The Mandarin grammatical particle "le" — one or many?

« previous post | next post »

When I was learning Mandarin over half a century ago, the more grammatically minded Chinese language teachers argued that historically and functionally there were multiple "le" particles that just happened to end up being written with the simple two-stroke character 了.  Then a contrary movement set in, and linguists tried to prune down all the "le" into two or even one, claiming that all of the different 了 developed out of an ur-了.

The irony of it all is that, before the 20th century, there was no established, systematic, explicit grammar for Sinitic languages in indigenous sources.

See, inter alia, Victor H. Mair (1997), "Ma Jianzhong and the Invention of Chinese Grammar," in Chaofen Sun, ed., Studies on the History of Chinese Syntax. Monograph Series Number 10 of Journal of Chinese Linguistics, 5-26.  (available on JSTOR here)

Mǎshì wéntōng 馬氏文通 (conventionally rendered as "Ma's Grammar", though it would probably be closer to the original meaning in Chinese to translate it as "Written Language Unobstructedness"; 1898)

Just as we have seen in a recent post, before the 20th century there was no Chinese concept of "word":

"HouseHold GarBage" (12/6/19)

Which leads to the question:  can you have grammar without words?

There have been countless papers, articles, dissertations, and monographs on le 了.  Here I'm going to introduce two dissertations on le 了 written within the last few decades and the latest monograph on le 了 as representative of what has been happening with regard to the conceptualization of this protean particle in recent times.

The first dissertation is that of John Rohsenow, "Syntax and Semantics of the Perfect LE in Mandarin Chinese", University of Michigan 1978 (University Microfilms; Ann Arbor, Michigan):

John's brief description of his dissertation:

"Syntax & Semantics" is/was a code word for a 1970s model of grammatical analysis championed by Jim McCawley and George Lakoff called "generative semantics", which was popular in the 1970s in Ann Arbor when I was a student there because they (had) both taught there at that time. My "thesis" was that the verbal particle LE marks "perfect (relative) tense" as well as perfecTIVE (completed) aspect.–The Gen Sem model went out of fashion shortly thereafter, when the "grammar wars" of the late 1960s-70s in linguistics terminated.

Next to be considered is Ziqiang Shi's 1988 UPenn dissertation, "The present and past of the particle 'LE' in Mandarin Chinese".  Ziqiang often came to consult with me on the historical background (especially the medieval period) of le 了.


In the field of Tense and Aspect, semantic categories such as perfectivity and imperfectivity have been assumed to be universal. However, these categories have not always been explicitly defined, and furthermore, their purported universal status has not been tested in languages other than the well-studied few. Researchers studying the particle LE in Mandarin Chinese have assumed the universality of such categories as perfectivity, inchoativity and current relevance. The particle LE may occur post-verbally or sentence-finally. When occurring after the verb, LE often signals completion, and when at the end of the sentence, it usually marks new situations. It has therefore been concluded that there are two different LEs in Chinese, one a perfective marker and the other a marker of inchoativity or current relevance. However, the obvious alternative to this conclusion, i.e. there is only one LE which encodes some semantic distinction underlying both perfectivity and inchoativity/current relevance, is left unexplored. In this dissertation, I seek first, to examine this alternative analysis of the particle LE. In the process, attempts are made to define such semantic categories as perfectivity and inchoativity explicitly. It will be argued that the function of the particle can be identified as marking relative anteriority, and that perfective and inchoative interpretations are a result of the interaction between the boundedness of situations and their relatively anterior status as signalled by the aspectual marker. Second, I explore the origin and development of the particle LE through time. It will be shown that the particle started out as a main verb taking sentential subjects and then went on to replace competing forms in the language. Once the function of the new grammaticalized abstract particle was established, it became stable over a long period of time. Finally, I investigate the variation of the particle LE in actual use. Possible factors conditioning the variation are identified and explanations of the variation proposed.

Now we have Chungeng Zhu's brand new monograph, just published in October of this year:

Chinese Aspectual Particle le: A Comprehensive Guide (Hong Kong:  Hong Kong University Press, 2019).

HKU Press provides this description of the book:

The Chinese language has no tense but has aspects. It relies on aspectual particles to express how an action or state relates to the flow of time. Among all Chinese aspectual particles, le (了) is the most frequently used and is notoriously elusive for non-native speakers to grasp.

Chinese Aspectual Particle le: A Comprehensive Guide is entirely devoted to le. It presents a systematic analysis of le and includes detailed illustrations of its usage. Breaking le down into le₁ and le₂, this book illustrates how le₁ forms the perfective aspect: when it is obligatory, incompatible, or optional, and when le can be interchangeable with guo (过) and zhe (着). It shows how le₂ denotes a change of state, performs as a modal particle, and is used in discourse. By comparing Chinese grammar with that of English whenever relevant, the book makes the usage of le more assessible to English speakers. It also contains a plethora of illustrative sentences, a wide range of vocabulary, and abundant cultural information.

And I wrote the following "Foreword" for the book:

The biggest challenge of Chinese grammar for students and instructors of Mandarin is arguably how to cope with the protean particle le (了). Is it one morpheme or more than one morpheme? What are its origins? What are its functions? Does it indicate tense? Aspect? Mode? Something else? Does it only have one pronunciation? It is clear that, right from the outset, the learner of Mandarin is confronted with a host of imponderables, and the teacher is stymied by how to present this superficially monolithic particle to his or her students in the most efficient way while not needlessly confusing them with extraneous ideas and superfluous information.

Many great (and some not so great) linguists have tackled le, both as a linguistic topic and as a pedagogical subject. Countless dissertations, theses, and articles have been published on le, but they almost always focus on one facet of this enigmatic particle or on a closely related group of features concerning it. One of the most distinctive qualities of Chinese Aspectual Particle le: A Comprehensive Guide is its unswerving determination to achieve complete comprehensiveness in covering the quintessential particle le, as is signaled by the subtitle.

The author, Chungeng Zhu, has had long experience in the classroom, so he is familiar with the types of problems and difficulties that English-speaking students face when striving to master the complexities of Mandarin grammar. With that in mind, he coauthored A Chinese Grammar for English Speakers, which has been very successful in helping learners come to grips with the grammatical system as a whole. Having finished that task, he then conceived of an entire book devoted to le as the linchpin for advanced studies of Chinese grammar.

In the course of this extensive volume, the author leaves few le stones unturned. He basically breaks le down into le1 and le2, the former signifying perfective aspect and the latter change of state. The first part of the book concentrates on the former and the second part on the latter. A separate chapter demonstrates how le1 and le2 complement each other. The final chapter deals with the lexicalization of le, which shows how modern lexemes have been formed on the foundation of le.

One of the most outstanding characteristics of Chinese Aspectual Particle le: A Comprehensive Guide is the author’s method for presenting all example sentences. He gives first the sentence(s) in Chinese characters, then the transcription in Hanyu Pinyin (the official romanization of the People’s Republic of China), followed by the English translation. The illustrative sentences are apt, and the English translations are accurate and made with an eye to con- veying the grammatical issues at play. What pleases me most of all, however, and what is missing from many textbooks and guidebooks for the study of Sinitic languages, including Mandarin, is the inclusion of the Hanyu Pinyin transcription.

The author takes Hanyu Pinyin very seriously, and I commend him for that. He has relied on the official “Basic Rules for Hanyu Pinyin Orthography” (Hànyǔ Pīnyīn zhèngcífǎ jīběn guīzé 汉语拼音正词法基本规则) promulgated by the Commission for Pinyin Orthography of the State Language Commission of the PRC. These rules are readily available in the back of all bilingual Chinese dictionaries in the ABC Series published by the University of Hawai‘i Press. The author has repeatedly checked the transcriptions of the Chinese sentences to ensure the greatest possible accuracy.

I cannot emphasize too strongly how important it is to have reliable Hanyu Pinyin transcriptions for all proper nouns, titles, illustrative sentences, and cited texts in Chinese characters. In the first place, there is no guesswork or ambiguity about how to pronounce the Chinese characters, and students do not have to waste time looking up the readings of characters, some of which can be highly refractory.

In addition to the ten chapters of the book that present the diverse roles of le in modern Mandarin grammar, there is a straightforward Preface that provides a lucid exposition of the practical aims of the volume and an illuminating Introduction that offers insightful comparative remarks regarding English and other Indo-European languages vis-à-vis Sinitic languages (the author is cognizant of this theme throughout the book), a succinct overview of aspect in Chinese, and a sketch of the development of le throughout history, tracing the grammatization of le to the Eastern Han dynasty (25–220 CE) and demonstrating how it evolved from a verb into an aspectual particle, with precise quotations from relevant texts along the way. The Introduction concludes with a conceptual approach that puts le1 and le2 in the context of their use in Mandarin.

The layout of the book is designed to make it easy for the reader to detect the fine nuances of the various applications of le, with shaded headings and colored type for salient items. Aside from the Hanyu Pinyin romanizations that I have highlighted above, another unusual distinguishing feature of this volume is that the hundreds of example sentences are interesting in their own right. They are not, as is often the case with many language textbooks, dreary and dull and boring. Rather, they offer the prospect for learning a considerable amount of valuable knowledge about the intersection of Western and Eastern culture. In other words, with his example sentences, the author is not merely concerned with conveying some grammatical point, but he also takes delight in presenting to his readers materials that are edifying and stimulating.

The Bibliography at the end of the volume affords evidence of the substantial scholarly resources upon which the author has drawn in the development of this impressive volume.

Everyone involved in the study and teaching of Mandarin can be grateful to Chungeng Zhu for having written Chinese Aspectual Particle le: A Comprehensive Guide. On the surface, it may seem as though he was only writing an esoteric, virtuoso tome on a two-stroke character (了). As one is drawn through its captivating pages, however, one soon realizes that what he has achieved is nothing less than the ample explication of one of the thorniest conundrums in Chinese linguistics—how to understand le.

What, after all, is 了?

According to zdic, it can be pronounced liǎo, which means "understand" (but that pronunciation can also mean "finish", which is the original signification of the character [see below under the etymological note]).  It's curious, though, that 了完, which means "finished" (with both morphemes meaning "finish"), is customarily pronounced "le wán".

Pronounced "le", it signifies / 
indicates completed action, change of state, future certainty, speeding up or stopping; also serves as an interjection and has half a dozen other usages and is highly productive in the formation of idiomatic usages (for an example, see the last four characters of the last paragraph of this post).

The complexities of le 了 are adumbrated in this etymological note from Wiktionary:

Verb “to finish; to be completed” > perfective aspect particle (了₁, weakened form) > change-of-state modal particle (了₂).

Two kinds of particle uses of can be distinguished: the perfective aspect particle after verbs (conventionally written as 了₁) and the sentence-final modal particle (了₂). It is generally accepted (Wu, 1998) that these two uses of are derived from the concrete verb “to finish”. The grammaticalisation of this verb had become common in the Tang Dynasty, initially in the form of ‹verb + (object) + perfective › to indicate the completion of an action.

The perfective particle subsequently underwent further grammaticalisation to become the sentence-final change-of-state modal particle; Liu (1985) has demonstrated that this last step may have involved the coalescence of sentence-final with in certain Mandarin dialects, as the pronunciations of 了₁ and 了₂ are distinct in these lects, with 了₂ rhyming with .

This word is cognate with Thai แล้ว (lɛ́ɛo, to be finished; already; then, afterwards), Lao ແລ້ວ (lǣu, to finish; to be completed; perfective particle) (Schuessler, 2007).

Do Sinitic languages have grammar?  You bet!  Here are three good descriptions of Mandarin grammar:

Yuen Ren Chao, A Grammar of Spoken Chinese (Berkeley:  University of California Press, 1968)

Charles N. Li and Sandra A. Thompson, Mandarin Chinese: A functional reference grammar (Berkeley:  University of California Press, 1981)

Jeroen Wiedenhof, A Grammar of Mandarin (Amsterdam, Philadelphia:  John Benjamins, 2015)

And here's one for Cantonese:

Stephen Matthews and Virginia Yip, Cantonese, A Comprehensive Grammar (London:  Routledge, 2011 [2nd edition])

The Chinese title of Chungeng Zhu's book is instructive:

Hànyǔ dòngtài zhùcí “le”:  Yīliǎobǎiliǎo

汉语动态助词“了”:  一了百了

Sinitic Aspect Particle "le":

One finished, all is finished
All troubles end when the main trouble ends
As the main item has been solved, all the rest are therefore solved, too

[VHM:  translations of the idiomatic subtitle are from Baidu Fanyi.  Make of them what you will.]


Selected readings

"Why Literary Sinitic is so darn hard" (5/30/19)

"A new grammar of Mandarin" (10/18/15)

"The Westernization of Chinese" (9/6/12)


  1. Chris Button said,

    December 7, 2019 @ 8:58 pm

    According to zdic, it can be pronounced liǎo, which means "understand" (but that pronunciation can also mean "finish", which is the original signification of the character [see below under the etymological note]). It's curious, though, that 了完, which means "finished" (with both morphemes meaning "finish"), is customarily pronounced "le wán".

    I had always assumed that the "le" pronunciation was the unstressed form of "liǎo". First the distinctive tone goes and then, presumably due to it being a grammatical particle, schwa reduction follows.

  2. Chris Button said,

    December 7, 2019 @ 9:05 pm

    Liu (1985) has demonstrated that this last step may have involved the coalescence of sentence-final 了 with 也 in certain Mandarin dialects, as the pronunciations of 了₁ and 了₂ are distinct in these lects, with 了₂ rhyming with 也.

    I haven't read Liu's work, but the proposal seems eminently plausible. It connects with some of the comments made recently here:

  3. Jeffrey said,

    December 7, 2019 @ 10:14 pm

    This entry is fantastic. As I've mentioned, I'm studying Chinese at a university here in China and we've covered both uses of -le. The teachers haven't assigned them names, but I write PF-le and CS-le, for the perfective and change-of-state interpretations of -le. The history of little -le looks interesting. I'll be digging in.

    One issue is that our teachers avoid tackling the difference between a word like gongzuo and a word like shuijiao. In my first semester, as we started to learn the basics of -le, I figured out that you can write shuilejiao but not gonglezuo. I asked the teacher why. She said you just can't. End of story.

    To non-native students, gongzuo and shuijiao look exactly the same. I now know that shuijiao is a liheci, but our teachers have never talked about them. Is there any way of looking at the two parts of the word to figure out if the second part is separable?

    My guess is that the differences are deep and automatic for Chinese speakers, the way the complexities of phrasal verbs are deep and automatic for English speakers. But as teachers, they need to be able to bring those deep aspects to the surface for non-native speakers, just as they do with teaching how to use -le.

    I've used LI and Thompson's book, but not Yuen Ren Chao's A Grammar of Spoken Chinese. I'll have to show that title to my listening and speaking teacher, the one who believes spoken Chinese has no grammar.

  4. Jonathan Smith said,

    December 7, 2019 @ 11:32 pm

    Fascinating. Based just on the above, Shi's dissertation ("semantic categories such as perfectivity and imperfectivity have been assumed to be universal. However…") sounds startlingly contemporary, while Zhu's book (sounds like it) works very much within the traditional "le1/le2" framework…

    Is "lewan2 了完" part of living language? If not, I'm not sure what we are to make of this pronunciation… but it could well be.

    @Jeffrey Interestingly, if I recall correctly, Li & Thompson (1981) remark e.g. that 革命 is "inseparable", but it is certainly separable these days… so it seems there is a Chinese predilection to reanalysis certain verbs as structurally "V+O"… besides 睡觉, words like 学习, 游泳 and a host of others which historically are in no way verb + noun get subjected to this treatment… I remember hearing a young child say "Ta1 yao4 tai1 huir3 mao4" to explain that a doll was being punished with a "timeout" :D Needless to say there is lots of literature on this topic…

  5. Jonathan Smith said,

    December 7, 2019 @ 11:43 pm

    One thought is that given that V is directly followed by information such as duration / instances / manner / complement (when present), for learners the vast majority of heard examples of (e.g.) "do thing for a while" will take the form monosyllabic verb + duration + object because of the ubiquity of monosyllabic everyday verbs in combination with "dummy" O, e.g., kan4 huier3 shu1… so perhaps it makes sense that one would start to put the odd disyllabic verb into this same basket and say, e.g. shui4 huir3 jiao4 rather than shui4 jiao4 huir3, etc.

  6. Victor Mair said,

    December 8, 2019 @ 1:12 am

    Suggestions from Diana Shuheng Zhang for how to translate Mǎshì wéntōng 馬氏文通. From her side, she sees two options:

    1) Ma's Chinese Grammatology. γράμμα (grámma) = γράφω "I write" + -μα "result noun suffix", means "writing, that which is in written form". From this ancient Greek root the English "grammar" is derived. And 馬氏文通 is indeed a grammatological work on the "-ology" of "gramma" 文.

    2) Ma's Chinese Textual Conspectus. This one is more word-to-word literal: 文 textual 通 conspectus. For 通, con- "together, strengthen"; spect- "to see, examine". Therefore, conspectus: A detailed survey or overview of a subject.

    So, one Greek-derived translation versus one that is Latin-derived.

  7. Jeffrey said,

    December 8, 2019 @ 4:51 am

    @Jonathan Smith

    That little girl really said, "Ta yao tai huir mao"? That's so funny — and great.

    I like your idea of reanalysis. I think you're right. Chinese speakers must have the choice to analyze the second part of a two-syllable verb as either a verb or a noun. Due to a lack of suffixes signalling the function in the word, the invariant Chinese words are very fuzzy. For example, ganmao can be a verb or a noun: Wo ganmao le, or Nide ganmao hao le ma? Must be the same here. And by the way, the fuzziness of Chinese words was the first obstacle I had to cross as a beginning student.

    In my Pleco dictionary, the two parts of a liheci verb are separated by double forward slashes, which helps me a lot: du//shu, zou//lu, and sao//di. I can see the nouns: shu, lu, and di. These are clearly V+N liheci verbs. But there are others, like shuijiao and liaotian that are not as easy for me to interpret. And that's what I'm trying to figure out now.

    I have a long list of V+V and V+N verbs. I will occasionally check them with my teachers. Sometimes they really have to think which column a verb belongs to.

  8. Michael Watts said,

    December 8, 2019 @ 5:00 am

    The Chinese language has no tense but has aspects.

    Then what exactly is the 有 in e.g. 我没有看到他 expressing?

    Pronounced "le", it signifies / indicates completed action, change of state, future certainty, speeding up or stopping; also serves as an interjection and has half a dozen other usages and is highly productive in the formation of idiomatic usages

    The "le" pronunciation is not universal. The popular song 小苹果 includes these lyrics:

    春天又来到了,[Spring has come again]
    花儿开满山坡 [blooming flowers cover the hillside]

    That 了 at the end of the first line must be the change-of-state marker, but it's pronounced liao.

  9. Jeffrey said,

    December 8, 2019 @ 5:05 am

    @Jonathan Smith

    I'm only a beginning student of Chinese, but I can follow your idea of how adding duration to the dummy object might have created that basket, as you say.

    In our class, we learned pretty quickly the pattern: Zuotian wanshang wo shuile qi ge xiaoshi de jiao. My classmates and I ask, "Okay, what's a jiao?" Teacher: "Sleep." Huh? Us: "Last night I slept a seven hour sleep?" Teacher: "Dui." Students: Hmm. Yup. There's your dummy object.

  10. Michael Watts said,

    December 8, 2019 @ 5:19 am

    My guess is that the differences are deep and automatic for Chinese speakers, the way the complexities of phrasal verbs are deep and automatic for English speakers.

    Sure, but V+O verbs in Chinese have a much more direct analogue in English; we do exactly the same thing. How do you know that the past tense of "catch fire" is "caught fire" rather than "catch fired"? The unsatisfying explanation is still the correct one — some verbs are in one category, some are in another, and you know which is which by memorizing them all. As discussed upthread, it's not a foolproof system.

    In my Pleco dictionary, the two parts of a liheci verb are separated by double forward slashes, which helps me a lot

    Hm, I don't see this in my Pleco. Regardless, I recommend you spring for the ABC dictionary, which will note V+O verbs with the part of speech "V.O.". (It's also better than the free Pleco dictionary — this is a case of "you get what you pay for".)

  11. Michael Watts said,

    December 8, 2019 @ 5:28 am

    I had always assumed that the "le" pronunciation was the unstressed form of "liǎo". First the distinctive tone goes and then, presumably due to it being a grammatical particle, schwa reduction follows.

    Hmmmm. Is the vowel of le reduced? I certainly think of it that way, but I generally assume that's a problem in me based on the English schwa being the closest English vowel.

    But le rhymes with she, and she contrasts with shi, and shi is the syllable traditionally thought of as having a zero rime. It looks difficult to claim that she features a reduced vowel…? :-/

  12. Jeffrey said,

    December 8, 2019 @ 6:16 am

    @Michael Watts

    Okay, got it. Mostly idiosyncratic. It's necessary to memorize the liheci verbs. I can do that. Thanks.

    I'm using a Pleco version that a Chinese person downloaded onto to my shouji here in China. That ABC dictionary sounds good. I'll check it out.

  13. Philip Taylor said,

    December 8, 2019 @ 6:33 am

    Victor — ' Victor H. Mair (1997), "Ma Jianzhong and the Invention of Chinese Grammar" '. May I ask whether, 20 years on, you would still speak of the "invention" of Chinese grammar ? I ask because it seems to me that natural languages (at least, those that have evoled beyond grunt-and-point") will have a grammar, even if it has never been formally recorded or described; the grammar will be the set of mutually-agreed (if unspoken) rules by which two or more of its speakers successfully communicate — sharing a vocabulary is insufficient, there must also be a shared grammar in order for communication to be possible.

  14. Antonio L. Banderas said,

    December 8, 2019 @ 7:14 am

    Can somebody please mention some kind of academic address of the author Chungeng Zhu?
    After googling for an hour, I haven't found any form of contact whatsoever.

  15. Chris Button said,

    December 8, 2019 @ 7:58 am

    @ Michael Watts

    In many respects, I think schwa should really be treated as a feature of syllabification rather than a "vowel" per se. As such it's not easy to pin down precisely phonetically.

    It is logically therefore what we should expect to find at the ultimate end of a process of reconstruction that treats the syllable as the basic functional unit. Hence the Proto-Indo-European default "e" vowel is often treated as /ə/, which with its ablaut variant "o" (/a/) structurally parallels the Old Chinese alternation between /ə/ and/a/ and other proto-languages worldwide once reconstructed back far enough.

  16. Antonio L. Banderas said,

    December 8, 2019 @ 8:48 am

    @Chris Button
    But ə is synchronically a Mandarin phoneme, isn't it?

  17. Victor Mair said,

    December 8, 2019 @ 9:30 am

    "this is surely just a slightly out of place use of 'words' to mean 'lyrics'"

    No, it is not "surely" what you say it is. She was instinctively thinking of cí 詞 in its contemporary, linguistic sense as "word", instead of in its historical, literary sense of "lyric". Given the context of the class and the subject — "shīcí 詩詞" ("poems and lyrics") is a common expression in literary studies — it is clear that it does not mean "poems and words". She just got confused and made a slip-up.

  18. Chris Button said,

    December 8, 2019 @ 9:31 am

    @ Antonio L. Banderas

    Sure, although it depends on your analysis of course. Take many varieties of American English where the "cup" vowel can essentially be treated as schwa despite it not being "reduced". Schwa as a "phoneme" and schwa as "syllabification" aren't contradictory concepts.

  19. Victor Mair said,

    December 8, 2019 @ 9:33 am

    From Tsu-Lin Mei:

    You should read 梅祖麟语言学论文集 and Sun Chaofen, Grammaticalization and word order change in Chinese grammar. The seminal paper is my 现代汉语完成貌句式和词尾的来源 originally published in 1981 in 语言研究。 It is this paper which established my reputation as a student of historical grammar in mainland Chinese and got the attention of Alain Peyraube and Sun Chaofen. Alain came to Cornell to work with me in 1981 (?) and Sun Chaofen especially wanted to come to Cornell to be my graduate student. The theory established is quite simple and readily confirmed by textual evidence, e.g. 敦煌变文集 and 祖堂集。Modern Chinese has 吃了饭了or V le(1) O le(2). Le (1) is post verb and before object and it indicates perfect aspect 完成貌and le(2) is sentence final le indicating new situation. The origin of both le (1) & le(2) is the intransitive verb 了liao meaning “completed , finished.” Before 900 CE the Chinese language has only VO liao. Then liao is grammaticalized, and as a grammaticalized particle it occurs in the position of V le O. This change occurred around 10th or 11th century. And later, the liao which remained in the position of VOliao also underwent grammaticalization and become le (2). The Chinese language developed V-Resultative Complement (e.g. 弄小,打破) sometime around 5th or 6th century. Before that time there was only V(1) V(2) in coordinate construction. With the rise of V-RC the Chinese verb can have a subordinate element following it (e.g. V-RC O) and this opened the way for the rise of V le O. This point came to me when I was giving a seminar at Alain Peyraube’s institute, CRLAO.

  20. Victor Mair said,

    December 8, 2019 @ 10:00 am

    Addendum from Tsu-Lin Mei:

    The title of Sun Chaofen’s book is Word-Order change and grammaticalization in the history of Chinese. This is the expanded and much improved version of Sun’s Ph.D. dissertation at Cornell (1988). Sun had the benefit of learning from Elizabeth Traugott, a distinguished scholar in the history of English language and general linguistics, e.g. on the notion of grammaticalization. So in our work on 近代汉语,I and my Chinese colleagues did the textual work and Sun Chaofen and Alain Peyraube provided the theoretical underpinning, i.e. the theory of grammaticalization.

  21. Victor Mair said,

    December 8, 2019 @ 10:46 am

    2nd addendum from Tsu-Lin Mei:

    I hasten to add one bibliographic item. In my previous messages I talked about the origin of le (1) and how the verb liao became grammaticalized and got into the position of post verb and pre object (i.e. V le O). All that is my work and the book by Sun Chaofen gives a pretty good account in English.. I vaguely indicated that le (2), the sentence-final le had a similar origin but did not state the place you can look up a clear and convincing account. Here it is: 近代汉语虚词研究(1992) , p. 43ff is about 动态助词 “了”, namely le(1). P. 111 ff is about 事态助词 “了”, namely le (2) and here you can find a clear account of the origin of le(2) and its early appearance , mostly in 祖堂集and Song texts. Both sections are by 曹廣顺, a fine scholar who corrected some of my mistakes in historical grammar. I went to Peking U to teach Chinese historical grammar in 1983. 刘坚,江蓝生,曹廣顺all came to my class and in 1992 they edited 近代汉语虚词研究 . When I went to Peking in 1983, they do not know there is a book called 祖堂集。 I bought ten Taiwan cheap reprints and distributed them judiciously and in 1992 all Chinese scholars know 祖堂集 and some of them did ground breaking analysis, e.g. le (2).

  22. Victor Mair said,

    December 8, 2019 @ 11:33 am

    I quite agree with Tsu-Lin about the importance of Zutang ji 祖堂集 for studying the development of Middle Vernacular Sinitic (MVS). Along with the Zhuang Zi and Shishuo xinyu, it is one of my favorite Chinese books. Here are some resources for becoming better acquainted with Zutang ji 祖堂集, drawn from this website:

    靜 Jing (n.d.) & 筠 Yun (n.d.)
    祖堂集 Zutang ji

    (Koreai átírás:) 정 Chŏng/Jeong & 균 Kyun/Gyun: 조당집 Chodang chip/Jodang jip
    (Rōmaji:) Jō & In: Sodōshū
    (English:) A Collection from the Halls of the Patriachs / Patriarch's Hall Anthology / Ancestral Hall Collection
    (Magyar:) Cu-tang csi (szerk. Csing & Jün): Az ősatyák csarnokának gyűjteménye / Az ősök csarnoka gyűjtemény

    Compiled in 952, published in Korea in 1245, biographies of 253 figures

    Compiled in 952 in the kingdom of Southern Tang (937–975), Zutang ji 祖堂集 is an invaluable source of information about the formative history of the Chan school and the gradual evolution of Chan literature. Long lost and forgotten in China, only to be rediscovered during the early part of the twentieth century among the woodblocks of the Buddhist canon stored at Haein Monastery 海印寺 in Korea, the text represents an outline of earlier Chan “history,” written from a regional perspective. Among the text's prominent features is its inclusion of unique materials not found in other Chan collections.

    柳田聖山 Yanagida Seizan (1922-2006)

    Yanagida Seizan 柳田聖山, 《《祖堂集》の本文研究》 Zengaku kenkyū 禪學研究 54 (1964), 11-87

    Zutang ji 祖堂集索引. 3 vols. Yanagida Seizan 柳田聖山, ed. Kyoto Daigaku Jinbun Kagaku Kenkyūjo, 1980.

    Seo Kyung-bo
    "A Study of Korean Zen Buddhism Approached Through the Chodangjip"
    Ph.D. dissertation. Temple University, 1960; mimeographed reprint, Seoul, Poryo ̆n'gak, 1973.

    PDF: Lineage and Context in the Patriarch's Hall Collection and the Transmission of the Lamp
    by Albert Welter
    In: The Zen Canon: Understanding the Classic Texts (2004), Chapter 5

    Christoph Anderl
    Studies in the Language of Zu-tang ji 祖堂集
    2 volumes. Oslo, Unipub, 2004, (ca. 980 pages)

    This is study of the grammar of Late Middle Chinese / Early Vernacular Chinese, exemplified by the 10th century Chan text Zǔtáng jí 祖堂集 [Collection From the Patriarchs' Hall]. The first volume includes a long introduction to the text, and the grammar. The second volume includes selective translations, the bibliography, and an index.

    Studies in the Language of Zu-tang ji 祖堂集, VOLUME 1: GRAMMAR (and Introduction, Lineage Charts, Variant Characters)

    Studies in the Language of Zu-tang ji 祖堂集, VOLUME 2: TRANSLATION (and Glossary of Linguistic Terms, Bibliographies, Index)

    Index of Buddhist terms, function words, proper names and glosses in "Studies in the Language of Zu-tang ji 祖堂集"

    Bibliography and Linguistic Glossary Concerning the Study of Early Vernacular Chinese with an Emphasis on the Language of Zǔtáng jí 祖堂集 'Collection from the Patriarchs' Hall' (10th century)

    PDF: Le Recueil de la Salle des patriarches
    "Tsou-t'ang tsi" par Paul Demiéville
    T'oung Pao, Second Series, Vol. 56, Livr. 4/5 (1970), pp. 262-286.

  23. Victor Mair said,

    December 8, 2019 @ 11:35 am

    See also:

    The Annals of the Ancestral Hall (Zutang ji) Study Group

    The Annals of the Ancestral Hall 祖堂集, which predates by fifty-two years the better-known Jingde-era Record of the Lamp, is the oldest complete Zen Lamp Record known. As such it is of inestimable value as an early example of the literature of Tang-dynasty Zen, the Zen from which all modern traditions stem. In comparison with the Jingde-era Record of the Lamp, edited by the leading intellectuals of the Northern Song Dynasty, the Annals of the Ancestral Hall contains much raw material rich in rustic charm. In recent years this material has drawn attention as a resource for the study of colloquial language. The study group commenced research on the Annals of the Ancestral Hall forty years ago under the guidance of Prof. Iriya Yoshitaka and Prof. Yanagida Seizan. The results of this work were published in 2003 by the International Research Institute for Zen Buddhism under the title Annals of the Ancestral Hall: With Notes and Japanese Readings 訓注祖堂集. The study group’s research presently centers on clarifying the background of the text by investigating the Zen masters of the Xuefeng-lineage, who were responsible for the Annals’ compilation. This year the study group will complete reading the chapter on Yunfeng in the seventh fascicle and turn its attention to the successors of Xuefeng, beginning with Xuansha Shibei and Changsheng in the tenth fascicle.

    Past Activities: This group, under its previous identity as the Tang Dynasty Recorded Sayings Study Group A, studied a wide variety of texts under the guidance of Prof. Iriya Yoshitaka and Prof. Yanagida Seizan.

    VHM: See the website for extensive bibliographical references.

    I was very fortunate to read this text with Prof. Iriya Yoshitaka 入矢義高 and to discuss it with Prof. Yanagida Seizan 聖山柳田, two of its foremost Japanese exponents.

  24. Jake said,

    December 8, 2019 @ 2:17 pm

    I thought Thai and Lao were of different genealogies than Mandarin, does the connection here mean (a) I'm wrong, (b) one's borrowed from the other, or (c) it's just one of those things?

  25. Michael Watts said,

    December 8, 2019 @ 3:45 pm

    I have another question about "sentence-final" 了 — it's true that there is a class of sentence-final modal particles, but, on thinking about it, it's not obvious to me that 了 is a member of that class, because it combines with them.

    已经解决了吧 [it's already taken care of 吧]
    你准备好了吗?[Are you ready?]
    如果,两年后,这个男人放弃了呢?[What if, after two years, this man gives up?]

    And of course it combines with sentence-final 啊 so commonly that it forms the contraction 啦 in that case.

    Can any other of the modal particles appear together in the same sentence like this?

  26. Jonathan Smith said,

    December 8, 2019 @ 3:46 pm

    Re: le, a piece of Zhu's book is at; it is interesting to read and builds on the tradition which Tsu-Lin Mei outlines above… it seems he was formerly at U. of North Georgia but don't know about now… in a word, I would say I don't think it's obvious that le1/le2 is a necessary synchronic analysis of Mandarin or a useful teaching device.

    Yes I agree with Michael Watts that from the point of view of the learner, V-O words simply need to be listed as such in the mental lexicon / textbook index, but examined more closely, it is clear that these statuses are changing rather fast and can be variable across speakers and items. Even one item could be treated differently by a single speaker; e.g., adding "complement of manner" to the verb 学习 with 得 ("学习得很努力") remains fine, to which contrast 睡觉, but the former can be heard treated as "V+O" in some instances as with duration…

  27. Zev Handel said,

    December 8, 2019 @ 4:16 pm

    You are correct that the current general consensus among Western scholars is that Tai languages and Sinitic languages belong to two different families that do not have a common genetic ancestor. The entry that is cited from Schuessler 2007 (p. 357, under liǎo₁ 了) , gives related forms not just in Thai and Lao, but also in Vietnamese. Schuessler's implication is that this word was an early borrowing into Chinese. Schuessler also notes that the Chinese word is "medieval", i.e. not attested in Old Chinese sources. However, the r- in the Vietnamese form rồi 'finished', corresponding to Chinese l- < Old Chinese *r-, raises the possibility that the word was already in Chinese in the Old Chinese period despite there being no textual attestation. Although Schuessler doesn't say so, we can infer that he believes this is an Austoasiatic word in origin, inherited by Vietnamese, and borrowed into Tai languages and Sinitic.

    It's worth noting that their are other hypotheses about the relationship between Tai and Sinitic. Most mainland Chinese scholars still subscribe to the older belief that Tai is a branch of Sino-Tibetan, co-equal with Sinitic, Tibeto-Burman, and Hmong-Mien.

    Recently, an idea has gained currency among Western scholars that Tai is a branch of Austronesian. Laurent Sagart has proposed that Austronesian and Sino-Tibetan are themselves two branches of a larger family (this is the "STAN" hypothesis). Under his proposal, Tai and Sinitic have a genetic relationship, but one that is much more distant.

  28. John Swindle said,

    December 8, 2019 @ 4:20 pm

    Jumping back a couple of steps to an easier question … Michael Watts, there's liao meaning "understand," like in 了解 (both parts of which again mean about the same thing, although I suppose that could be disputed) and there's the protean le, but isn't "liao" for "le" in songs just what you do in singing?

  29. Michael Watts said,

    December 8, 2019 @ 5:48 pm

    isn't "liao" for "le" in songs just what you do in singing?

    It might be a thing you can do in singing, like pronouncing 的/地 as di (even when pronounced de elsewhere in the same song), but I don't think it's a default or anything.

    For example, the lyric in 爱的就是你, “像喝了奇怪药水中了魔法”, is pronounced with both 了 as le.

  30. Michael Watts said,

    December 8, 2019 @ 5:49 pm

    (I guess I should note that 像喝了奇怪药水中了魔法 features two aspectual 了 and zero change-of-state 了.)

  31. Jonathan Smith said,

    December 8, 2019 @ 8:06 pm

    [At a party]
    A: When do you think Prof. Mair will get here?
    B: [opens door and sees Prof. Mair] 他来了!
    [During party]
    C: Has Prof. Mair shown up or what? I haven't seen him.
    B: 他来了。
    [Day after party]
    D: Did Prof. Mair come to to party?
    B: 他来了。
    So I suppose (1) is "le (2)" aka change of state / 事态助词了 and (3) is "le (1)" aka perfect aspect / 动态助词了, but I'm not sure about (2)… is B saying "He came", "He has arrived", or "He is here", and does it matter? It seems fine to translate respectively (1) "He's here", (2) "He's here", and (3) "He was there", i.e., no sign of "perfect aspect" per se at all. In this light, note Michael Watts'(s) two song lines are certainly different; in 喝了药水 one is done drinking; in 中了魔法 one is now under the spell's effects…

    And for (2) and (3) above, if the answer were "no" we could of course say "他没来" in both cases, so are there also 没2=hasn't and 没1=didn't? Would such a 1 vs. 2 be a useful explanatory device or theoretical analysis above and beyond plain discussion of examples of shades of meaning and pragmatics? How many 吧 are there? 您是梅教授吧?… 还是告诉梅教授吧 … 那就选梅教授吧… How many 呢? 你呢? … 玩儿呢 … 手机呢?… etc. etc.

  32. Michael Watts said,

    December 8, 2019 @ 9:38 pm

    In this light, note Michael Watts'(s) two song lines are certainly different; in 喝了药水 one is done drinking; in 中了魔法 one is now under the spell's effects…

    I can't agree with this; they're parallel in every way. It's not even possible to assume that one is currently under the spell's effects — whether you were cured later or not, you were hit by the spell when you drank the potion.

    Note that the verbs you're worried about, 来 [come] and 中 [(be) hit], themselves express a change of state. On purely semantic/logical grounds, to say that a change of state has finished happening is to say that a new state obtains. But I see no reason to assume that just because it's possible to apply perfective 了 to verbs indicating a change of state means that perfective 了 itself indicates a change of state in those cases. That weight can be borne by the verb.

  33. Jonathan Smith said,

    December 8, 2019 @ 10:53 pm

    As you almost say, if "something happened" then "a new state obtains", and so I'm personally not convinced that it is possible or useful to strictly divide 了 into an item that marks the former versus an item that marks the latter.

  34. Jonathan Smith said,

    December 8, 2019 @ 10:56 pm

    Also re: above "呢吗" I suppose though some might scorn it

  35. Antonio L. Banderas said,

    December 9, 2019 @ 6:02 am

    @Michael Watts
    What is the technical meaning of 啦 being a "contraction" of 了啊?

  36. liuyao said,

    December 9, 2019 @ 11:33 am

    The way that shuijiao 睡觉 has become a VO verb may in part be influenced by such words as wujiao 午觉 ("noon nap").

    For xuexi 学习, there are zixi 自习, yuxi 预习, fuxi 复习, etc. [And now there is
    Xi the emperor.]

    For youyong 游泳, the four strokes are named wayong 蛙泳 ("frog swim", or breaststroke), yangyong 仰泳 (backstroke), dieyong 蝶泳 (butterfly), and ziyouyong 自由泳 (freestyle).

    Not knowing Pleco, I think it may be helpful if, for such VO verbs, you get to see some of the (bisyllabic) nouns that end with the same second character.

  37. Michael Watts said,

    December 9, 2019 @ 4:25 pm

    What is the technical meaning of 啦 being a "contraction" of 了啊?

    That it is formed from a relaxed pronunciation of 了啊 as a single syllable. I believe it is common for 啊 in specific to form contractions with unstressed words that may (otherwise) end sentences — most prominently 了, but we also have e.g. 的+啊 = 哒.

    More practically, if you ask people what 啦 means, they'll tell you it's 了+啊. The two are fully synonymous.

    I'm also aware of contractions for buyao 不要 -> bie (别), buyong 不用 -> beng (甭), and xuyao 需要 -> xiao (?). Sometimes when people say shíhòu 时候 it sounds to me like the single syllable shòu, but the one person I asked about that denied it.

    buyao -> bie would seem to support Chris Button's theory of liao -> le, in that the obvious first step of contracting buyao would be biao, not bie.

    Not knowing Pleco, I think it may be helpful if, for such VO verbs, you get to see some of the (bisyllabic) nouns that end with the same second character.

    It's possible to do something like this in Pleco, but the interface is bad — you'd be looking at an alphabetized list of all words (with dictionary entries) ending in a particular character. I don't find it useful.

  38. Jake said,

    December 9, 2019 @ 6:18 pm

    @Tev: Thanks!

  39. Alison said,

    December 9, 2019 @ 7:19 pm

    On the topic of 睡觉 shuìjiào and similar verbs… I am not sure about the linguistics, but as a student of Chinese I found it helpful to think of them as German separable verbs like anfangen or mitnehmen.

    In German the usage of separable verbs tends to be that one part of it must always end the sentence. So mitnehmen (to take with) is structured in a sentence like "will you take your bag with?" In Chinese, 睡觉 is also often used that that way. In German there are other verbs that look like they should be separable but they are not. Same goes for Chinese.

    As a result – and I am not sure if this is grammatically correct – 睡觉 lives in my head alongside verbs 看见 kànjiàn or 拿走 názǒu which are those verbs that if you put a different result word on it then they have a different meaning than they would on their own. This makes sense to me because you can also 睡着 (fall asleep) or just 睡 by itself in certain contexts.

    Anyway this is a bit of a tangent. I don't have anything insightful to say on 了, although I will say that I find the whatever-不了 bùliǎo construct awkward and irritating. To me it comes across sounding like baby talk, like those "cute" doubled up words (e.g. 去看看 qùkànkàn). I find myself going out of my way to avoid it because I don't want to sound twee. I'm curious about the history of that construct.

  40. Michael Watts said,

    December 10, 2019 @ 4:03 am

    I will say that I find the whatever-不了 bùliǎo construct awkward and irritating. To me it comes across sounding like baby talk, like those "cute" doubled up words (e.g. 去看看 qùkànkàn).

    Do you happen to know why you perceive it this way? Reduplicated nouns really are diminutive — they feature prominently in Chinese baby talk (as in, how adults talk to babies). And it is my understanding that reduplicated verbs partake of some similar "cute" or "diminutive" quality.

    But (in)ability infixes aren't like that at all. :-/

  41. Alison said,

    December 10, 2019 @ 8:56 am

    That's a good question. In my office, 不了 is frequently used in work-related contexts, but for some reason when i hear it i still associate it more with whining children or 撒娇的 women. I associate a certain sing-song version of 不可能 with the same thing. I don't mind writing it, i just don't like saying it. This might be just one of those odd things that happens when you learn a foreign language and you haven't been desensitized to a phrase enough to shake the original (or most "sticky") association.

  42. Chris Button said,

    December 10, 2019 @ 11:31 pm

    buyao -> bie would seem to support Chris Button's theory of liao -> le

    Thanks, although I'm pretty sure the proposal must have been made countless times before in the literature. I don't think it's a radical idea.

    Hence the Proto-Indo-European default "e" vowel is often treated as /ə/, which with its ablaut variant "o" (/a/) structurally parallels the Old Chinese alternation between /ə/ and/a/ and other proto-languages worldwide once reconstructed back far enough.

    To be clear, the suggestion here isn't that language nowadays functions any differently from proto-languages. Rather that it is only through the comparative historical method that we can get a deep enough insight into the underlying phonological structure.

    Sometimes the ə/a is clear in the current phonology (although not on a surface phonetic level of course where we have all manner of vowels) as in some NW Caucasian languages. Other times we only need to go back to a certain (internal) time depth, such as in Burmese where inscriptional evidence unequivocally demonstrates an origin in earlier ə/a. And then in cases like Old Chinese, we need to reconstruct quite a long way back.

    The question of time depth is only relevant in so far as often being able to gain an insight into the underlying structure. The comparative method is inherently a historical process even when it only involves comparing modern languages. In short, I'm suggesting that the underlying ə/a structure (or in more extreme terms, phonologically–but not phonetically–"vowelless" structure) itself persists throughout language–sometimes and in some languages it is just more heavily obscured than at other times and in other languages.

RSS feed for comments on this post