Mount a chariot

« previous post | next post »

This has always been a bone of contention with me ever since I started studying Buddhology and Sinology in the late 60s and early 70s, when everybody I knew — Chinese and foreigners, scholars and laypersons alike — pronounced 大乘 and 小乘, the Chinese equivalents of Mahāyāna and Hīnayāna, respectively as dàchéng and xiǎochéng.  But that didn't make sense to me, since Mahayana means "Great Vehicle" and Hīnayāna means "Small Vehicle", i.e., modifier + noun construction, so I formed the opinion that, in Modern Standard Mandarin (MSM) they should be pronounced as dàshèng and xiǎoshèng.  Consequently, I began to use these pronunciations — dàshèng and xiǎoshèng — for Mahayana and Hinayana, rather than dàchéng and xiǎochéng.  At first it seemed odd, causing editors and reviewers to "correct" me.  Slowly, however, over the decades, other scholars began to adopt these readings, dàshèng and xiǎoshèng, until now most knowledgeable Buddhist specialists use them, although the lay public, by and large, still pronounce them dàchéng and xiǎochéng.

It is generally understood — at least in Modern Standard Mandarin — that 乘 as a verb ("mount; ascend") is read chéng (2nd [rising] tone) and as a noun ("war chariot; imperial carriage") is read shèng (4th [departing / falling] tone).  Yet, we can find examples in the verse of Du Fu (712-770) and other Tang poets of it being a noun in píngshēng 平聲 ("level tone", which would include the 2nd / rising tone) — see below for evidence.

What's going on here?

  1. chéng

(BaxterSagart): /*Cə.ləŋ/ (Zhengzhang): /*ɦljɯŋ/

  1. shèng

(BaxterSagart): /*Cə.ləŋ-s/ (Zhengzhang): /*ɦljɯŋs/

From Jonathan Smith

My first nihilistic thought — perhaps a reflection of a more general frame of mind I find myself occupying of late :D — is that the so-called noun sheng4 'chariot' isn't really Mandarin per se but just old pedantry — that said/whined, it would certainly be interesting to look at the poetic tradition to see how much the ping ("level") vs. ze ("oblique") split can tell about how these items were used over the years. Were there particular contexts you had your eye on? Perhaps it will turn out that there is not much of a pattern at all… the very first random article I look at says that in "xī rén yǐ chéng / shèng huánghè qù 昔人已乘黄鹤去" ("Riding a yellow crane, gone long ago was the sage of old" [source]) we should read "shèng", which if true would go to show how little I know about the matter…

From Chris Button:

All I can add is that the oracle-bone form shows a person in a tree, and is used in the name of the military “scout” Wang Cheng. So any sense of “chariot, carriage” is clearly derived later from an earlier notion of climbing/mounting a tree.

Nominalization via the -s suffix is of course well attested. It doesn’t necessarily preclude a nominal use without an -s suffix though. I’m curious how you can be so sure that the rhyming is ping-sheng? Phonological preciseness and approaches to rhyming don’t always have to go hand in hand. Or is this a clear-cut case?   [VHM:  see data from Ning Wu cited below]

An hour after sending me the above two paragraphs, Chris dug up this earlier comment of his from two years ago (he is a master archivist of previous Language Log posts):

Quoting this comment from Sasha Lubotsky:

In my contribution to the "Mummies" conference, I argued that the Chinese word for "chariot" [*kl^yag†] corresponds _not_ to Toch.B kokale (which deviates rather strongly from the Chinese form), but from Toch. B klen̄ke, A klan̄k 'vehicle, Skt. yāna-, vāhana-', Toch. AB klān̄k- 'to ride, travel (by vehicle)', PIE *kleng- (cf. Modern German lenken 'to guide, conduct', Wagenlenker 'charioteer').

Victor H. Mair, ed., The Bronze Age and Early Iron Age Peoples of Eastern Central Asia (Washington, D.C.: Institute for the Study of Man Inc. in collaboration with the University of Pennsylvania Museum Publications, 1998). 2 vols, p. 382.

I was quite confused by this comment until I took a look at the paper from the conference. The comparison is with 乘, which would have been something like OC *lə̀ɲs (from earlier *ljə̀ŋs), rather than chē 車 ("car; wheeled vehicle / machine").

The problem with an external association with 乘 is that the meaning "chariot" derives from the original sense of 乘 *lə̀ɲ "mount, ride, ascend". The morphology with the suffixal -s shows the nominal sense to be derived from the verb, and the graphic form depicted in the oracle-bone form leaves us in no doubt about that.

From South Coblin:

We all know that today the distinction between shèng “carriage” and chéng “mount or ride (a vehicle)” is not well maintained in actual speech. In fact, some people aren’t even aware of it. Instead, there is a strong tendency to use chéng everywhere, especially since shèng has become essentially a text-based or literary form rather than a spoken one.

Now, I suspect that this tendency had already begun by Tang times. And poets, when they needed a word in a certain tone in order to achieve correct prosody in the genre in which they were writing, were perhaps willing to take liberties in favor of colloquial usage. How far this tendency had advanced by that period is something that could be investigated. Of course, there would have been purists back then, as there are today, who took umbrage at this failure to observe correct Classical usage. I seem to recall that there are passages in certain phonological lexica where this issue comes up. But, communis error facit ius. Today shèng is essentially gone even from such set phrases as Dàshèng Fójiào, as I’m sure you are aware.

Note from VHM:

Indeed, except that yours truly has fought a decades long rearguard action on behalf of shèng, always insist on it for myself, and have actually won many adherents among eminent Buddhologists.

From Axel Schuessler:

As I have endeavored to show in vain, the original final *-s had the function to form passives; often these derivations were nouns, therefore the superficial notion that qusheng forms nouns. From the OC perspective, shèng originally meant ‘that which is being mounted’ (pass.), from chéng ‘to mount’. Yet one can just as well refer to a vehicle as to something ‘that one mounts’ (active). Now this is OC. Tang tones, 1000+ years later, are a different matter. Perhaps South has some insights into that. — In Tibetan verbs, -s marks past tense. Compare this to English -ed: this is being ‘learned’ (past passive participle), and also ‘he has learned it (the whole phrase is active).

But… by Tang times that was lost, and qùshēng 去聲 ("departing tone") had become an all purpose derivational device.


Selected readings


Addenda for specialists

From John Carlyle:

One reason qu derivation is a tricky issue in Chinese historical phonology is because we are missing the qu tone volume of the early Qieyun manuscripts. It is hard to tell what was added later by consulting various yinyi and other rime books and which words Lu Fayan himself listed in qu tone. A good place to look for qu tone readings retained in the early reading tradition is Jingdian shiwen.

Sure enough, there is a case of 乘 read as zyingH (in Baxter's notation) in the sound glosses Lu Deming gave for Zuozhuan. In Yin Gong year 1, we find the 乘 of


glossed as

乘。繩證反 zyingH。 注及下同。

​which also applies to the 乘 of the following sentence


​Both are the nominal derivation. My impression is that qu readings like this were fossilized retentions in the reading tradition of certain classical texts and were no longer used in the vernacular, similar to jū  reading of 車 nowadays. I agree with Prof. Schuessler that the derivational process was unproductive (or at least moribund) by Tang times.

Out of curiosity, I also checked Zhuànlì wànxiàng míngyìng 篆隸萬象名義 (which retains a good deal of the fanqie sound glosses of the original Yupian) and Mengzi yinyi to see if I could find other instances of the zyingH reading but could not find any. I suspect the reading was already rather specialized and unusual by Lu Deming's (556?-630) time.

From Ning Wu, visiting professor from Sun Yat-sen University, who is auditing my Poetry and Prose class:

After searching the Complete Tang Poetry yesterday noon on and some other websites for the verses that rhyme with the character 乘, there were several results, including those by Du Fu 杜甫 and Cen Shen 岑参. The character 乘 in these verses rhymes with an even tone, and the “大乘" that appears in these places is exactly the expression "Mahayana" of Buddhism. Therefore, I think it is possible to draw a conclusion that the “乘” of “大乘” should be pronounced in an even tone rather than a falling tone among those literati of the Tang Dynasty.

Professor South Coblin suspected that that tendency had already begun by Tang times, which sounds reasonable, but I don’t think it likely that all Tang poets deliberately used chéng instead of shèng only in order to write their poems, although there have been similar cases when poets wrote poems. If the whole society pronounced the character like shèng, but the literati only used the sound chéng to rhyme when they were writing poems, then, firstly, there was no respect for Buddhism, which had a very high status at that time, and, secondly, the poets who believe Buddhism also used this character to write poems, and they would not have used the even tone chéng intentionally only for the purpose of rhyming.

Five poems that can be retrieved from the Complete Tang Poetry all use chéng to rhyme. In fact, poets after the Tang Dynasty also rhymed with chéng, and without exception, they all rhymed in the even tone. Although rhyming with falling tone is less common when writing Tang poems with rhyme schemes, according to incomplete statistics, no poet has used shèng to rhyme since the Tang Dynasty. Please see those five Tang poems and one Qing poem below.


唐 岑參

























唐 張祜








唐  劉得仁

像教得重興,因師說大乘。 從來悟明主,今去證高僧。

蜀國煙霞異,靈山水月澄。 鄉閭諸善友,喜似見南能。




唐 薛能





清 姚瑩





People often say shàng chéng 上乘, and xià chéng 下乘 in Mandarin, which are from Buddhism, too. The meaning of the character乘 here is exactly the same as the characte 乘 in the 大乘、小乘 and 三乘. Then, why is its pronunciation shèng when the 乘 is in 大乘、小乘、三乘,  but  chéng when the 乘 in 上乘、下乘?


  1. Victor Mair said,

    November 2, 2022 @ 10:36 am

    From a distinguished Buddhist scholar:

    I’ve always followed you (and 漢語大辭典 and 國語辭典) on the proper pronunciation in modern Mandarin. The Language Log post is really helpful on all counts.

  2. John Swindle said,

    November 2, 2022 @ 10:57 am

    In some English-speaking medical/surgical circles in the US, especially in obstetrics, the first vowel in “centimeter” is pronounced in faux-French. It sounds real educated.

  3. Daniel Barkalow said,

    November 2, 2022 @ 2:17 pm

    I could imagine that this is a pronunciation that was lexicalized at some point when that pronunciation was standard or at least common, with its meaning referring to the Buddhist concept, distinct from saying in MSM what the Sanskrit means. This would be like "wrought iron" having fixed an old conjugation and usage of "work": what was originally a descriptive phrase become an opaque name for the substance, such that people can identify the substance well by the idiom but don't get any insight into the process that produces it.

  4. John Swindle said,

    November 2, 2022 @ 4:21 pm

    @Daniel Barkalow: Your explanation via lexicalization is better than mine.

  5. Wolfgang Behr said,

    November 2, 2022 @ 11:59 pm

    As much as I agree with Victor and Axel that a denominal/passive (*-s > qùshēng) reading as a rendering for Skt. yāna makes much more sense, the evidence from poetic rhyming that 乘 was mostly read as píngshēng already in Early Middle Chinese is pretty strong. There is a still useful little article by the late Shào Róngfēn 邵榮芬 (1922-2015) on this problem, putting together the available evidence for such MC readings:

    "'Fóchéng' zhī chéng dú píngshēng zhèng" 「佛乘」之「乘」讀平聲證,in: Zhōngguó Yǔwen Yánjiū (HK) 中國語文研究 8 (1986): 141-146.

    (Unfortunately the volume is not available digitally at the T.T. Ng Centre [] in HK, nor has it been included in Shao's three 文集; and I can't access the library here in Japan before next week. I'll be happy to send out a pdf scan to everyone interested in due course.)

    So, I think, John Carlyle's comparison with the maintenance of jū for 車 as a largely obsolete “learnéd” reading practice today,is quite apt. As I can attest from reactions to my own pronunciation as fóshèng, dàshèng etc. you tend to come across as a highfalutin wisenhimer in Buddhist and Buddhological circles in China doing that.

    Aside: If Lubotsky was right in assuming that this is a borrowing from Toch. B kleṅke, A klaṅk or its PT ancestors, despite the transparent inner-Sinitic derivation from 'to mount', and if Lù Démíng is right in assigning a qùshēng reading to nominal usages of 乘 in texts like the Zuǒzhuàn, as John has shown, this would demarcate an interesting time frame for the contact of Chinese with Tocharian at a time when the final *-s had already been transformed into a suprasegmental feature. It would also contradict some other evidence assembled by Pulleyblank on the _late_ ( > 6th c.) disappearance of *-s in some varieties of Chinese (“Some further evidence regarding Old Chinese *-s and its time of disappearance”, BSOAS 36, 1973, (2): 368-373.) In any case, it would be good to have at least some inner-Chinese evidence for the *k- element of the IE cluster on the Chinese side in Lubotsky's etymology. The only thing I can think of are extremely rare loan relationships like yǐn 隱 <'j+nX < *[ʔ](r)ə[n]ʔ written in the Han stone classics for 乘 in the Shūjīng 書經, where the laryngeal or uvular initial would then have to be interpreted as a reflex of foreign *k-. This doesn't chime well with other early loan relationships in excavated texts, however, where we find shēn 申 < MC < syin < OC *l̥in for 乘 in Chu personal names attested in several texts of the Shàngbó corpus.

  6. Chris Button said,

    November 3, 2022 @ 7:00 am

    @ Wolfgang Behr

    The loan use of 隱 is a weird one. It brings to mind the possible graphic association of 殷 *ʔə̀n with 身 *ɬə̀ɲ

    I don’t see any evidence for a velar in 乘 *lə̀ɲ. The closest you can get from a phonetic perspective in terms of auditory perception is probably the voiceless lateral fricative ɬ- in 申 (and 身). Baxter and Sagart’s fully voiceless l̥- should really be preaspirated ʰl- surfacing as the fricative ɬ-. But 乘 itself is just l-.

RSS feed for comments on this post