once again, the irregular reflex is not about sound changes randomly failing to apply. It's just that the (superficial) irregularity of high-frequency words needs to be looked at more closely from the perspective of actual speech. Otherwise, from a strict "regularity" approach:
的 should be pronounced zhī
不 should be pronounced fú
什麽(物) should be pronounced shíwù
Chris Button said,
May 18, 2025 @ 5:47 pm
Incidentally, I was curious why 什物 could become a question word. Jerry Norman makes a nice comparison in his 1988 book of Italian "cosa" (thing) being used to mean "cosa?" (what?).
Words like "thing", "item, "goods", "wares" have various etymologies. Taking 東西 literally is likely a mistake in my opinion. The solution likely lies in something sounding similar, which was the point made at the top of the original post here.
KIRINPUTRA said,
May 24, 2025 @ 12:12 am
@ Jonathan Smith
Vanishingly unlikely that the Mandarin would have been informed by the Hoklo. Keep in mind that SIÁᴺ-MIH (back syllable is "linguistically" Tone Class 4) is a pretty recent formation. SĪ-MIH is the "canonical" form, and still used in Teochew, apparently. SÍM-MIH is SĪ-MIH evolved, via *SĪM-MIH. SIÁᴺ-MIH doesn't seem to follow from SÍM-MIH; could it be from *SĪ HÁᴺ-MIH?