European dictionaries such as the Webster's English Dictionary or the Oxford Dictionary often have detailed etymologies of each word entry, together with dates of oldest attested usage and history of each affix; either this, or the strongest conjectures thereof. Why is it that Chinese dictionaries have nowhere near this amount of etymological information? At most, scholarly dictionaries might have information on individual characters. For example, they might say 破:形聲。从石,皮聲。本義:石頭開裂, or copy an entry directly from the Shuo-Wen-Chieh-Tzu (說文解字).

However, for words of two or more characters, I can hardly find any etymological information. Some words, such as 剎那 are described as being from Sanskrit, but how old is the oldest attested example of its use? Was the character 剎 created solely for translating this word? How about 電話? I have been told that this two-character word is from Japan; but when was it first coined? When and why was 幾何 repurposed to also mean "geometry?" Barely any of this kind of etymological information can be found in Chinese dictionaries. Etymological dictionaries of Classical Greek and Latin are hundred-fold more detailed, so its seems that despite being dead languages, the histories of their lexicon are more alive than that of Chinese.

I know of a dictionary called <<近現代漢語新詞詞源詞典>>, a copy of which I don't have, but given the name of its title, I am guessing that it only contains etymological information on words recently coined. How about older words, such as 駱駝? Where can I find etymological info on this? If Chinese is truly both a classical and a modern language, then I would expect its etymology to deserve better treatment than this. By the way, I do not include four character expressions here, because the etymologies of four character expressions tend to be much more detailed.

I am also aware of the 漢語大詞典, the best dictionary I have. It is the next best thing, but it only includes examples of word uses from relatively famous authors or archaeological discoveries, so I don't know whether the oldest known examples of them are among the printed examples. For words of foreign origin, it is lacking information on countries from which they were borrowed and dates of borrowing.

4 Answers 4


It is true that you need different dictionaries for older and newer words. For newer words, 近現代漢語辭源 published by 上海辞书出版社 in 2019 is quite comprehensive as far as I know. I believe it is a revised and expanded version of 近現代漢語新詞詞源詞典 by basically the same editors.

For older words, 辭源 is indeed a good dictionary. 漢語大字典 and 漢語大詞典 are also good. "it only includes examples of word uses from relatively famous authors or archaeological discoveries." -- The truly ancient documents (like those written more than 2000 years ago) are almost all very famous. The editors even include new archaeological discoveries. Isn't that good? For relatively newer words (like those that appeared about 1000 years ago), the editors clearly also cite many not very famous authors.

Another thing to consider is that a Chinese "word" is not a natural concept or a concept that can be clearly and easily defined. The boundaries between 字, 词, and 词组 are not so clear. Many words cannot be found in dictionaries because editors don't consider them to be words.

I agree there could be a better etymology dictionary that includes them all, old or new, word or phrase. For now, we have to consult multiple dictionaries and we need background knowledge to form the whole picture.

  • "Many words cannot be found in dictionaries because editors don't consider them to be words." For example?
    – r13
    May 27 at 14:44

Finding out etymologies of words is a complex process, and frequently correct information is impossible to obtain anywhere, even if it would be actually known. I will describe my general outline of actions if I need such data, which might be altered in a perticular case.

  1. Go to Wiktionary. Sometimes, especially if the history of the word is highly idiosyncratic (or especially if it is a hidden borrowing), it will just have the answer. (Example: 和尚)

  2. Go to the 漢語大詞典 and check the meanings in their evolution as it lists them and the earliest attestation it has. At least, that would allow to establish the higher bound of the world (=the moment to which it already definitely existed). Frequently, especially if the world has their classic attestations and a pile of meanings clearly connected between themselves, that would be the end of the story. (Example: 勉强)

However, the 漢語大詞典 is pretty famous for not listing the real earliest attestations of words, sometimes being amiss for centuries! As Endymion Wilkinson bemoans in his Chinese History: A New Manual (Third Edition Harvard University Press, 2013 - see the table on p. 90), it is prone to list a word known from pre-Tang sources and only illustrate it by a sentence from a twentieth-century writer. So, if there are no ancient attestations, a further investigation is needed.

  1. Now we must make sure the word is not a Meiji-Japan coinage reimported to Chinese. As most of the words connected with modern world or Western concepts are, it is a very probable occurrence (and even if the word is ancient, the current meaning may be Japan-coined - see 銀行). So, we go to kotobank.jp and try to find the word in the electronic copy of the 日本国語大辞典. This dictionary has more care towards the actual period of first attestation (even though has the problem of mixing the first appearance of a Sinitic word in a Chinese text in Japan and within a Japanese text). It might lead to additional revelations: say, if a word suddenly has attestations in the Nihon Shoki of the eighth century while the 漢語大詞典 does not, it is a reason to suspect the latter is incomplete. However, if the word only has Meiji attestations, it is usually a clear sign that it is really Meiji-coined. (Looking at the Korean and Vietnamese counterparts can be useful at this moment.)

  2. If we could not conclusively establish the word as a late, Japanese or other, coinage, we should start looking for it in the earlier corpus. Here two corpora would help. First, in what manner does the word appear on ctext.org, especially in the pre-Qin and Han database. If it appears there, the investigation is over. (See 推理)

  3. No matter if it appears or not, the word should also be checked in the Buddhist texts database, such as the Taishō Tripiṭaka at cbetaonline.dila.edu.tw . It is a veritable treasure trove of less rigidly Classical usage of Chinese; finding a non-Classically attested word in, say, a collection of Chán Buddhist collected dialogues is a clear indication that a word is a vernacular Song coinage; sometimes the word could be found in Japanese texts but not in Chinese, pointing at the kind of vernacular usage that was lost in China. (Generally checking if the word in whole is on www.buddhism-dict.net also helps: it might, say, provide the information that a word is a Buddhist Sanskrit calque.)

Usually, this is enough to get the general gist of how the word was developed.


Visit: Masini, Federico. “The Formation of Modern Chinese Lexicon and Its Evolution Toward a National Language: The Period from 1840 to 1898.” Journal of Chinese Linguistics Monograph Series, no. 6 (1993): i–295. http://www.jstor.org/stable/23887926.

for 19th and early 20th century vocab

Read the works of Edwin Pulleybank for ancient Indo-European Loanwords


Have you tried the Outlier Dictionary of Chinese Characters? Alas it requires payment, NOT free. And as at May 28 2022, it is available merely on Smart Phones.

I am bemused by Alexander Z.'s answer recommending Wiktionary! It appears hit and miss for me. Sometimes it is correct. But other times it is wrong — just peruse some of our and questions here.

Go to Wiktionary. Sometimes, especially if the history of the word is highly idiosyncratic (or especially if it is a hidden borrowing), it will just have the answer. (Example: 和尚)

  • Please read my post carefully. The title says, "Chinese etymology dictionary for words of more than one character." That is, MORE THAN ONE (>1).
    – KureKotake
    May 30 at 5:42

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.