An Anatomy of Chinese

An Anatomy of Chinese
Rhythm,
Metaphor,
Politics
perry link
Rhythm, Metaphor, Politics
p e r ry l i n k
h a rva r d u n i v e r s i t y p r e s s
Cambridge, Massachusetts, and London, England
2013
Copyright © 2013 by the President and Fellows of Harvard College
All rights reserved
Printed in the United States of America
Library of Congress Cataloging-in-Publication Data
Link, E. Perry (Eugene Perry), 1944–

An anatomy of Chinese : rhythm, metaphor, politics / Perry Link.
p. cm.
Includes bibliographical references and index.
ISBN 978- 0-674- 06602-1 (alk. paper)
1. Chinese language—Rhythm. 2. Chinese language— Metaphors.
3. Chinese language—Terms and phrases. 4. Chinese language—Semantics.
5. Chinese language— Political aspects. I. Title.
PL1279.L483 2013
495.1'16—dc23 2012015096
To the sprightly spirit of Y. R. Chao—departed,
and yet not—and to everyone else who loves
the sounds and structures of spoken Chinese
Contents
Introduction 1
1 Rhythm 21
The Prevalence of Rhythmic Patterns
in Daily-Life Chinese 24
Is Rhythm Unusually Common in Chinese? 37
Speakers’ Awareness of Rhythm 40
Are There Fads in Rhythms? 44
The Roots of Rhythms 49
“External” Rhythms: Dominant and Recessive 54
Recessive Rhythms of Favor 60
How Recessive Rhythms Affect Structure 68
How Universal Are the Preferred Rhythms
of Chinese? 74
Do Rhythms Have Meanings? 82
What Other Formal Features Contribute to
Meaning? 94
Can the Users of Rhythm Be Unaware of
Its Effects? 109
2 Metaphor 113
How Do Metaphors Work in Ordinary Language? 115
viii Contents
Metaphor and Thought 128

Time 136
Color 147
Up and Down 155
North and South 162
Consciousness 169
The Self in Ancient Thought 171
Privilege in Dyads 174
Metaphors That Chinese and English (Pretty Much)
Share 183
Metaphors in Chinese That Diverge from
English in Significant Ways 198
Conceptual Differences That Are Rooted in
Metaphor 209
Can Conceptual Metaphors Generate Philosophical
Problems? 215
The Significance of Similarities and Differences among
Conceptual Metaphors in Different Languages 231
3 Politics 234
A Bifurcation 235
Characteristics of the Official Language 243
The Language Game 278
How the Game Is Played: From the Side
of the Rulers 295
How the Game Is Played: From the Side of the
Ruled 321
Effects of the Language Game in the Mao and the
Post-Mao Eras Compared 341
Epilogue 349
Acknowledgments 357
Index 359
Introduction
This book has grown from files that I have kept for more than three de-
cades on items that have fascinated me about the Chinese language.
When I dug into those fi les a few years ago I found that, in order to inter-
pret their contents responsibly, I would need to read in a number of
fields—prosody, cognitive science, philosophy of mind, comparative poli-
tics, even music theory—that were largely new to me. I asked colleagues
for introductions to those fields and, with my store of Chinese examples
serving as test cases, found the resulting exploration both pleasurable and
rewarding.
Academic specialties in the early twenty-first century bristle with their
own jargons, and nearly every area I turned to required an investment of
time before I felt I was getting what was there to be got. Each investment,
in the end, was worthwhile. Special jargon does serve a purpose when it
allows expression of thought that could not have been put as precisely in
ordinary language. But that said, I should say as well that I often found
the jargon of subfields to be not entirely necessary: the same thought
could often have been put, just as clearly or more so, in plainer language.
Why we academics like jargon is an important question. It relates to why
many humanities and social science disciplines are becoming more self-
contained, growing as if in parallel universes; why students can be puzzled
as they move from economics to anthropology to literature, even if their
teachers claim to be sharing a subject (“China,” for example); and why
2 An Anatomy of Chinese
deans announce programs and deliver homilies to encourage “interdisci-

plinary” approaches. But an adequate analysis of the jargon problem is
beyond my scope here. I mention it briefly only in order to suggest why I
have tried to minimize it in this book. I admire the view of the eminent
Chinese linguist Y. R. Chao, who wrote:
[Usually] I shall prefer to use a familiar term, with a warning against

making unwarranted inferences, in preference to using unfamiliar
terms, which, though safe from being misunderstood, are often also
safe from being understood.1
I will use technical terms only where I think plain language will not do.
But my aim is clarity, not, alas, complete avoidance of tedium. Scholarship
depends on attention to detail, and the demands of detail can make sen-
tences and whole paragraphs sometimes seem dry, even if they are jargon
free.
I want to start you where I started—noticing a few interesting little
facts about the Chinese language that can lead, if one pursues them, into
much larger areas.
In the fall of 1988, shortly after I arrived in Beijing for a year of work in
the Beijing office of the Committee on Scholarly Communication with
China (administered by the U.S. National Academy of Sciences), I noticed
a sign that was intended for pedestrians crossing Haidian Road. In other
countries such a sign might have said “Caution” or “Look both ways.” But
this one read: Yi kan, er man, san tongguo ϔⳟˈѠ᜶ˈϝ䗮䖛 ‘First look,
then go slowly, then cross’.2 The phrase is not only rhythmic but exhibits the
1–2, 1–2, 1–2–3 pattern of syllables that is at least as old as mirror inscrip-
tions of the Han period3 and that has pervaded not only elite poetry but
1. Yuen Ren Chao, A Grammar of Spoken Chinese (Berkeley: University of California

Press, 1968), p. 136.
2. The Chinese examples in this book are given in either traditional or simplified
characters, 㐕储ᄫ or ㅔԧᄫ. Since examples are drawn from a range that extends
from Zhuangzi to Hu Jintao, uniform restriction to either form might in some cases
produce an unnatural feeling.
3. See Lin Suqing ᵫ㋴⏙, “Liang-Han jingming suojian jiyu yanjiu” ܽ⓶䦵䡬᠔㽟
ঢ়䁲ⷨお [Research into auspicious phrases seen in Han-period mirror inscriptions], in
Introduction 3
folksongs, proverbs, and storytelling in many later eras. (In Chinese it is

called qiyan ϗ㿔 ‘seven speakings’, and I will ask the non- Chinese-
speaking reader to adopt it as a technical term.) A banner stretched above
the road’s northbound lane, apparently intended for vehicles headed out
of the city, read:
Gaogao xingxing chu cheng zou 催催݈݈ߎජ䍄

An’an quanquan hui jia lai ᅝᅝܼܼಲᆊᴹ
Have a happy trip leaving the city, and be very safe in coming home.
Here was a couplet that exhibited not only the qiyan rhythm but gram-
matical parallelism and “semantic antitheticality” (i.e., paired opposites in
meaning: chu ‘exit’ versus hui ‘return’ and zou ‘leave’ versus lai ‘come’) of a
kind favored by classical poetry. The message seemed somehow more
formal and exalted than if it had been put in ordinary language.
Formal? Exalted? I crossed the street and saw a public toilet. A sign
warned: Jinzhi suidi daxiaobian ⽕ℶ䱣ഄ໻ᇣ֓ ‘Don’t just relieve yourself
anywhere you like’. Qiyan again. The pattern seemed useful in a variety of
contexts, exalted or not, but in any case seemed to bear a kind of authority.
Its partner wuyan Ѩ㿔, the equally classical 1–2, 1–2–3 syllabic pattern,
was also widely in evidence. A television advertisement for cockroach
killer promised: Zhanglang siguangguang 㶥㵖⅏‫‘ ܝܝ‬Cockroaches dead
to the last one!’ Somehow the poison seemed a bit more lethal in wuyan.
A notice for a childbirth class promised wutong fenmianfa ᮴⮯ߚ࿽⊩
‘pain-free delivery’. Could wuyan mollify even labor pain? No, I thought.
But it was apparent that someone, somewhere had felt that wuyan could
add credibility to a claim about pain reduction.
These uses of classical rhythms were fairly obvious, I thought, and
I guessed that both writers and readers of such phrases might have been
Handai wenxue yu sixiang xueshu yantaohui lunwenji ⓶ҷ᭛ᅌ㟛ᗱᛇᅌ㸧ⷨ㿢᳗䂪᭛䲚

[Collected papers of the Academic Conference on Han Literature and Thought], ed-
ited by Guoli Zhengzhi Daxue Zhongwen Xisuo, 161–188 (Taipei: Wenshizhe Chu-
banshe, 1991). Also “Liang-Han jingming chutan” ܽ⓶䦵䡬߱᥶ [Preliminary in-
quiry into Han-period mirror inscriptions], Academia Sinica, Bulletin of the Institute of
History and Philology, 63, no. 2 (1993): 325–370.
aware of them. But sometimes I noticed wuyan and qiyan at work intuitively,
as it were—buried inside phrases where they certainly made a difference
but perhaps neither writer nor reader was consciously aware of them.
Why, for example, do we fi nd the opening line of Lu Xun’s famous story
“Kong Yiji” so lovely, so mellifluous? Luzhen jiudian de geju, shi he biechu
butong de 元䦂䜦ᑫⱘḐሔ, ᰃ੠߹㰩ϡৠⱘ ‘The layout of the taverns in
Lu-town was different from those in other places’. The sentence is com-
posed of two phrases of seven syllables each—not exactly parallel, to be
sure, yet readable in something close to a 1–2, 1–2, 1–2–3 rhythm. Was Lu
Xun conscious of the pattern when he wrote the line? Probably not, I
guessed, although I think the line must have “felt right” to him even if he
did not stop to examine the reasons. (Certainly I, as one reader of the line,
sensed its rhythmical beauty before I ever thought of counting syllables.)
Near the lugubrious end of Xu Zhenya’s novel Yulihun, the narrator con-
fesses Yu yi shangxinren, xie ci duanchangshi ԭѺ‫ڋ‬ᖗҎ, ᆿℸᮋ㝌৆ ‘I, too,
am a grief-stricken person, writing this heart-breaking tale’.4 The line is
wuyan, and exhibits parallelism, but it is not presented in the text as po-
etry. For Shangzhong hongling xia zhong ou Ϟ⾡㑶㧅ϟ⾡㮩, the title of a
novel by the celebrated contemporary novelist Wang Anyi, in English, we
must settle for something unrhythmical like “planting red water nuts
above and lotuses below,” but the phrase’s structure in Chinese includes
qiyan, parallelism, and even a ping-ze ‘level-oblique’ tonal pattern.5 Wang
Anyi may (or may not?) have been consciously aware of these details when
she created them. But how aware are her readers? Most, I would guess,
feel that the result is pleasant but do not ask why, and are not consciously
aware of the role that rhythm and parallelism are playing.
In musing over such questions, I found some of the most dramatic ex-
amples in the Mao Zedong era. During the Cultural Revolution, Mao
exhorted the Chinese people to “smash the four olds”: old customs, old
culture, old habits, and old ideas. Certainly wuyan and qiyan should count
as “old culture,” yet Red Guards who gathered in Tiananmen Square at
4. Xu Zhenya ᕤᵩѲ, Yulihun ⥝Ṽ儖[ Jade Pear Spirit] (Shanghai: Minquan chu-
banshe, 1914), p. 165.
5. On ping-ze and other technical matters, non-Sinologists may refer to James J. Y.
Liu, The Art of Chinese Poetry (Chicago: University of Chicago Press, 1962), pp. 20–38.
Introduction 5
the time chanted Women yao jian Mao Zhuxi! ៥Ӏ㽕㾕↯Џᐁ ‘We want to
see Chairman Mao!’ in a 1–2, 1–2, 1–2–3 pattern. Were they aware that
they were using an example of the “four olds” in order to praise the lead-
ing opponent of the “four olds”? Was Mao himself aware? It seemed im-
possible. It seemed that no one noticed the irony, even as everyone was
intuitively enjoying the lilt and rightness of the phrases. Then I began to
notice that many phrases from the Cultural Revolution era used qiyan:
Funü neng ding banbiantian ཛཇ㛑乊ञ䖍໽ ‘Women can hold up half the
sky’;6 Linghun shenchu gan geming! ♉儖⏅໘ᑆ䴽ੑ ‘Make revolution in
the depths of your soul!’; Dahai hangxing kao duoshou ໻⍋㟾㸠䴴㠉᠟
‘Sailing on the seas relies on the helmsman’. Wuyan seemed just as com-
mon. Gongye xue Daqing, nongye xue Dazhai ᎹϮᄺ໻ᑚˈ‫ݰ‬Ϯᄺ໻ᆼ ‘In-
dustry should learn from [the] Daqing [oilfields], agriculture should learn
from [the] Dazhai [commune]’ was not only wuyan but parallel. It was in-
teresting to me that these were all carefully crafted political slogans,
things on which the creators must have spent considerable conscious
effort. Have any cultural products—not only in Chinese history but in
world history— ever been more closely scrutinized for political correct-
ness than the model operas of the Cultural Revolution era? But one opera
was titled Hongse niangzi jun 㑶㡆࿬ᄤ‫( ݯ‬Red detachment of women) and
another Zhiqu weihushan ᱎপ࿕㰢ቅ (Taking Tiger Mountain by strat-
egy). Both are 1–2, 1–2–3 wuyan.
Whether or not they were used intentionally, the rhythms seemed to
add something to the phrases they inhabited. What was it? Should we call
it “meaning”? Can rhythms by themselves “mean”? Women yao jian Mao
Zhuxi! somehow sounds more exalted, more righteous, than a plain Women
shi lai kan Zhuxi de ៥‫ץ‬ᰃ՚ⳟЏᐁⱘ would sound. Similarly, Yi kan, er
man, san tongguo ‘First look, then go slowly, then cross’ sounds more for-
mal, more authoritative, than a casual dajia xiaoxin guo jie, a! ໻ᆊᇣᖗ䘢㸫,
䰓! ‘Everybody be careful crossing the street, okay?’ Certainly something
is added by the rhythm, but one problem with calling this added thing
“meaning” is that, as we have just seen, rhythms can be used without
conscious awareness. Red Guards feel that they are praising, and Mao
6. I discuss the misleading translation of this statement as “women hold up half the
sky” in Chapter 3.
Zedong feels praised, but neither side (most likely) is at all aware of think-
ing “We are using this rhythm for this purpose.” Of words, we normally
assume that they need to be used consciously in order to be “meant.” Is
this not true of rhythms? If we say that rhythms do not have “meanings,”
then what is the right word for what they add?
The way rhythms can “mean” seemed to me parallel to the sense in
which certain grammatical constructions— all by themselves, indepen-
dently of vocabulary— can, like rhythms, convey implications. In English,
for example, “I threw her the ball” and “I threw the ball to her” both de-
scribe me as throwing a ball in her direction. But the former implies that
she caught (or otherwise received) the ball, and the latter does not. The
grammar seems primarily responsible for the implication of how success-
ful the effort is. The same effect can be seen in “Jack sent Jill a package”
and “Jack sent a package to Jill”; or “I taught Gladys Chinese” and “I taught
Chinese to Gladys.”7
I also began to wonder if rhythmical patterns were more common in
ordinary Chinese than in ordinary English. Every human language uses
stress, of course, and sometimes stress patterns take on aesthetic qualities
when they “just feel right”—not merely because the accents are all on the
right syllables, but because of something more than that. For example,
most English speakers would probably say that “bright and shiny” sounds
better than “shiny and bright,” because TAH-ta-TAH-ta sounds better
(has more natural balance than?) TAH-ta-ta-TAH.8 The phenomenon
is clear in things like slogans, chants, marches, and advertising jingles.
Among the latter, we should note the pretechnological forebears known
as hawkers’ calls. China was once rich in hawkers’ calls, and a few survive
even in the West, for example in American ballparks, where a rhythmic
pattern like “GET-cha HOT-dogs!” clearly assists in the hot dog vendor’s
delivery of a message. Here, I thought, the rhythm itself helps hearers to
know instantly what the topic is. Rhythms in language must be universal
and part of being human, yet I still felt that they seemed especially com-
7. I am indebted to Adele E. Goldberg for this insight. See her “Constructions: A

New Theoretical Approach to Language,” Trends in Cognitive Sciences 7, no. 5 (May
2003), 219–224.
8. I am indebted to James Richardson for this example (email message to author,
September 18, 2002).
Introduction 7
mon in Chinese. Beijingers could take Yi kan, er man, san tong guo in stride;
what would New Yorkers think of being told in rhythm how to cross the
street? What would they do if exhorted in parallelism to have a nice time
in the country? How would they respond to a sign reminding them, in
pentameter, where not to pee?
The public buses in Beijing, during that autumn of 1988 when I was
working there, had ticket-sellers who would call out the names of bus
stops in rhythmic patterns. One of these patterns was Xia yi zhan, X-X-X,
mei piao mai piao! ϟϔキ X-X-X, ≦⼼䊋⼼ TAH-ta-TAH, TAH-ta-TAH,
ta-TAH-ta-TAH! “Next stop is X-X-X, get a ticket if you don’t have one!”
Many of the bus stop names were three syllables (zhongguancun Ё݇ᴥ
‘middle-gate village’, shuangyushu ঠ὚ᷥ ‘twin elms’, dongwuyuan ࡼ⠽ು
‘zoo’, baishiqiao ⱑ⷇ḹ ‘white-stone bridge’ etc.), which allowed the trisyl-
labic X-X-X almost always to fit. Where a stop had a longer name, like
Nongye Kexueyuan ‫ݰ‬Ϯ⾥ᄺ䰶 ‘Academy of Agricultural Sciences’, it got
a three-syllable abbreviation, Nongkeyuan, and the ticket-seller still
could say Xia yi zhan, Nongkeyuan, mei piao mai piao! Conductors on New
Jersey Transit trains between Princeton and New York City also some-
times enter a car and say, “Tic-KETS!” with a distinctive lilt. But if they
went on for ten syllables, using the rhythm that their counterparts in
Beijing do, it would not work. “Next stop IS, New BrunsWICK, show
your TIX QUICK!” might frighten people off the train. In Beijing, even
the three syllables of the stop names had a distinctive internal grammati-
cal structure. Almost all were “two-syllable modifier plus one-syllable
modified”: nongke + yuan; dongwu + yuan, tianan + men, wangfu + jing,
baishi + qiao, and so on. This structure clearly helped the rhythm. Or was
the rhythm in charge of creating the structure? One should look into this,
I thought.
Yet another aspect of that ticket-seller’s call was interesting. She, like
everyone who speaks Chinese, said “xia” yi zhan ‘one stop “below’ ” for the
“next” stop. Time was going “down.” “Next week” is xia ge xingqi ϟϾ᯳
ᳳ, “next month” xia ge yue ϟϾ᳜, and “next time” xia ci ϟ⃵. Similarly,
“last month” is shang ge yue ϞϾ᳜ ‘the month above’, “last time” is shang
ci Ϟ⃵ ‘the time above’, and so on. When you continue talking about
something, you shuo xia qu 䇈ϟএ ‘go down with your talking’. Do we
ever do this in English? I asked myself. We do say “down through the
ages” and that inheritances get “passed down”— although for the latter
case I wondered if “down” meant “later in time” or “lower in family sta-
tus.” In any case, even if we sometimes do use “down” for “future” in En-
glish, we don’t use it as much as Chinese does. In English the time line
seems horizontal. We say “front” for future: we look forward to things, to
the glorious future that lies before us, and so on. But wait: Chinese does
this, too. After Mao, Deng Xiaoping urged the Chinese people: Xiang
qian kan ৥ࠡⳟ ‘Look forward’— clearly meaning “Look to the future.”
Was this usage a borrowing from Western language? No, because classi-
cal Chinese has qian zhan ࠡⶏ ‘forward outlook’ for looking toward the
future, and a term like qiantian ࠡ໽ ‘the day in front’, meaning the day
before yesterday, is not a borrowing from Western language.
But wait, again: qiantian refers to the past? Didn’t qian ‘front’ refer to the
future in xiang qian kan ‘look forward’? How can you look forward to the
day before yesterday? And why, in order to say “the day after tomorrow,”
does Chinese say houtian ᕠ໽ ‘the behind day’? The “behind day” is in
the future? Is Chinese confused? “Future generations” are houdai ᕠҷ
‘behind generations’. We all hope that future generations will have good
“futures,” or good qiantu ࠡ䗨 ‘forward paths’. So we worry about houdai
de qiantu ᕠҷⱘࠡ䗨, literally “the forward paths of the behind genera-
tion.” Shouldn’t we, perhaps, worry about our logic fi rst?
English, it turns out, is no better. In English we look forward to the fu-
ture, but our forefathers reside in the past. They came before us, and there-
fore can be of no help with the problems that lie before us. Hmmm.
It suddenly seemed strange to me that human beings could talk with
one another in either Chinese or English and remain clear about what was
being said. Obviously they can, however. There must be rules that help
people to keep things straight, even if they are not consciously aware of
those rules.
I ran across Metaphors We Live By by George Lakoff and Mark Johnson,
and it helped to clarify these problems, if not to solve them.9 It was also
my entrée into the field of cognitive linguistics, where the study of meta-
phor has grown considerably since Lakoff and Johnson published their
9. George Lakoff and Mark Johnson, Metaphors We Live By (Chicago: University of

Chicago Press, 1980).
Introduction 9
book in 1980. The book discusses the use of space as a metaphor for time
(i.e., the question of which direction time “goes”) and many other exam-
ples of daily-life metaphor that not only reflects the way we conceive the
world but “structures” the way we are likely to continue conceiving it as
new experience comes along. They point out, for example, that in English
“consciousness is up, and unconsciousness is down.”10 Thus we fall asleep
and wake up. We go under hypnosis or sink into a coma. The structuring
power of this metaphor was evident when psychoanalysis came along and
a subconscious was metaphorically conceived as “below” the conscious level.
I read Lakoff and Johnson asking myself whether the “we” in their
phrase “metaphors we live by” included Chinese speakers. Do different
languages have different structuring metaphors that reflect, perhaps,
partly different worldviews? It was interesting to me that a French trans-
lation of Lakoff and Johnson’s book could use many, but not all, of the
book’s original examples. How different is Chinese? How important are
the differences? In Chinese, for example, we do not use “up” and “down”
for moving into and out of consciousness. When we do use spatial meta-
phors for this purpose, the movement is conceived as crossing a border
within a single horizontal plane. Thus, in fainting we yunguoqu ᱜ䘢এ
‘faint and cross away’, and in awakening we xingguolai 䝦䘢՚ ‘awake and
cross toward here’. This is not terribly different from “pass out” and “come
to” in English, but it was interesting to me that “up” and “down” were not
involved in Chinese, except in modern terms, like xiayishi ϟᛣ䄬 ‘subcon-
scious’, that are clearly borrowings from Western languages. I thought of
Zhuangzi’s famous story of dreaming that he was a butterfly and then
waking to wonder whether he was, indeed, a man who had dreamt of be-
ing a butterfly or was now a butterfly dreaming he was a man. Is the co-
nundrum more puzzling when one thinks of Zhuangzi crossing over a
line within a single plane than if one thinks of him rising and falling be-
tween levels? It felt so to me, but it was hard to say exactly why.
I became curious in the opposite direction as well: might certain West-
ern philosophical conundrums appear differently if approached using Chi-
nese metaphorical apparatus? For example, Lakoff and Johnson stress the
importance of what they call “ontological metaphors” that turn complex
10. Ibid., p. 15.

processes into “entities,” use nouns to label them, and thus make them
easier to talk about. “Inflation,” they say, is not originally an entity, but
if we use an “ontological metaphor” to conceive it as such we can then
“measure inflation,” “combat inflation,” and see inflation as an actor that
“lowers our living standards,” “takes a toll at the checkout counter,” and
so on. All of this seemed right to me, but then Lakoff and Johnson write:
“ontological metaphors like this are necessary for even attempting to deal
rationally with our experiences,”11 and here I had my doubts. Students of
Chinese philosophy have often noted that Chinese thinkers like to talk
about “process” more than “thing,” and, vaguely speaking, I had always
felt that Chinese is by nature more verb-rich while English is more noun-
rich. Thus in English we can find it natural, if not elegant, to utter a
phrase like “the beginning of the development of the process of construc-
tion of bilateral relations” (in 1979 I had to translate this phrase in China
for the vice-chancellor of an American university). The phrase strings
together nouns that in a sense might be viewed as verbs (“begin,” “de-
velop,” “proceed,” “construct,” “relate”) in disguise. If we put the noun-
rich phrase directly into Chinese and say liangbian de guanxi de jianshe de
guocheng de fazhan de kaishi ܽ䙞ⱘ䮰֖ⱘᓎ䀁ⱘ䘢⿟ⱘⱐሩⱘ䭟ྟ, we are
grammatically correct but sound horrid, indeed so horrid that the mean-
ing is not easy for a Chinese speaker to grasp. “Ontological metaphors,” it
seemed to me, just aren’t as common or natural in Chinese as in English,
and whenever they appear in abundance in Chinese, the Chinese takes on
a flavor of “translatese.”
But if that is so, then we need to doubt that ontological metaphors are
all that “necessary” for “even attempting to deal rationally” with life, as
Lakoff and Johnson write. How much a language uses them might vary
quite a lot—with no difference in how “rationally” people can get through
life. Indeed, to turn things around, it might be that Western languages
talk about “entities” rather too much—perhaps thereby creating problems
where there needn’t have been problems, or at least not such tough ones.
Western philosophers have long wrestled with what we mean by terms
like “the good,” “mind,” “reality,” and “existence.” These are nouns, and
we might ask how much of Western puzzlement over them has had to do
11. Ibid., p. 26.

Introduction 11
with trying to figure out what “things” they are. In Chinese it is extremely
awkward to translate “the good” as a noun; “reality” and “existence” as
nouns are marginally more possible, but still are more easily discussed us-
ing verbs or other parts of speech. The Western “mind-body problem”
somehow feels less problematic in Chinese; nouns like xin ᖗ and shen 䑿
are available, to be sure, but their use in grammatical context does not eas-
ily conjure the sort of radical mutual separateness of conceptual category
that preoccupied René Descartes. I began to wonder if Chinese grammar
might help with Western philosophical problems—not by “solving” them
so much as suggesting ways they needn’t be seen as problems in the first
place.
Despite some very interesting contrasts, however, on the whole I found
more similarities than differences in comparing the conceptual meta-
phors of Chinese and English. The puzzles about “before” and “after” as
spatial metaphors for time led to very similar answers for the two lan-
guages. So did other examples, like being “red with anger” in English and
mian hong er chi 䴶㋙㘇䌸 in Chinese. Lakoff and Johnson suggest that the
“experiential basis” of metaphor can often explain such similarity. Since it
is true of human beings generally that intense anger causes blood to rush
to the face, it is probably no accident that a red face should mean anger in
several languages. Even something as basic as “high is more” (high level,
high octane, etc.) can be seen as having the simple experiential basis that,
originally, the more physical objects one puts in a place, the higher a pile
becomes. Other theorists have gone further, claiming that it is not just
common experience but the hardwiring of the human brain that leads to
commonalities in perception. Kant claims this for concepts of space and
time, and Chomsky for fundamental grammatical structures.
I was surprised to learn that contemporary neuroscience has found
there are universal “best examples” of colors like “red” and “yellow.” I had
always assumed that the labeling of hues along the color spectrum was
arbitrary in the sense that each language could do this as it pleased. In
Chinese huang 咗, for example, is not coterminous with “yellow,” because
it spans from yellow all the way through tan to brown. The Huanghe
咘⊇‘Yellow River’is brown. But it turns out that this is only part of the
truth. There are shades of certain colors like red, green, blue, and yellow
that native speakers of different languages will tend to pick out as the best
examples of the range within which they fall. These “focal colors” tend to
be the same, it seems, because of the physiology of the human eye.12 And
so it happens that the two words red and hong ㋙, for example— despite
all of their different cultural and political connotations in English and
Chinese— still lead native speakers of the two languages to identify the
same “best example” on the color spectrum.
I began to wonder if there were parallel ways in which preference for
certain rhythms might also be common to humanity at large. Is it only
chance that rhythmic storytelling in China has often used 4:4 time, just
as Western hymns and chants do? Might that be because all of us humans
walk on two legs? Is it merely coincidence that the ten-syllable rhythm I
heard on that Beijing bus (TAH-ta-TAH, TAH-ta-TAH, ta-TAH-ta-
TAH) happens to be the same as the one I use when I say my ten-digit
telephone number in the United States? This must, of course, be partly
coincidence. Both ticket-sellers and telephone numbers can, and do, em-
ploy other rhythms. But is there, as it were, a certain repertoire of rhythms
that the human brain prefers, and of which this is one? Here is another
example: The Cultural Revolution song “The East Is Red” uses a rhythmic
pattern of 1–2–3, 1–2–3, 1–2–3–4–5–6–7 (Dongfang hong, taiyang sheng,
Zhongguo chu le [ge] Mao Zedong ϰᮍ㑶ˈ໾䰇छˈЁ೑ߎњϾ↯⋑ϰ).13 So
does the Western nursery rhyme “This old man, he plays one, / he plays
knick-knack on my thumb.” So do Chinese riddles. So do a number of
poems by the eighth-century Chinese poets Li Bai and Du Fu. It is hard
to postulate “borrowing” over such a range of cases.
In thinking about whether rhythms might have “meanings” that users
are not fully aware of, and also noticing how established metaphors seem
to “structure thought,” I wondered how such factors might relate to po-
litical uses of language. Some very astute observers of contemporary Chi-
nese language use have raised the issue of how political usages—including,
but not limited to, structural metaphors and rhythmic slogans— serve
12. Paul Kay and Chad McDaniel, “The Linguistic Significance of the Meanings of
Basic Color Terms,” Language 54, no. 3 (September 1978), 610– 646.
13. There are syncopated beats in some of the lines, but this does not change the
basic prosodic structure.
Introduction 13
to shape thought in contemporary Chinese society. The Swedish scholar

Michael Schoenhals, in writing about how to “do things with words in
Chinese politics,”14 unfolds many subtleties in the power engineering of
political language in Communist, and especially Maoist, China. In a
number of essays, the eminent literary critic Li Tuo has wondered whether
Chinese writers, once acculturated to a “Mao literary form,” can extricate
themselves from its worldview— or even become aware that they in fact
remain inside it. Li’s observations in some ways recall eastern Eu ropean
writers like Miklos Haraszti and Czeslaw Milosz. In describing Hungar-
ian writers in the 1950s as living inside a “velvet prison,”15 Haraszti re-
ferred not only to the system of material rewards that lured and held them
but also to the velvet unreality of official language that dulled their intel-
lectual work. In The Captive Mind, Milosz went even further, explaining
how language, gesture, and other aspects of personal presentation created
a kind of public “role playing” that could become so thoroughgoing that
eventually, “after long acquaintance with his role, a man grows into it so
closely that he can no longer differentiate his true self from the self he
simulates.”16
What Milosz says of Communist Poland was, we know from literary
accounts and memoirs, substantially the same in China during high Mao-
ism. In the immediate post-Mao years, when I first lived in China, the all-
consuming power of official language had receded somewhat, but there
was still a clear— and very interesting—bifurcation between official and
unofficial language. The official language was used in newspapers, on the
radio, and at political meetings, and its distinctive features separated it
clearly from everyday talk, which was used for buying fish, scolding chil-
dren, gossiping about one’s sister-in-law, and other such daily-life activities.
Left alone, people preferred ordinary language, but in official contexts they
needed to use officialese. All governments, of course—including premod-
ern Chinese governments—have used officialese, but there are interesting
14. Doing Things with Words in Chinese Politics: Five Studies (Berkeley: Center for
Chinese Studies, 1992).
15. The Velvet Prison: Artists under State Socialism, trans. Katalin and Stephen
Landesman (New York: Basic Books, 1987).
16. The Captive Mind (New York: Knopf, 1953), p. 55.
differences of degree in how and how much it has been used. In Qing
China, the guanhua ᅬ䁅 ‘official talk’ was left to officials. Ordinary
people didn’t need it for daily life. In Mao’s China, politically correct ver-
bal expression was not optional. Even for several years following Mao’s
death, political study meetings were mandatory, and one’s “performance”
had to be correct at them. Moreover, in order to get certain things that
you needed in daily life—like bicycles, job assignments, or permission to
marry— officialese had to be used correctly. The ordinary citizen had to
abandon ordinary talk and play a kind of official “language game” in such
contexts. For example, a professor I knew at Zhongshan University in
Guangzhou, where I was doing research in 1980, wanted a bigger apart-
ment for his family. The government had recently directed that universi-
ties treat professors better, in order to bring them back into the fold after
their severe mistreatment during the Cultural Revolution. My friend
knew about this policy. He went to his Party secretary, but did not ask
“Can I have a bigger apartment?” He asked: “Do you think we can con-
cretize Party Central’s policy on intellectuals?”17
In graduate school I had learned much about different kinds of Chinese
language: classical and vernacular, different forms of ancient language,
and the many very different versions of modern oral Chinese, which are
more nearly different languages than just “dialects,” as they are some-
times called. But now I began to be interested in another axis on which
the Chinese language divides into different versions—the official and the
unofficial. I found some systematic differences. In vocabulary, for exam-
ple, official Chinese used more borrowings from Western languages (pri-
marily the neologisms introduced to China via Japan in the late-Qing
years, with later additions from English, German, and Russian) than
daily-life Chinese did. Terms like xingshi ᔶ࢓ ‘situation’ or dongxiang
ࡼ৥ ‘trend’ had an official flavor. They carried an aura of correctness but
were usually abstract enough that one did not know exactly what they
meant. This made them useful in obfuscating sensitive things. For exam-
ple, to describe the persecution of people during the Cultural Revolution—
17. Neng bu neng luoshi yixia zhongyang de zhishifenzi zhengce? 㛑ϡ㛑㨑ᅲϔϟЁ༂ⱘ

ⶹ䆚 ߚᄤᬓㄪ?
Introduction 15
events for which ordinary language might use lively phrases like hunfei
posan 儖亯儘ᬷ ‘soul flies and spirit scatters—be scared out of one’s wits’
or jiapo renwang ᆊ⸈Ҏѵ ‘home wrecked and person perished’—the offi-
cial language could retreat into a phrase like caiqu cuoshi jinxing zhengdun
᥵প᥾ᮑ䖯㸠ᭈ䷧ ‘adopt measures to carry out reorganization’. Cuoshi
and zhengdun are modern neologisms, and also what Lakoff and Johnson
call “ontological metaphors,” but the “entities” to which they are sup-
posed to refer could hardly be more vague. I was reminded of George
Orwell’s famous essay “Politics and the English Language,” in which he
writes that political language “is designed to make lies sound truthful and
murder respectable, and give an appearance of solidity to pure wind.”18
Another interesting aspect of the official language was its implicit claim
to moral weight. No matter how vague the cognitive meaning of caiqu cuoshi
jinxing zhengdun might be, the implication was clear that it was a “correct”
thing to do. Words that were not originally ethical terms could take on
moral weight. For example, Mao Zedong liked the word zui ᳔ ‘most’, and
he liked to use it in series. In 1940 he called the great Chinese writer
Lu Xun the “most brave, most correct, most firm, most loyal, and most
ardent national hero.”19 In the high Maoism of the Cultural Revolution
years, anything zui in the official language had to be good. Mao Zedong was
women xinzhong zui hong zui hong de hong taiyang ៥ӀᖗЁ᳔㑶᳔㑶ⱘ㑶໾
䰇. An appearance of three zui in a row all but guaranteed that the adjec-
tive20 that followed described a wonderful quality. Thus zui, zui, zui hong
᳔᳔᳔㑶 ‘most, most, most red’ and zui, zui, zui zhengque ᳔᳔᳔ℷ⹂
‘most, most, most correct’ both made sense, and indeed were common, in
the official language; but to say zui, zui, zui fandong ᳔᳔᳔ডࡼ ‘most, most,
most reactionary’—although the phrase works fine grammatically and
18. “Politics and the English Language,” in A Collection of Essays (Garden City, N.Y.:
Doubleday Anchor, 1954), p. 177.
19. Widely quoted in later years, the phrase first appeared in January 1940 in “On
New Democracy.” See Selected Works of Mao Tse-tung, vol. 2 (Peking: Foreign Lan-
guages Press, 1967), p. 372.
20. Terms that I here call “adjectives” are different from adjectives in Western lan-
guages because they can follow subjects directly to form sentences, without any other
verb. Some grammarians call them “stative verbs.”
lexically—would have sounded wrong in the official language. There

would have been something oddly contradictory about it, because zui, zui,
zui primes the hearer for something good, and then fandong contradicts
the priming. On the other hand, certain other phrases, although value-
free in ordinary language, in the official language implied ‘bad’. From the
early 1950s (and perhaps earlier), the phrase jishaoshu ᵕᇥ᭄ ‘tiny minor-
ity’ in official language meant not just “small in number” but (morally or
politically) “wrong” as well. When student demonstrators crowded Tian-
anmen Square in the spring of 1989 and Premier Li Peng charged that
they were being manipulated by a “tiny minority” of troublemakers, he
intensified both the cognitive and evaluative meanings by duplicating the
phrase: a jishaoshu de jishaoshu ‘tiny minority within a tiny minority’ was
causing all the trouble. The phrase could not have been used for some-
thing he favored. If he had wanted to say, for example, that Mao Zedong
had been “a minority within a minority” of great world leaders, he would
have had to say it some other way. To say Mao Zedong shi jishaoshu de jishao-
shu de weida lingxiu ↯⋑ϰᰃᵕᇥ᭄ഄᵕᇥ᭄ⱘӳ໻乚㹪 would sound
contradictory. In using this example in Chinese in lectures to Chinese
audiences, I have found that it sometimes induces laughter.
Rhythms, too, I came to sense, had evaluative components in official
language. To take another example from Cultural Revolution language, it
was almost de rigueur in the late 1960s to refer to the Communist Party
as weida de guangrong de zhengque de gongchandang ӳ໻ⱘ‫ܝ‬㤷ⱘℷ⹂ⱘ݅
ѻ‫‘ ܮ‬the great, glorious, correct Communist Party’. The three modifiers
had to be in exactly that order for the result to sound right, and the
rhythm was also fi xed: ta-TAH-ta, ta-TAH-ta, ta-TAH-ta, TAH TAH
TAH. The rhythm itself was so well associated with greatness, glory, and
correctness that, I felt, even if one were to take the vocabulary out of it, a
shadow of the meaning might still remain. I tested this notion by insert-
ing opposed vocabulary to see what would happen. If we substitute
Kuomintang for Communist Party, for example, and say not that it is
great, glorious, and correct but shady, stingy, and disgusting, we can, in
an identical rhythmic pattern, say heian de xiaoqi de taoyan de guomindang
咥ᱫⱘᇣ⇷ⱘ㿢ঁⱘ೟⇥咼. This phrase, too, immediately elicits chuck-
les from people who know Cultural Revolution language. The message is
contradictory: the denotative component of meaning says “bad” while
Introduction 17
the rhythm says “good.” A similar test can show that even if we restrict
ourselves to positive meanings, the rhythmic contribution to meaning
emerges only in the official language, not the language of ordinary life. If
we say (to be as positive as we can) “my beautiful, peaceful, comfortable
home” and put the words into the same official-language rhythm, we
could get something like meili de anjing de shufu de wo jia 㕢呫ⱘᅝ䴰ⱘ㟦
᳡ⱘ៥ᆊ. But this result is so awkward that, if heard in actual daily life,
the hearer would wonder what is wrong with the speaker.
In any human language, of course, the patterns of set phrases (slogans,
proverbs, idioms) are distinctive, and it is not hard to generate a sense of
incongruity by inserting different vocabulary into them. In finding this
phenomenon in official Chinese, we should not assume that it does not
exist elsewhere. But there are differences of degree, and there are some-
times differences in form as well. It began to interest me that Chinese
slogans, grammatically speaking, are usually presented as subjectless
predicates: dadao sirenbang ᠧ‫צ‬ಯҎᐂ ‘Down with the Gang of Four’;
quanxin quanyi gao sihua ܼᖗܼᛣ᧲ಯ࣪ ‘Give your full heart and mind
to the Four Modernizations’.21 In English, it is much more natural to use
imperatives. And imperatives, grammatically speaking, are complete sen-
tences. They don’t take subjects; you cannot grammatically say “I down with
the Gang of Four” or “You down with the Gang of Four.” But Chinese
slogans, as floating predicates, “invite” a subject and then leave the spot
blank. Grammatically, one certainly could say wo dadao sirenbang ‘I knock
down the Gang of Four’ or ni dadao sirenbang “you . . .” or zanmen dajia
yikuair dadao sirenbang અӀ໻ᆊϔഫ‫ܓ‬ᠧ‫צ‬ಯҎᐂ ‘all of us together . . .’
Leaving the subject blank in Chinese slogans subtly gives a very different
effect from the imperative form of English slogans. In English an implicit
“I” is telling an implicit “you” to do something. Chinese has more the feel
of “would that it be that [some result come about].” It is left unstated who
does the action, or who tells whom to do it; the focus is on the end result.
From here, I began to notice other ways official Chinese had what
might be called a “goal orientation.” This was reflected not only in gram-
mar (as in the slogan example) but in vocabulary as well. The “filler verb”
gao ᧲, for example, was in widespread use by the end of the Mao era, in
21. We can note in passing that these two slogans are wuyan and qiyan.
both official language and daily-life uses. (Mao Zedong favored the word,
and it is used much less in Taiwan and overseas Chinese communities.)
Gao is so flexible in its uses that one would need a fairly long list of En-
glish words to supply its counterparts: do, make, effect, perform, pursue,
mess with, and so on. Fundamentally, it is just a device to transfer our at-
tention to the goal expressed in its direct object. We have seen earlier the
example quanxin quanyi gao sihua, which I translated (using an imperative,
in English) as “Give your full heart and mind to the Four Moderniza-
tions.” To reflect gao, I could have added a word like “pursue” before “the
Four Modernizations,” but it would not have mattered much. The point of
gao is just to say “bring the goal about.” How? Gao makes no comment.
Just get it done. This characteristic of gao has made things convenient for
policy-makers and bureaucrats. To issue a policy to gao shehuizhuyi ᧲⼒Ӯ
ЏН ‘do socialism’, gao huanbao ᧲⦃ֱ ‘do environmental protection’, or
gao shuangbai fangzhen ᧲ঠⱒᮍ䩜 ‘go the double-hundred direction—
give more latitude in literary expression’ has allowed bureaucrats to tell
what a certain result should be without having to take the responsibility
(and therefore the political risk of a “mistake”) for saying exactly how the
result should be pursued. A Maoist slogan said zhua geming, cu shengchan
ᡧ䴽ੑ, ֗⫳ѻ ‘grasp revolution, stimulate production’. Exactly what to
“grasp” and exactly how one might “stimulate” were problems for the slo-
gans’ audience, not its issuers.
In sum, I came to feel that contemporary official Chinese did have a
number of characteristics by which one could distinguish it from daily-
life Chinese—not as different languages, of course, but as two distinct
registers, or idiolects, of Chinese. The two idiolects were held together
not only because they shared a common base but because there was fre-
quent interchange, and sometimes borrowing, between the two. In unof-
ficial contexts people talked, in ordinary language, about how they should
play the chess game of the official language. Conversely, there could be
talk, in the official language (at political meetings, for example), about
how someone had spoken— or more likely, misspoken—in the unofficial
language. There was also seepage of official vocabulary (as well as gram-
mar, rhythms, and metaphors) into daily-life uses. Gao, for example, came
to be used in a wide variety of daily-life contexts: looking for a spouse
Introduction 19
became gao duixiang ᧲ᇍ䈵; cleaning the toilet gao weisheng ᧲ि⫳;
studying physics gao wuli ᧲⠽⧚; hatching a plot gao yinmou ᧲䰈䇟; cook-
ing up a couple of dishes, gao liang ge cai ᧲ϸϾ㦰; and many more.
In the post-Mao years, some people began to worry, similarly to the
way Li Tuo worried about writers, that the thinking of ordinary people
had been shaped by unnoticed infusion of official language into daily-life
language. Mao favored military metaphors, for example. Originally his
movement annihilated the enemy (xiaomie diren ⍜♁ᬠҎ). But later the
focus turned to annihilating errors (xiaomie cuowu ⍜♁䫭䇃), annihilat-
ing revisionism (xiaomie xiuzhengzhuyi ⍜♁ׂℷЏН), annihilating the
rightist tendency to reverse correct verdicts (xiaomie youqing fan’anfeng
⍜♁েؒ㗏Ḝ亢), and so on. Eventually, in daily life, people found them-
selves near the ends of meals urging one another to fi nish off the remain-
ing food by suggesting “annihilation” of it. Xiaomie shengcai ⍜♁࠽㦰 en-
tered the language as “finish off the leftovers.” Chinese people in Taiwan
and overseas do not use this metaphor and can find it startling when they
first hear it. Mainland people who do use the metaphor, and who come to
be aware of it, seem to feel a bit sheepish but not seriously upset. No one
really approaches leftovers with violent intent, after all. But for some, the
worry lingered. If we did not notice this militarism creeping into our
thinking, how much else have we not been noticing? In the summer of
1989, when a group of Chinese dissidents, refugees from the June Fourth
massacre in Beijing, met in Paris to draft a Declaration of the Chinese
Democratic Front, they got into heated debates over language. Demo-
cratic front? Isn’t “front” a military term? Mao used it, but should we?
Everyone wanted to end the dictatorship of the Communist Party, but not
everyone wanted to use terms like tuifan ᥼㗏 ‘overthrow’ or dadao ᠧ‫צ‬
‘knock down’. Two factions formed on the question, and each took itself
to be the more radical. One side said, in effect, we do not compromise; we
are willing to come right out and say “Down with the Communist Party.”
The other side said: you compromise by accepting the Communist Par-
ty’s language; it is we who do not compromise; we say only “End one-
party rule.”
These Chinese were concerned, I realized, with the same kind of ques-
tion that had intrigued me as I wondered about the “meanings” of rhythms
and structural metaphors: how do these aspects of language that we use

with ease every day, and that work very well for us in getting certain
things communicated (even though we remain largely unaware of how
they are doing so) relate to how we think? The chapters that follow ex-
plore three themes—rhythm, metaphor, and political rhetoric—in more
detail.
1
Rhythm
The stress and intonation patterns in speech, which linguists call prosody,
and which I am also calling, less formally, “rhythm,” are universal in hu-
man languages, in fact essential to them. If you hear syllables pronounced
in a manner that aims at uniformity in matters of stress, pause, and pitch,
the utterances will seem to you to be coming perhaps from a computer, or
an imaginary alien. In any case, they will seem “not human.” Under a
musicologist’s rigorous defi nition of “rhythm,” all phrases in spoken lan-
guage necessarily have a rhythm, even if it is perfectly uniform or utterly
random. The examples I discuss in this chapter might best be called “con-
ventional rhythmic patterns”; in calling them “rhythms” I am simply us-
ing shorthand.
In English we speak of syllables within words “receiving accents”— and,
for longer words, which syllables get primary and secondary accents. On
a scale where 1 means most stress and 4 means least, in American English
the word “constitution” is usually said 2–4–1–3. If the stress pattern of
“constitution” strays very far from 2–4–1–3, native speakers of American
English will find that it “sounds funny.” They will get this sense instantly,
whether or not they have any conscious awareness of the pattern.
Stress patterns also play roles in phrases containing two words, three
words, or more. For example, the name “Joey Davis,” in American English,
is usually said 2–4–1–3, using the same stress pattern as “constitution.”1
Again, a native speaker would find a radical diversion from this pattern to
sound extremely odd. Said in a 4–2–3–1 pattern, the name “Joey Davis”
might startle you, even cause you to keep your distance from Joey the
person, and you might literally find it hard to pronounce the name in a
pattern like 1–4–3–2. The ways some stress patterns naturally sound bet-
ter than others can cause us, without being aware that we are doing so, to
choose certain word orders over others. In English, for example, if we
have a one-syllable word and a two-syllable word and want to connect
them with “and,” we find that it sounds better to put the one-syllable word
first. “Salt and pepper” sounds better than “pepper and salt”; “bright and
shiny” sounds better than “shiny and bright.”2 In general, TAH-ta-
TAH-ta is more agreeable than TAH-ta-ta-TAH.
Syllabic “stress” can mean several things. It can mean that a syllable,
compared to others near it, is (1) higher in pitch, (2) longer in duration, or
(3) louder. Human languages, including the regional languages of China,
differ in how “stress” is composed among these three elements. For Man-
darin Chinese, Y. R. Chao has written that stress is “primarily an enlarge-
ment in pitch range and time duration and only secondarily in loudness.”3
In Cantonese, where “there is no neutral tone” (i.e., no syllable on which
there is “completely weak” stress) as there is in Mandarin,4 somewhat dif-
ferent rules apply. In mellifluous Cantonese, pitch plays a bigger role than
it does in Mandarin, so even without neutral tones, there is plenty of room
1. Example is from Mark Liberman, “The Intonation System of English” (Ph.D.

diss., Massachusettts Institute of Technology, 1975), quoted in Matthew Chen, “Met-
rical Structure: Evidence from Chinese Poetry,” Linguistic Inquiry 10, no. 3 (1979), pp.
371–420, p. 413.
2. These examples are from James Richardson, professor of English, Princeton
University. Email message to author, September 18, 2002.
3. Yuen Ren Chao, A Grammar of Spoken Chinese (Berkeley: University of California
Press, 1968), p. 35.
4. Ibid., p. 38.
Rhythm 23
for “rhythm.” In this book, my focus is on Mandarin, but only because I

do not know other Chinese languages intimately enough to feel confident
about examples.
I want to make clear, too, a general methodological point that will run
through this chapter— and to a certain extent through the chapters that
follow as well. It is this: I am not seeking rules that have no exceptions
(although I welcome such rules if I find them); I am interested in patterns
whether or not they have exceptions, and remain interested whether or
not I can explain the exceptions. Take, for example, the pattern according
to which “Joey Davis” is normally said with 2–4–1–3 stress. If someone
mistook Joey Davis for his brother Louie, Joey might say, “No, I’m Joey
Davis,” and then he would probably use a stress pattern of 1–3–4–4. This
pattern would be quite normal; both speaker and listener would fi nd it
so, and indeed would fi nd it odd not to use such a pattern in such a case.
Y. R. Chao offers the example zhima da de shaobing 㡱咏໻ⱘ➦仙, which,
when said with slightly more stress on da, means “hot biscuits on which the
sesame seeds are large” but with slightly more stress on zhi means “sesame-
seed-sized [i.e., tiny] hot biscuits.”5 In simple cases such as these, we obvi-
ously are not dealing with an exception to a rule but with a more complex
picture of how several rules apply. It is possible even to imagine a situation
in which a native speaker of English, in saying “Joey Davis,” might use the
4–2–3–1 stress pattern that I said is so unnatural-sounding as perhaps to
frighten you away from Joey. Let’s imagine that Joey, several hours ago,
exited his house leaving the air conditioner on and all the windows open.
His roommate, returning home, is disgusted to fi nd that Joey has done
this yet again, for the fourth time this week. The roommate hisses sarcas-
tically, “Jo-EY Da-VIS!” in a 4–2–3–1 pattern. This pattern would be
highly unusual, statistically speaking— but normal in context. My point
is that this kind of “exception,” or the much more common “exception” of
1–3–4–4 (for Joey Davis), does not undermine an interest in, or analysis of,
the most common pattern of 2–4–1–3.
Could one, with care, delineate all the rules of stress in all imaginable
contexts so that the rules applied with the same kind of exceptionlessness
that we expect from mathematics? I think so. I doubt there is anything
5. Ibid.
ultimately mystical about stress patterns. But for my purposes in this

book, the relevant question is “which level of analysis is most useful for
showing what I want to show?” Do I want, primarily, to study forests or
trees? For stress patterns, I do not fi nd the leave-no-exceptions pursuit of
apodictic certainty to be the most fruitful one for understanding the gen-
eral cases, which in turn are often the best cases for illustrating how stress
affects communication. So a few exceptions (or complexities, to put it
more precisely) do not bother me. Of course, too many exceptions would
indeed be a problem, because then a generalization itself comes into ques-
tion. For my purposes here, a pattern that holds 80 or 90 percent of the
time will be respectable; even if it holds only 60 percent of the time it
might be worth noting. But the main point I want to leave with the reader
is this: when I discuss a rhythmic “pattern,” I am never claiming that ex-
amples that fall outside it are therefore “unsayable” or inauthentic.
The Prevalence of Rhythmic Patterns in Daily-Life Chinese
Rhythms in oratory and literature can be lengthy, complex, and varie-

gated; they can be creative and therefore not at all “standard” in a sense in
which one might say they are 80 percent or any other percent representa-
tive of anything else. In the next-to-last chapter of Lao She’s novel Camel
Xiangzi, for example, an elderly street vendor of fried cakes philosophizes
at length, intermittently using creative rhythm that contributes a languid
mood as well as a sense of depth to his soliloquy. He begins with some
trochees: Ni xiang duzi hunhao? Դᛇ⤼㞾⏋ད? TAH-ta TAH-ta, TAH-ta,
then Shei bushi name xiang de? 䂄ϡᰃ䙷咑ᛇⱘ TAH-ta-ta TAH-ta
TAAA(elongated)-ta? Then, a moment later: Yige ren neng you shenme beng?
Kanjianguo mazha ba? ϔ‫ן‬Ҏ㛑᳝⫮咑䐺? ⳟ㽟䘢㵲㲅৻?6 TAH-ta-ta,
TAH-ta, ta-ta-TAH. TAH-ta-ta TAH-ta-ta. The rhythm, although not a
standard pattern, oozes from the language and enchants the reader. There
are no formal signs to mark the rhythm, but readers who are native speak-
ers of Chinese, asked to read the sentences aloud, consistently converge on
6. Lao She 㗕㟡, Luotuo Xiangzi 俅侱⼹ᄤ (Hong Kong: Xuelin shudian, n.d.),
pp. 284–285.
Rhythm 25
the same patterns.7 That the rhythms emerge from the syllabic patterns of
the language, and not just from its meanings, is evident when one compares
English translations of the same lines, in which the original rhythms are
entirely lost: “So you think getting along on your own is best, do you? . . .
Who doesn’t think that way? . . . How far can one man hop? Have you ever
seen a grasshopper?”8 The English has rhythms, to be sure, but they are
different from the originals and deliver a somewhat different mood.
This kind of creative and complex rhythm is not the focus of this chap-
ter. (It is a worthy topic, but one cannot do everything.) Here we will fo-
cus on shorter phrases that have relatively standard rhythmic patterns and
are used in a wide variety of contexts within Chinese culture. The pat-
terns are seldom creative, and are repetitive from case to case, but these
facts do not prevent the patterns’ details and uses from becoming ex-
tremely complex and interesting.
We can begin by sketching the wide variety of contexts in which rhyth-
mic phrases appear. Five-character wuyan Ѩ㿔 patterns and seven-
character qiyan ϗ㿔 patterns have been extremely common in poetry and
folksongs for a long time. This point hardly needs further comment, ex-
cept perhaps to note that wuyan and qiyan patterns have persisted in twen-
tieth- and twenty-first-century poetry much more than is commonly sup-
posed. “Modernist” revolutions in the 1920s and 1980s brought free verse,
French symbolism, and Western-flavored “misty” poetry to China, but
these influences were always a matter of vanguard art and academic study.
It remained true that when people wanted to express strong feelings about
partings, deaths, or political events, they continued— overwhelmingly—
to use wuyan and qiyan patterns. In 1976, after Zhou Enlai died, crowds of
mourners gathered at Tiananmen Square seeking to reveal that, down
deep, they preferred Zhou to the Maoists who had survived him. They
brought thousands of poems to the square, nearly all of which were wuyan
or qiyan.9 Fiction writers, as well, have borrowed wuyan and qiyan, especially
when they have wanted to emphasize the depth, solemnity, or splendor of
7. I have run the experiment a few times. You can do it, too.
8. Lao She, Rickshaw, trans. Jean James (Honolulu: University of Hawaii Press,
1979), pp. 228–229. James’s translation is excellent.
9. See Tong Huaizhou ス់਼, Tiananmen Shiwenji ໽ᅝ䭔䀽᭛䲚 (Collection of
Tiananmen poetry) (Beijing: Beijing chubanshe, 1979), 2 vols.
something. In writing of the Nanjing massacre of 1939, Wang Huo subti-

tled the three volumes of his Zhanzheng he ren ᠄⠁੠Ҏ (War and people)
with qiyan phrases.10 Some of the chapter titles use the pattern as well. As if
to underscore the heartfelt sense that the pattern can convey, one chapter
head begins with a sigh: A! Xueyu xingfeng Nanjingcheng 䰓! 㸔䲼㜹乼फҀජ
‘Ah! Bloody rain and rancid wind course through Nanjing city’.
But wuyan and qiyan are not just for elite expression and solemn contexts.
As any Chinese-reading visitor to tourist sites in China can observe, these
patterns are favorites of graffiti artists as well. Phrases such as daoci yiyou,
hexu gui? ࠄℸϔ␌ԩ䷜⅌ ‘visiting here, what need is there to leave?’ are not
hard to find. At the famous Mount Emei in Sichuan in the 1980s, an official
display featured a qiyan poem by the celebrated poet and Communist elder
Guo Moruo, who had extolled the beauty of the view. Scrawled on a post
nearby was the sarcastic comment of someone who apparently thought the
view had not been worth the hike. The graffiti artist took issue with Guo
Moruo’s aesthetic judgment but not with his choice of qiyan form:
Bushi Guo lao chui de xiong ϡᰃ䛁㗕਍ᕫ‫ܛ‬

Na ge jiuzi cai lai you? ા‫ן‬㟙ᄤᠡ՚␌
If our great Mr. Guo had not puffed this place up

What monkey’s uncle would have wanted to come?11
Incidentally, in this book I often use what others sometimes call “free”
translation, especially for items of popular culture and items for which
rhythm and rhyme are important. In doing so, I do not see myself as sac-
rificing fidelity but, on the contrary, as trying to preserve fidelity to a va-
10. I.e., Yue luo wu ti shuang man tian ᳜㨑Р୐䳰⒵໽ [Moon falls, raven cries, frost
fi lls the sky], Shan zai xu wu piao miao jian ቅ೼㰮᮴㓹㓜䯈 [Mountains dimly discern-
ible in the mists], and Feng ye di hua qiu se se ᵿ৊㥏㢅⾟⨳⨳ [Maple leaves and reed
flowers rustling in the fall]. Wang Huo ⥟☿, Zhanzheng he ren ᠄⠁੠Ҏ [War and
people), 3 vols. (Beijing: renmin wenxue chubanshe, 1993).
11. I am indebted for this item to a marginally educated woman from Sichuan who
lives in New York but prefers to remain nameless. Even in 2007, she did not wish to be
known as passing along a criticism of a Communist icon like Guo Moruo. She could
not remember Guo’s poem, only the satiric answer to it.
Rhythm 27
riety of things at once: meaning, rhythm, rhyme, register, and the holistic
life of a phrase. I do not believe that a translation that kills a lively phrase
and pickles its literal meaning has more “fidelity,” overall, than one that
keeps the life at the cost of some literal precision. This is why, for the
couplet above, I allow myself the phrase “monkey’s uncle” as an attempt
to match the spirit of the line in popular American usage even though
there is no monkey, but only a nondescript male relative, in the Sichuan-
ese original. (I do not ask my reader to share my views on this point, but
just to be aware of what they are.)
The “cultural T-shirts” (wenhuashan ᭛࣪㸿) that began to appear in
Chinese cities in the 1990s also make use of standard rhythms. In a piece of
startling satire of plainclothes police, a T-shirt that appeared in 2006 read:
baoan xunluo zhiyuanzhe ֱᅝᎵ䙣ᖫ丬㗙 ‘volunteer security patrolman’.12
Some kinds of popular culture in China have shown a preference for
four-syllable combinations. Menus, for example, easily show the cultural
preference—not requirement, but preference—for four syllables in the
naming of dishes: gongbao jiding ᆂֱ䲲ϕ ‘kung-pao chicken’, mapo doufu
咏ယ䈚㜤 ‘pock-marked lady tofu’, and so on. It doesn’t matter much if the
names are literal (qingzheng liyu ⏙㪌冝儮 ‘steamed carp’) or fanciful (mayi
shangshu 㵲㷏Ϟ‍ ‘ants climbing a tree—pork and vermicelli Sichuan
style’)—in any case four syllables are preferred. Their internal stress pat-
terns can vary somewhat (to my ear gongbao jiding is close to 1–4–3–3 and
mayi shangshu close to 1–4–2–1), but, grammatically and semantically, nearly
every string splits into 2 + 2, not 1 + 3 or 3 + 1. The 2 + 2 pattern suggests
and maintains a principle of balance.
Although poetry, fiction, graffiti, and menus are normally written forms,
their rhythmic features suggest strong connections to spoken language.
Rhythm in oral Chinese is even easier to find than in writing. Marches
and chants, as in other languages, observe obvious rhythms. Chants that
are used for popular purposes in Chinese sometimes take on the patterns
of wuyan or qiyan and other patterns that are associated with poetry and
song. I understand from an eyewitness (or should we say ear-witness?)13
12. Www.signese.com, July 9, 2006. Viewed June 21, 2012.

13. My source is Professor Hu Ch’ang-tu 㚵ᯠᑺ, Teachers College, Columbia
University.
that in the 1940s, basketball cheerleaders at Nankai University in Tianjin

chanted:
Yi, er, san; san, er, yi ϔѠϝ, ϝѠϔ,

Yi, er, san, si, wu, liu, qi ϔѠϝಯѨ݁ϗ
Jia-you, jia-you, duo yong li! ࡴ⊍, ࡴ⊍, ໮⫼࡯!
One, two, three; three, two, one

One, two, three, four, five, six, seven
Step on it, step on it,14 go all out!
The third line here is qiyan, and the first two are a pattern that I call 3–3–7
and will discuss in more detail below. The pattern 3–3–7 occurs in ancient
poetry, and in modern times appears not only in basketball cheers but in
many things, from the Cultural Revolution anthem “The East Is Red” to
things like popular riddles that describe, in the following example, a peanut:15
Ma wuzi, hong zhangzi 咏ሟᄤ, ㋙ᐇᄤ

Li bian zhuzhe (ge) bai pangzi 㺣䙞ԣⴔϾⱑ㚪ᄤ
A room of hemp, a curtain of red,

A little white fatty lies in bed.
It appears as well in nursery rhymes:
Ni pai yi, wo pai yi Դᢡϔ, ៥ᢡϔ

Yige xiaohair kai feiji ϔ‫ן‬ᇣᄽ‫ܓ‬䭟亯″
You pat one, I pat one

One little child fl ies the plane
Ni pai er, wo pai er ԴᢡѠ, ៥ᢡѠ

Liang ge xiaohair diu shoujuanr ܽ‫ן‬ᇣᄽ‫϶ܓ‬᠟㍍(‫)ܓ‬
14. In midwestern American slang, “step on it” means “step on the accelerator,”
which approximates the Chinese jiayou ‘add gas’.
15. In the second line, the ge shares a beat with the preceding zhe. This sharing does
not change the overall rhythmic structure of the line. I have used “lies in bed” instead
of “lives inside” in my translation for the sake of fidelity to the rhyme.
Rhythm 29
You pat two, I pat two

Two little children lose their hankies.
(and further verses)
The historical roots of these rhythmic patterns lie in popular sayings,

songs, and proverbs as well as in elite poetry. It is not clear how much the
seepage of such patterns was “top down,” socially speaking, and how much
was “bottom up.” (Some have argued that popular sayings and folksongs
might have become elite patterns when written down by literati scribes.) In
any case, there is no doubt that wuyan and qiyan are found in many kinds
of nonelite material. Common wisdom on how to predict the weather may
have originated among farmers:
Ri mo yanzhi hong ᮹≦㛁㛖㋙

Wu yu biyou feng ⛵䲼ᖙ᳝乼
When the sky is red at the end of the day

Wind if not rain will be on the way.
Popular sayings often include the “poetic” conventions of parallelism as

well as those of rhythm:
Jian ren shuo ren hua 㽟Ҏ䁾Ҏ䁅

Jian gui shuo gui hua 㽟儐䁾儐䁅
Speaking human language to a person

And the dev il’s tongue to a dev il
Qiyan is as easy to find as wuyan in sayings at the popular level. Examples are
Guilin shanshui jia tianxia Ḗᵫቅ∈⬆໽ϟ ‘the scenery at Guilin beats any in
the world’ and si zhu bupa kaishui tang ⅏䉀ϡᗩ䭟∈➭ ‘dead pigs aren’t afraid
of boiling water’ (a punishment, once it is applied, loses its deterrent effect).
Many contemporary examples can be found in the satiric “popular dit-
ties” (shunkouliu ䷚ষ⑰) that are passed around in Chinese society, orally
and authorlessly, rather as jokes are passed around in Western societies.
This example from the late 1990s protests the plight of elderly state-
enterprise workers:
Qingchun xian gei dang 䴦᯹⤂㌺咼

Lao le mei ren yang 㗕њ≵Ҏ仞
Shuo shi kao ersun 䁾ᰃ䴴‫ܦ‬ᄿ
Ersun xia le gang ‫ܦ‬ᄿϟњያ
I worked my whole life for the Party,

And had nothing at the time I retired;
Now they tell me to live off my kids,
But my kids one by one have been fired.
Qiyan examples of shunkouliu are even more numerous than wuyan exam-
ples. This one looks back on the course of the Communist revolution:
Xinxinkuku sishi nian 䕯䕯㢺㢺ಯकᑈ

Yi zhao hui dao jiefang qian ϔᳱಲࠄ㾷ᬒࠡ
Jiran hui dao jiefang qian ᮶✊ಲࠄ㾷ᬒࠡ
Dang nian geming you wei shui? ⭊ᑈ䴽ੑজ⚎䂄?
For forty-some years, ever more perspiration

And we just circle back to before Liberation
And speaking again of that big revolution,
Who, after all, was it for?
With the rapid spread of handheld telephones in Chinese cities in the first
decade of the twenty-first century, shunkouliu, which began as an oral
medium, turned partly into the written medium of text messaging. The
change did not affect the prevalence of rhythmic patterns.
Traditional popular arts such as storytelling, clapper-tales, drumsing-
ing, and other forms known as quyi ᳆㮱 ‘song art’ are waning in contem-
porary China, but the most adaptive among them, xiangsheng Ⳍ㙆 ‘come-
dians’ dialogues’,16 has remained popular in altered form. In quyi, including
16. The common mistranslation of xiangsheng as “crosstalk” appears to be based on

a misreading of Ⳍ in the fourth tone, where it means ‘looks’ or ‘appearance’, for the
same character in the fi rst tone, where it means ‘mutual’ or ‘each other’. My rendition
of “comedians’ dialogues” avoids this pitfall, but has the flaw that while the great ma-
jority of xiangsheng pieces are performed by two people, some are done by one, three,
four, or even more performers.
Rhythm 31
xiangsheng, rhythms of many kinds are consciously woven into the textures
of language. An iconic xiangsheng piece called Xiju yu fangyan ᠆࡛㟛ᮍ㿔
(Drama and dialects),17 made famous by Hou Baolin and Guo Qiru, opens
with a 4:4 beat:
Zuo ge xiangshengr yanyuan, ‫Ⳍןخ‬㙆‫ܦ‬ⓨવ

Ke burongyi; ৃϡᆍᯧ
qima de tiaojianr, 䍋⺐ⱘṱӊ‫ܦ‬
dei hui shuo hua. ᕫ᳗䂀䁅
Performing xiangsheng is no easy matter;

At the very least, you have to know how to talk.
It uses qiyan to say that their adorable popular dialect of Beijing is:
luoli luosuo yida dui ಝ䞠ಝ૚ϔ໻ේ
a big pile of loquaciousness
But returns to 4:4 in saying that “refi ned” Beijing dialect is
duan xiao jing han (er) ⷁᇣ㊒ᙡ (㗠)

luojixing qiang 䙣䔃ᗻᔋ
terse, forceful, (and) powerful in logic
Elsewhere, 4:4 is prefaced by a heavy beat at the beginnings of two paral-

lel lines:
Zher wumen yi xiang 䗭‫ܦ‬ሟ䭔ϔ䷓

Nar fajue yi wen . . . 䙷‫ⱐܦ‬㾎ϔଣ
17. Hou Baolin փᅱᵫ and Guo Qiru 䛁ਃ‫ۦ‬, “Xiju yu fangyan” ៣࠻㟛ᮍ㿔 , in
Wang Wenzhang ⥟᭛ゴ, ed., Hou Baolin biaoyan xiangsheng jingpinxuan փᅱᵫ㸼
ⓨⳌໄ㊒ક䗝 (A selection of Hou Baolin’s xiangsheng per for mance pieces) (Beijing:
Wenhua yishu chubanshe, 2003), pp. 35–49.
Here, the door to the room makes a sound;

There, somebody notices and asks . . .
The couplet appears, periodically in the structure of the piece, five times.
The Maoist pressure to remake Chinese literature and art, and to get
rid of “old” things, brought much change to xiangsheng but hardly got rid
of its rhythms. Ma Ji’s piece called Baigujing xianxingjiⱑ偼㊒⦄ᔶ䆄
‘White-boned demon revealed’ complains that under the Gang of Four,
whatever you did could be called wrong, even to the extent that (in wuyan):
Shang ye shang buqu ϞгϞϡএ

Xia ye xia bulai ϟгϟϡ՚
Huo ye huo buliao ⌏г⌏ϡњ
Si ye si bucheng ⅏г⅏ϡ៤
Go up?—You can’t go up
Down?—You can’t go down
Live, you can’t live
Die, you can’t die!
This piece is said to have been created underground during the fi nal years
of Mao Zedong’s rule and then, after his death, appropriated for official
use when the new leadership found it useful in discrediting the “all-evil
[Maoist] Gang of Four.” Whether or not these claims are true for this par-
ticular piece and this particular rhythm, it is easily demonstrable, in gen-
eral, that Chinese governments have adopted popular rhythms in their
“propaganda work.” The phenomenon was especially obvious during the
Mao era, but it did not begin then. The Nationalists used rhythmic slo-
gans in the 1930s and 1940s, and continued to use them in Taiwan after
1949. A public billboard in Taipei in the 1960s read, in wuyan:18
renren zuo hao ren ҎҎ‫خ‬དҎ

riri zuo hao shi ᮹᮹‫خ‬དџ
18. I am indebted to Andrew Plaks, professor emeritus of East Asian Studies,

Princeton University, for this example, October 4, 2006.
Rhythm 33
Everyone be a good person.

Every day do good deeds.
Mao Zedong is said to have favored rhythmic phrases in his personal

speech. Whether or not this is so, it is certainly true that he used them in
political slogans at many points in his life. On August 5, 1966, Mao issued
a “big-character poster” that was later viewed as a turning point in bring-
ing on the violent phase of the Great Proletarian Cultural Revolution.
The poster’s title was a wuyan phrase followed by a qiyan phrase: paoda
silingbu: wode yige dazibao ⚂ᠧৌҸ䚼: ៥ⱘϔϾ໻ᄫ᡹ ‘smash the head-
quarters: a big-character poster of mine’. Then, in calling on young peo-
ple to support him with their own big-character posters, Mao chose qiyan
again: daming, dafang, dazibao! ໻勈໻ᬒ໻ᄫ᡹ ‘big outcry, big release,
big-character posters’. I have cited other examples of wuyan and qiyan in
Maoist language in the Introduction, so will not list more here, but I will
speculate that Mao’s preference for these patterns may help to account for
why they became so prevalent in political struggle throughout Chinese
society during the late Mao years. To cite just one of a great number of
examples, in the 1957 Anti-Rightist Movement, He Chi, a gifted writer of
xiangsheng pieces, was attacked for “hating socialism” and “organizing an
anti-Party clique.” During a “struggle session” organized by local officials
to humiliate him, he heard:19
Dadao He Chi dayoupai ᠧ‫צ‬ԩ䙆໻ে⌒

Mandi dagun shua wulai ⓓഄᠧⓒ㗡⛵䋈
Knock down the big rightist He Chi

Writhing around on the ground in his lies
To what extent have Chinese officials been conscious of the ways they use
rhythms? This is hard to say, and almost certainly varies from case to
case. When officials during the Cultural Revolution used old rhythms in
order to attack “old culture” (which certainly would include old rhythms),
19. He Chi, He Chi zizhuan (Autobiography of He Chi) (Beijing: Zhongguo minjian

wenyi chubanshe, 1989), pp. 287– 88.
they likely were not very conscious of the self-contradictory nature of their
activity. (In the political context of the time, it would have been frighten-
ing to arrive at such consciousness.) Yet they must have been aware of the
rhythms at some level, because the rhythms in their phrases must have
been part of what made them feel that the phrases were fitting.
There is good evidence that at least some people in the government
have been highly conscious of rhythm and have fashioned it consciously
in propaganda work. A man named Peng Ruigao, of the Shanghai Munici-
pal Propaganda Bureau, told New York Times reporter Craig Smith in
2001 that “slogans require the writing techniques and rhythms of classical
poetry to make them palatable to the people.”20 Putting rhythms into
slogans was part of Peng’s job. Signs of conscious word craft are apparent,
as well, at the very highest levels of Chinese government, for example in
the “Eight Honors and Eight Shames” (Barong Bachi ܿ㤷ܿᘹ) of the Hu
Jintao regime. These eight couplets, memorized by schoolchildren across
China, asked people to regard “loving the country (or state)” as an honor
and “hurting the country (or state)” as a shame; to regard “resolute strug-
gle” as an honor and “wallowing in luxury” as a shame; and so on. Each
line is seven syllables long (although the rhythmic pattern is not the same
as in qiyan), and the evidence of conscious craft is overwhelming. The first
syllable of each line is yi ҹ, and the sixth is wei Ў ‘take [something] as
[something else]’; the seventh syllable in each line is either rong ᾂ ‘honor’
or chi ᘹ ‘shame’ in alternating lines; and the four syllables from number
two to number five are all balanced 2 + 2 phrases, the first five of which
are a two-syllable transitive verb followed by a two-syllable object, and
the last eleven of which are parallel two-syllable verbs. There can be no
doubt that the ideological content of the Eight Honors and Eight Shames
was carefully debated at high levels; there also can be little doubt that
their form was consciously crafted as well.
Government slogans on other topics often made use of wuyan and qiyan
patterns as well. Rhyme and parallelism could be involved, as the follow-
ing examples from the years 2006 and 2007 make clear.21 Central Chinese
Television (CCTV) sometimes exhorted youth to idealism with the wuyan
20. Craig S. Smith, “Shanghai Journal; Political Power Grows from the Point of
His Pen,” New York Times, June 14, 2001.
21. I am grateful to Mao Sheng for most of these examples.
Rhythm 35
phrase rexin xian shehui, zhenqing ai renjian ⛁ᖗ⤂⼒Ӯ, ⳳᚙ⠅Ҏ䯈 ‘dedi-

cate yourself to society with ardor, love humanity sincerely’. Warnings to
drive safely, written on public walls and blackboards in several towns and
county seats around the country, read shouwo fangxiangpan, shike xiang
anquan ᠟ᦵᮍ৥Ⲹ, ᯊࠏᛇᅝܼ ‘grip the steering wheel, always think of
safety’ and siji yibei jiu, qinren liang hang lei ৌᴎϔᵃ䜦, 㽾Ҏϸ㸠⎮ ‘one
cup of wine for the driver, two streams of tears for the family’. Walls and
blackboards also carried calls to protect the environment. In wuyan, there
was diqiu shi wo jia, lühua kao dajia ഄ⧗ᰃ៥ᆊ, 㓓࣪䴴໻ᆊ ‘the globe is
our home, greenification depends on everyone’, and in qiyan, there was
yihua yicao jie shengming, yizhi yiye zong guanqing ϔ㢅ϔ㤝ⱚ⫳ੑ, ϔᵱϔ৊
㨝ᘏ݇ᚙ ‘every flower and blade of grass is life, care for every branch and
leaf’. Family planning examples included the wuyan example jiating zinü
duo, xiaokang hui huapo ᆊᒁᄤཇ໮, ᇣᒋӮ⒥വ ‘with many children in a
family, middle-class lifestyle declines’ and zinü zhiliang gao, shenghuo shuip-
ing gao, xingfu zhishu gao ᄤཇ䋼䞣催, ⫳⌏∈ᑇ催, ᑌ⽣ᣛ᭄催 ‘quality of
children high, standard of living high, happiness index high’. In qiyan there
was chusheng quexian ganyu hao, bang ni sheng ge hao baobao ߎ⫳㔎䱋ᑆ䷤ད,
ᐂԴ⫳Ͼདᅱᅱ ‘it’s best to intervene on birth defects; it’ll help you pro-
duce a good little treasure’ and sheng nan sheng nü yiyang hao, nüer yeshi
chuanhouren ⫳⬋⫳ཇϔḋད, ཇ‫ܓ‬гᰃӴৢҎ ‘having boys or girls is equally
good; girls keep the family line going, too’. 22
It should be no surprise that commercial advertisements have also em-
ployed wuyan and qiyan rhythms. In 2008 Nescafé coffee was advertised
under the slogan weidao haojile ੇ䘧དᵕњ ‘the taste is excellent’. In En-
glish this might seem dull, perhaps even self-defeating as an advertise-
ment, but in Chinese, thanks in part to its wuyan structure, the phrase is
far more effective.23 A newspaper advertisement for a restaurant in Shao-
xing, Zhejiang Province, called Donghu Leyuan (East Lake Paradise) ad-
vertised in 2006: jiewen jiujia hechu hao, donghu leyuan jiujia you! ‫׳‬ଣ䜦ᆊ
ԩ㰩ད, ᵅ␪ῖ೦䜦ᆊ‫‘ ۾‬May I ask where the good restaurants are? The
22. From “Jihua shengyu xin biaoyu chutai neiqing” 䅵ߦ⫳㚆ᮄᷛ䇁ߎৄ‫ݙ‬ᚙ (The
inside story on the promulgation of the new birth-control slogans˅, Nanfang zhoumo
फᮍ਼᳿ (Southern Weekend), August 23, 2007.
23. I am indebted to David Moser for this example. Email message to author, April
13, 2008.
East Lake Paradise restaurant is fine!’24 Even McDonald’s was using qiyan
on CCTV: shike changxiang maidanglao ᯊࠏ⬙ᛇ呺ᔧࢇ ‘always keep Mc-
Donald’s on your mind’. Qiyan could help sell tonics: xueqi chongzu cai jian-
kang, buxue renzhun jiuzhitang 㸔⇨‫ܙ‬䎇ᠡ‫ع‬ᒋ, 㸹㸔䅸‫ޚ‬б㡱ූ ‘health
requires that the blood-spirit be ample, and to bolster the blood you need
Jiuzhitang’; also jinnian guojie bu shou li, shou li zhi shou naobaijin Ҟᑈ䖛㡖
ϡᬊ⼐, ᬊ⼐াᬊ㛥ⱑ䞥 ‘don’t take gifts at New Year’s this year unless the
gift is Naobaijin’. If health fails, at least one can look good with cosmetics,
now offered in wuyan: xiangyao pifuhao, zaowan yao dabao ᛇ㽕Ⲃ㙸ད, ᮽᰮ
㽕໻ᅱ ‘if you want good skin, sooner or later you’ll need Great Treasure
Cream’. You’re obese? A guaranteed diet is offered in wuyan: jianfei bu
fantan, fantan bu shoufei ‫ޣ‬㙹ϡডᔍ, ডᔍϡᬊ䌍 ‘cut the fat and keep it
off, if the fat comes back there’ll be no charge’. An advertisement for sani-
tary pads on CCTV used the phrase nüren yue zuo yue kuaile ཇҎ䍞‫خ‬䍞
ᖿФ, making use not only of qiyan rhythm but a pun on yue, which could be
either 䍞 ‘more’ or ᳜ ‘month’, so that users could be either “happy every
month” or “more and more happy,” depending on how one took the phrase.
A poster advertising English lessons makes its appeal in a very Chinese
rhythm: Yingyu xuexi xintupo 㣅䇁ᄺдᮄさ⸈ ‘A new breakthrough in the
study of English’.25 Even the sale of gasoline warrants rhythm: jiayou zhong
shi hua, fangxin pao tianxia ࡴ⊍Ё⷇࣪, ᬒᖗ䎥໽ϟ ‘fill up with China Pet-
rol and go wherever in the world you want’.26 An email provider, in touting
the freedom of expression that its ser vices made possible, came up with
the qiyan phrase wode dipan wo zuo zhu ៥ⱘഄⲸ៥‫خ‬Џ ‘I am master in my
own domain’.
From an historical point of view, the widespread use of rhythm in com-
mercial advertising might be entirely expected, since its roots can be
found in the hawkers’ calls that lasted into the early twentieth century
and that almost invariably used distinctive rhythms. Hawkers’ calls in-
cluded an immense variety, perhaps because each hawker wanted to sound
distinctive. Refuse collectors who called out a phrase like shou polanr de!
ᬊ⸈⟯‫‘ ⱘܓ‬taking in junk!’ might truncate the po to make the effect shou
24. Zhejiang ribao ⌭∳᮹ฅ (Zhejiang Daily), March 12, 2006.

25. Available at www.signese.com, July 13, 2006. Viewed June 21, 2102.
26. All of the examples in this paragraph, unless otherwise noted, are from CCTV
in 2006 or 2007.
Rhythm 37
p’lanr de ta-ta-TAH-ta. It is extremely rare in Mandarin Chinese for the

first syllable of a compound to be unstressed (pronounced in “neutral
tone”), but that is what a refuse collector could do with polan, in order to
help everybody know exactly who was heading down the street.
In all such examples, it seems to me, the rhythm adds something to the
message. It makes the message seem not just more pleasant and memora-
ble but also somehow more natural, more right, more authoritative, more
exalted— or something like that. I will discuss the possible meanings of
rhythm below, as well as questions of which rhythms seem particularly
“Chinese,” and why. But first I want to address a prior question Chinese
linguists sometimes raise: “Why are rhythms so common in everyday
Chinese?” Is there something unusual about the nature of the language?
Is Rhythm Unusually Common in Chinese?
Rhythmic patterns in Chinese are sometimes determined by grammar.

Consider, for example, verb-object constructions of two syllables plus two
syllables (which I abbreviate as “2 + 2”). Phrases like yuedu baozhi 䯙䇏᡹㒌
‘read newspapers’ can regularly be shortened to 1 + 2 du baozhi or to 1 + 1
du bao but not, normally, to 2 + 1 yuedu bao. But 2 + 2 phrases in which the
first two syllables modify the second two normally follow a different rule:
shoubiao gongchang ᠟㸼Ꮉॖ ‘wristwatch factory’ can shorten to 2 + 1 shou-
biao chang or to 1 + 1 biao chang, but not to 1 + 2 biao gongchang. In all of the
above cases, moreover, the syllables come in neat packs of 2, 3, or 4. The
counterpart phrases in English include four and five syllables and cannot
be manipulated in regular patterns of the kinds Chinese allows. Noting
such examples, Lu Bingfu and Duanmu San conclude that “we are not
aware of any other language in which syntax and word length show such a
striking relation.”27 This has to do, they show, with the distinctive rhyth-
mic flexibility of Chinese.
In Chinese, morphemes, or minimal grammatical units, are almost al-
ways monosyllables. They combine easily with one another to produce
standard word lengths, and often produce balance as well. For example,
27. Bingfu Lu and San Duanmu, “Rhythm and Syntax in Chinese: A Case Study,”
Journal of the Chinese Language Teachers Association 37, no. 2 (May 2002), p. 134.
daxiang ໻䈵 ‘big-elephant’ is “elephant” and yachi ⠭啓 ‘tooth-tooth’ is

“tooth.” “Ivory” borrows one morpheme from each to make xiangya, in
two syllables, while “ivory chopsticks” can be xiangya kuaizi 䈵⠭ㅋᄤ or
xiangyakuai, three syllables or four, as one prefers.28 This kind of recom-
binatory power is easy to find in other languages, but Lu and Duanmu are
probably right that the neatness of syllabic structure is distinctive to Chi-
nese. Compare, for example, “electrocardiogram” and xindiantu ᖗ⬉೒
‘heart-electric-chart’. Although electro and dian, cardio and xin, and gram
and tu each can combine with other morphemes to make other words, in
Chinese, but not English, the morphemes that move around are reliably
one syllable (not two, three, or more). This makes creation of rhythm
much easier. A Chinese poet (or other phrase-maker) can quite easily pro-
duce lines of five syllables each; a French or English poet, attempting
comparable craft, might produce five beats per line, but to try to line up
rows of exactly five syllables apiece turns phrase-making into an arcane
exercise in which technical demands eclipse other aspects of art or mean-
ing. Written form only reinforces the difference, because, in Chinese, five
syllables are reliably five Chinese characters, which standardly yield lines
of exactly the same length, since each character is conceptually square. 29 A
French or English poet who tried to produce lines of exactly equal length
would feel almost impossibly constricted, and achievement of the goal, if
someone could do it, would likely appear to the reader as an oddity, a tech-
nical tour de force whose form, not content, would dominate attention.30
28. I am indebted to David Moser for this example. Email message to author, May
8, 2004.
29. Even a character that is not square, like ϔ ‘one’, is usually “conceptually square”
because it occupies the center of an imaginary square space on a printed page. See
Martin J. Heijdra, “Typology and the East Asian Book: The Evolution of the Grid,” in
Perry Link, ed., The Scholar’s Mind: Essays in Honor of Frederick W. Mote (Hong Kong:
Chinese University Press, 2009), pp. 120–125.
30. In an effort to illustrate this problem, David Bellos, professor of comparative
literature at Princeton University, has translated the shunkouliu above (“For forty-
some years, ever more perspiration / And we just circle back to before Liberation,”
etc.) in four lines of exactly twenty-one bits apiece: “Blood sweat and tears / Over forty
long years / Now it’s utterly over / Who stole the clover?” See David Bellos, Is That a
Fish in Your Ear? Translation and the Meaning of Everything (London: Par tic u lar Books,
2011), p. 135.
Rhythm 39
In Chinese, with its monosyllabic morphemes, lines of equal length are

much easier to produce. The phrase-maker has much more flexibility to
arrange content within a set form. A phrase can fall into a formal pattern
without seeming in any way odd, and there is no sense that the formal
requirements of a phrase are attracting too much attention to themselves.
Indeed there is something of the opposite sense: that the form of the
phrase actually adds to its depth or credibility. Something that in English
might seem “too cute” can seem perfectly natural in Chinese. And be-
cause this sort of naturalness has been around for a long time, Chinese
culture tends to value it. Consider the following mundane example. A sign
in the men’s toilet at CCTV in the early 2000s reminded users to flush
the toilet even if they were in a hurry. It read congcong er lai, chongchong er
qu ࣚࣚ㗠ᴹ, ‫ކކ‬㗠এ.31 The phrase has rhythm, rhyme, grammatical
parallelism, and antithetical semantic parallelism (lai ‘come’ versus qu ‘go’).
In Chinese it feels natural, even cheerful. But what happens if one tries to
put it into English? The meaning is “flush the toilet even though you’re
busy,” but how would one imitate the form? “Come in a flash, go with a
flush” captures the four-syllable structure, some of the parallelism, and
some of the rhyme. The lines are of equal length, but maybe a bit too long.
“Enter rushing, exit flushing” is terser, and also has the right number of
syllables and equal line length, and some rhyme. But note: we are now
playing a game in English, and in that game any effort seems awkward if
it is bad and too cute if it is good.
Chinese people have noticed the remarkable flexibility of their lan-
guage and sometimes make up games and puzzles that depend on the
monosyllabicity of morphemes. The recombinatory cleverness that re-
sults goes well beyond anything one could even contemplate in a West-
ern language. On the lid of a teapot, for example, the characters keyi
qingxin ye ৃҹ⏙ᖗг can be inscribed in a circle so that, wherever one
chooses to begin, one can read around the circle and the result makes
sense:32
31. I am indebted to David Moser for the example and for the “come in a flash, go
with a flush” translation cited below. Email message to author, May 8, 2004.
32. I am indebted to Hu Ping for this example. Email message to author, June 9,
2007.
ৃҹ⏙ᖗг You can clear the mind

ҹ⏙ᖗгৃ You can also take it as mind-clearing
⏙ᖗгৃҹ Clearing the mind is also possible
ᖗгৃҹ⏙ The mind, too, can be cleared
гৃҹ⏙ᖗ You can clear the mind, too
But if morphemes are almost always monosyllabic in Chinese, it is quite

another matter, and grossly incorrect, to say that words are. The often-
heard statement that Chinese is a “monosyllabic language” makes sense
only if one is speaking of morphemes. The word word is not always easy to
apply to Chinese,33 but whatever the defi nition, there are a great number
of unproblematic examples of polysyllabic words. Many of these, too, have
reliable stress patterns. In modern Mandarin doufu 䈚㜤 is ‘bean curd’. In
writing the two characters, a person might conceive 䈚 and 㜤 as “mono-
syllables.” But for most people most of the time (including buyers, sellers,
and eaters of bean curd who may have been illiterate), doufu was a two-
syllable word. Moreover it had a clear internal stress pattern: always
DOUfu, never douFU. Examples of other two-syllable words or phrases
are innumerable. Rhythmic patterns become more complex, as we shall
see, when three or more syllables are involved.
Speakers’ Awareness of Rhythm
If someone were to go to the market and ask for douFU, he or she would
probably get the stuff, but everyone within earshot would know “some-
thing’s wrong.” The ability to recognize that “something’s wrong” in
cases like this extends well beyond the small group of people who can
explain—in terms of syllables, stresses, and so on—why it is wrong. This
simple example shows that there are two levels in what we can mean by
“awareness” of stresses or rhythms. When a person hears or uses a stress
pattern that follows convention, it “sounds right” or “feels good.” When
used incorrectly, it sounds wrong. Should we say that the user is conscious
of this connection between rhythm and rightness? The question is subtle.
33. See Y. R. Chao, Grammar of Spoken Chinese, p. 136.

Rhythm 41
We cannot say that users are not consciously aware, because the sense of
rightness (or of wrongness, for mistakes) is indubitably a conscious feel-
ing. But it is also true that the great majority of users of stress patterns do
not make conscious decisions when selecting patterns and cannot explain
what they have done once the sounds have been pronounced.
The “tones” of Chinese are a good example of this phenomenon. If, fol-
lowing Y. R. Chao, we regard voice pitch as part of what we mean by
“stress,” then we can take Chinese tones as miniexamples of stress pat-
terns. A native speaker of Chinese—including any dialect—will immedi-
ately feel that something’s wrong when a tone is wrong. This feeling will
be just as clear as it is for an English speaker who hears “gat” when he or
she can infer that “get” is what the speaker meant to say. But here’s the
point: native speakers of Chinese, unless they are specially trained, usu-
ally cannot identify which tone was said and which was meant. Many will
have learned labels for the tones in school (first tone, second tone, etc.) but
still will not be able to say “that was second tone, it should have been
fourth,” or even “your voice pitch was rising and it should have been fall-
ing.” At that level, they normally are not conscious of pitch. But the same
people will be immediately aware of any mistake, and can very easily pro-
nounce what the proper sound should have been.
This two-leveledness in the awareness of pitch and stress can create
ironies. I noted in the Introduction how Red Guards, in the late 1960s,
assembled in Tiananmen Square chanting an “old” qiyan rhythm (women
yao jian Mao zhuxi) in praise of the man who was urging them to reject
everything old. I doubt that the Red Guards were aware of this
contradiction.
What about Mao himself? At which level, if any, might he have been
aware of such things? There is ample evidence that Mao was willing, in
general, to exempt himself from rules that he applied to others. In 1963,
for example, he published a volume of his classical poetry, in qiyan and
other forms, Mao Zhuxi shici sanshiqi shou ↯Џᐁ䀽䀲ϝकϗ佪 (Thirty-
seven poems by Chairman Mao). By 1974 the little book had been re-
printed fourteen times.34 For nearly everyone else in China during those
34. Mao Zedong, Mao Zhuxi shici sanshiqi shou ↯Џᐁ䀽䀲ϝकϗ佪 [Thirty-seven
poems by Chairman Mao] (Beijing: Wenwu chubanshe, 1974).
Cultural Revolution years, any public show of favor toward classical po-
etry might be dangerous, possibly even lethal. Mao’s book not only con-
tained classical poetry; it imitated traditional, a.k.a. “feudal,” style in
many other ways: it used traditional characters, arranged them vertically
in unpunctuated lines, opened from the right-hand side, and was com-
posed of pages bound with string between covers of blue paper. In many
ways it was made to resemble books of the Ming- Qing era. Its design fea-
tures were unavailable to any other mainland Chinese who produced
books during the late-Mao era, a time when many traditional-style books,
including genuine Ming-Qing volumes, were being burned. Can we imag-
ine that the whole production of Mao’s poetry book happened without the
conscious awareness of its makers of what they were doing? Hardly. Can
we suppose that Mao himself did not consciously approve? Again no. This
is almost impossible to imagine. It is much easier to imagine that the prin-
ciple “Mao is an exception to Maoist guidelines” was in use.
Yet, despite the obvious exceptionalism, I would guess that Mao sel-
dom, if ever, was consciously aware of Red Guard use of qiyan rhythm to
praise him. It seems more plausible that he just felt—rather as the user of
correct tones in ordinary Chinese just feels—that the phrases sounded
right. This hypothesis gains strength when we consider the many exam-
ples of his use of such rhythms in watchwords that he intended for others.
I noted several of these in the Introduction (Funü neng ding banbiantian
ཛཇ㛑乊ञ䖍໽ ‘Women can hold up half the sky’; Linghun shenchu gan
geming! ♉儖⏅໘ᑆ䴽ੑ ‘Make revolution in the depths of your soul’;
Nongye xue Dazhai ‫ݰ‬Ϯᄺ໻ᆼ ‘Agriculture should learn from Dazhai’;
etc.). When Mao wanted to warn that there are a lot of good-for-nothings
hidden in the shallow waters at the august Peking University, his phrase
Beida shui qian wangba duo ࣫໻∈⌙⥟ܿ໮ ‘The water is shallow and the
turtles are many at Peking University’ no doubt just felt better to him
because of its qiyan rhythm.
Qiyan and wuyan rhythms are in fact easier—not, as one might expect,
harder—to find in the public political language of high Maoism. These
rhythmic habits not only appeared in national campaign slogans and the
names of the model operas but seeped into daily-life political struggle as
well. At the two sides of a makeshift stage used for “struggling class ene-
mies” in Harbin, two vertical banners bore the wuyan matching couplet
Rhythm 43
buwang jiejiku, laoji xueleichou ϡᖬ䰊㑻㢺, ⠶䆄㸔⊾қ ‘don’t forget class

bitterness; remember forever blood-and-tears enmity’.35 In Sichuan an art
student remembers a chant, in qiyan, of yao zuo, yao zuo, hai yao zuo 㽕Ꮊ,
㽕Ꮊ, 䖬㽕Ꮊ ‘left, left, ever more left’.36 These are but random examples
in an ocean of others.
The irony of articulate intent versus inadvertent use of rhythm is sharp-
est for the Mao years but hardly limited to them. Fifty years earlier Hu
Shi, in his famous 1917 essay “My Humble Opinion on the Reform of
Literature” called on Chinese writers to “pay no attention to parallelism”
(bujiang duizhang ϡ䃯ᇡҫ).37 But Hu himself, in promoting the scientific
method,38 called for these actions:
Dadan de jiashe ໻㞑ഄ‫؛‬䀁 Hypothesize boldly.

Xiaoxin de qiuzheng ᇣᖗഄ∖䄝 Seek evidence carefully.
Hu’s phrases are not only wuyan but elegantly parallel (“big gall this, little
heart that”). We can imagine that Hu might explain this gap between his
theory and his practice by holding that his rule about parallelism applies
to literary essays, whereas aphorisms of the kind I have quoted here serve
a different purpose. Still, it is interesting that Hu seems to have felt, at
least for aphorisms, that wuyan rhythm and parallelism add something, or
somehow make things better. When his friend and fellow reformer Chen
Duxiu followed Hu’s “Humble Opinion” with a more hard-hitting essay
titled “On Literary Revolution,” Chen called for “knocking down” all
kinds of “ornate,” “rotten,” “extravagant” old forms and styles and creat-
ing “colloquial social literature.”39 Chen’s essay itself, though, was written
35. Exhibition of the photography of Li Zhensheng ᴢᤃⲯ, California Museum of

Photography, University of California, Riverside, April 28–July 7, 2007.
36. From the memory of Tang Shaoyun ૤㒡ѥ, professor of art, Xiamen
University.
37. Hu Shi 㚵䘽, “Wenxue gailiang chuyi” ᭛ᅌᬍ㡃㢏䅄, Xinqingnian ᮄ䴦ᑈ [New
youth], no. 1 (January 1, 1917).
38. Hu Shi 㚵䘽, “Qingdai xuezhe de zhixue fangfa” ⏙ҷᅌ㗙ⱘ⊏ᅌᮍ⊩ (1921) in
Hu Shi wencun 㚵䘽᭛ᄬ (Collected works of Hu Shi), collection 1, volume 2 (Hefei:
Anhui jiaoyu chubanshe, 2003) (1921).
39. Chen Duxiu 䱇⤼⾔, “Wenxue geming lun,” Xinqingnian ᮄ䴦ᑈ [New youth],
no. 2 (February 1, 1917).
in a classical style that can only be called fairly ornate and a bit extrava-
gant. Here, too, an ironic gap opens between “what I say should be done”
and “what I am doing.” Two levels of awareness—focused consciousness
and inadvertent habit— are in play.
Are There Fads in Rhythms?
Fads in language use can originate in a number of ways, some of them pur-
poseful, but they seem often to spread at the same less-than-conscious level
that explains much use of rhythm in language. People sense that it feels good
to use a certain word or phrase without specifically noting that they are
following a trend in doing so. In the modern West teenagers are held to be
major purveyors of language fads, but in fact the adult worlds of business,
government, and academe are also full of examples. In the late twentieth
century words like “interface” and “module” were stylish in American bu-
reaucratic language. In the twenty-first, “robust,” “compelling,” and “mul-
tiple” have received faddish attention. There are plenty of other examples.
It seems that not only words but rhythms can be involved in faddish-
ness. This notion first occurred to me for the case of contemporary Chi-
nese when I noticed how many slogans of the Great Leap Forward years
(1958– 60) seemed to be four-syllable phrases:
Duo kuai hao sheng! ໮ᖿདⳕ More, faster, better, thriftier

Chao Ying gan Mei! 䍙㣅䍊㕢 Surpass England and catch up with
America.
Li gan jian ying ゟビ㾕ᕅ Erect a pole and see a shadow—get
instant results.
Tu fei shui zhong ೳ㙹∈⾡ Earth, fertilizer, water, seeds
Mi bao guan gong! ᆚֱㅵᎹ Dense protect manage labor.
[Mao’s formula for miraculous crops]
And so on. To test my hypothesis rigorously one would need to count the
occurrences of different kinds of rhythms from a variety of randomly se-
lected texts (and where possible, recordings) from different time periods.
I have not done this kind of rigorous study, but have examined, in the
Rhythm 45
spirit of a pilot study in that direction, headlines and slogans in the People’s
Daily for the month of October in 1951, 1958, and 1966, respectively.40 I
chose these three years in order to highlight the slogans of three promi-
nent campaigns (i.e., those called Suppression of Counterrevolutionaries,
The Great Leap Forward, and The Great Proletarian Cultural Revolu-
tion). In using the People’s Daily as my source I was aware of the bias toward
official and political language as opposed to daily-life usage, but still, the
easy comparability of three months of the People’s Daily made systematicity
possible, so the effort seemed worthwhile.
I found standard patterns of four, five, and seven syllables in all three
periods. In October 1951 an unusually large number of paired slogans ap-
peared that used an eight-syllable line followed by a seven-syllable line.
For example:
Gonggu renmin minzhu zhuanzheng 䵣೎Ҏ⇥⇥Џᇜᬓ

Jianjue zhenya fan’geming ෙ≎䦂ວড䴽ੑ
Consolidate the people’s democratic dictatorship.

Resolutely suppress counterrevolutionaries.41
Kuoda chengxiang wuzhi jiaoliu ᫈໻ජ䛝⠽䊾Ѹ⌕

Fazhan gongnongye shengchan ⱐሩᎹ䖆ὁ⫳⫷
Expand material exchange between city and countryside.

Develop industrial and agricultural production.42
Patterns of nine syllables also seemed unusually common in 1951:
Dazhang qigu zhenya fan’geming ໻ᔉ᮫哧䦂ວড䴽ੑ
Roll out the banners and drums in suppressing counter-

revolutionaries.43
40. I am grateful to Mao Sheng for assistance in this work.

41. Renmin ribao Ҏ⇥᮹ฅ [People’s daily], October 3, 1951, p. 3.
42. Ibid., October 20, 1951, p. 2.
43. Ibid., October 3, 1951, p. 3.
Zhongguo renmin datuanjie wansui Ё೟Ҏ⇥໻೬㌤㨀ⅆ
Long live the great unity of the Chinese people.44
In 1958, by contrast, four-syllable lines seemed unusually common, and

they were often paired with a second line that might have four, five, or
seven syllables. An example of a four-four pair is:
Loushang louxia, diandeng dianhua ὐϞὐϟ⬉♃⬉䆱
Electric lights and telephones upstairs and down.
A four-five pair:
Qingzhu woguo nongye dafengshou ᑚ⼱៥೑‫ݰ‬Ϯ໻Єᬊ
Congratulate our country’s big agricultural harvest.45
And a four-seven pair:
Quanmin dongyuan ܼ⇥ࡼਬ

Baozheng gangtie fan yi fan ֱ䆕䩶䪕㗏ϔ⬾
Mobilize all the people to guarantee a redoubling of steel.46
In the fall of 1966 rhythms of all kinds were more common in the pages of
the People’s Daily than they had been in either 1951 or 1958. The Cultural
Revolution rhythms seemed, too, to carry a more palpable “beat” than
others. For example:
Weida de daoshi, weida de lingxiu, ӳ໻ⱘᇐᏜ, ӳ໻ⱘ乚㹪

Weida de tongshuai, weida de duoshou, ӳ໻ⱘ㒳Ꮩ, ӳ໻ⱘ㠉᠟
Mao Zhuxi wansui! ↯Џᐁϛቕ!
44. Ibid., October 5, 1951, p. 3.

45. Ibid., October 1, 1958, p. 4.
46. Ibid., October 1, 1958, p. 9.
Rhythm 47
Long live the great teacher, great leader, great commander, great
helmsman Chairman Mao!47
In this example the first four phrases are not only five syllables in length
but invite a heavy lilt: ta-TAH-ta-ta-TAH, ta-TAH-ta-ta-TAH. . . . The
crucial contribution of the little particle de ⱘ, which precedes the modi-
fied nouns, is worth noting. It not only prevents a reading of the five-
syllable lines in the customary pattern of 2 + 3, with a pause after the sec-
ond syllable (as would certainly happen if, for example, the line were
weida zongsiling ӳ໻ᘏৌҸ ‘great commander-in-chief’). It also, because
it is repeated in parallel, magnifies the lilt. English, which has no corre-
sponding marker, does not have the same lilt-inducing tool available to it.
“Great teacher, great leader, great commander, great helmsman” can be
said with rhythmic parallelism, to be sure—but not with the swing that
comes naturally to weida de daoshi, weida de lingxiu, weida de tongshuai,
weida de duoshou. Or consider the example I noted in the Introduction of
the Cultural Revolution phrase weida de guangrong de zhengque de gongchan-
dang ӳ໻ⱘ‫ܝ‬㤷ⱘℷ⹂ⱘ݅ѻ‫‘ ܮ‬great, glorious, correct Communist
Party’. Compared to the ta-TAH-ta, ta-TAH-ta, ta-TAH-ta rhythm of the
Chinese, the English phrase “great, glorious, correct” sounds flat.48
In 1958 a casual comment from Mao Zedong seems to have led to a
rhythmic pattern that took at least two different shapes during China’s
late Mao years. In early August 1958, when Mao was on an inspection
tour of prototypes of the communes (dashe ໻⼒) that he was planning, a
local official in Shandong reportedly asked him what these new organiza-
tions should be called, to which he reportedly answered: haishi jiao “renmin
gongshe” hao 䖬ᰃি“Ҏ⇥݀⼒”ད ‘It’s probably best to call them “People’s
Communes” ’.49 Mao’s grammar in this sentence was fi ne, but when his
sentence was shortened, as it was in People’s Daily headlines and in many
other places, to renmin gongshe hao! Ҏ⇥݀⼒ད!, the result was a saying
that observes wuyan rhythm but is not, when lifted from context, a gram-
matically natural phrase. (Without hen ᕜ or ting ᤎ or some other adverb
47. Ibid., September 27, 1966, p. 4.

48. I am indebted to David Moser for this insight on de. Email message to author,
April 13, 2008.
49. Renminwang Ҏ⇥㔥 [People.com], http://cpc.people.com.cn /GB/64162/64170
/4467343.html. Viewed June 23, 2012.
of degree preceding hao, the sentence technically means “the People’s

Communes are better [than some alternative].”) But the phrase renmin
gongshe hao was repeated so often that it not only came to seem natural but
also established the pattern A–B–C . . . hao, which spread, especially during
the Cultural Revolution, to refer to a variety of politically correct things:
shehuizhuyi hao ⼒ӮЏНད ‘socialism is good’; dazibao hao ໻ᄫ᡹ད ‘big-
character posters are good’; wenhua dageming hao ᭛࣪໻䴽ੑད ‘the Great
Cultural Revolution is good’; geming weiyuanhui hao 䴽ੑྨਬӮད ‘the revo-
lutionary committees are good’; fuke nao geming hao ໡䇒䯍䴽ੑད ‘it is good
to return to class to make revolution’;50 haishi laoshi dian hao 䖬ᰃ㗕ᅲ⚍ད
‘it’s better to be a bit more sincere after all’.51 The pattern survived even
as late as 2005 and 2006 in the family-planning slogans jihua shengyu hao
䅵ߦ⫳㚆ད ‘the planning of births is good’ and zhisheng yige hao া⫳ϔϾད
‘producing only one is good’.52
It seems that Mao added a permutation to the pattern with another in-
advertent comment in the summer of 1966, when he said that the rebel-
lious spirit of the Red Guards was hao de hen དᕫᕜ ‘really good’. The
August 23, 1966, issue of the People’s Daily published a front-page editorial
under the headline hao de hen, and in the ensuing days this headline ap-
peared three times: hongweibing de wuchanjieji geming zaofan jingshen hao de
hen! 㑶ि݉ⱘ᮴ѻ䰊㑻䴽ੑ䗴ড㊒⼲དᕫᕜ ‘The proletarian revolution-
ary rebellious spirit of the Red Guards is really good!’53 Then many other
things became hao de hen as well. On September 6, 11, and 24, it was re-
ported that foreign friends from Albania, Japan, Congo, Chile, Cuba, and
elsewhere were sending in messages that Zhongguo wuchanjieji wenhua
dageming hao de hen Ё೑᮴ѻ䰊㑻᭛࣪໻䴽ੑདᕫᕜ ‘China’s Proletarian
Cultural Revolution is really good’.54 Eventually other phrases that used a
50. Dongfang hong ϰᮍ㑶 [The east is red], November 3, 1967, p. 2.
51. Hongqi zhanbao 㑶᮫៬᡹ [Red flag warfare] (Xinjiang), January 20, 1968, p. 1.
For many of the examples in this paragraph I am indebted to Mao Sheng, “Geming
weiyuanhui hao” (The revolutionary committees are good), unpublished manuscript,
March 2007.
52. Mao Sheng, “Geming weiyuanhui hao.”
53. Renmin ribao Ҏ⇥᮹᡹ [People’s daily], August 24, 1966, p. 3; August 26, 1966,
p. 2; August 27, 1966, p. 3.
54. Ibid., September 6, 1966, p. 5; September 11, 1966, p. 4; September 24, 1966,
p. 4.
Rhythm 49
rhythmic pattern of TAH-ta-TAH (resembling hao de hen) seem also to

have become stylish. In the August 24 People’s Daily, right below a head-
line that ended in hao de hen!, another ended in gan de hao!: shoudu guangda
gongnongbing relie huanhu “hongweibing” gan de hao! 佪䛑ᑓ໻Ꮉ‫⃶⚜⛁݉ݰ‬
੐þ㑶ि݉ÿᑆᕫད ‘Broad ranks of workers, peasants and soldiers in the
capital cheer the “Red Guards” for doing well’.55
There is room for further research in this area, but in any case the hy-
pothesis that rhythms have gone through passing fads in contemporary
Chinese, in political language and elsewhere, seems plausible in a number
of ways.
The Roots of Rhythms
Rhythms in language can result from set forms, as in songs or chants, or

from less consciously observed rules that involve grammar, meaning, em-
phasis, or simply convention. In Mandarin Chinese it is easy, for example,
to notice conventional stress patterns in which the second syllable of a
two-syllable compound or phrase is an unstressed “neutral” tone: er.duo
㘇ᴉ ‘ear’, qing .ni 䇋Դ ‘invite you’ and so on. (Here and below I follow
Y. R. Chao in using a dot to precede a syllable that is pronounced in the
unstressed “neutral” tone.) It is also easy to note how an unstressed sylla-
ble can take on stress when a speaker’s meaning calls for it: wo qing ta, bu
qing ni ៥䇋ཌྷ, ϡ䇋Դ ‘I invited her, not you’.
Other things being equal (and I will consider some of those “other
things” below), Mandarin Chinese has a clear preference for maintaining
syllabic balance. One syllable plus one syllable naturally form a balance,
and so do two plus two. When three syllables are involved, it sometimes
(not always) happens that the three are reduced to two so that balance can
be maintained. For example, “school” is xuexiao ᄺ᷵ in two syllables. To
say “middle school” (which is high school, in American usage) the modifier
zhong Ё needs to come first, but this would make a three-syllable zhong-
xuexiao, which feels awkward. So Chinese drops the -xiao, and the word for
high school becomes zhongxue, where the xue stands in for all of xuexiao.
55. Ibid., August 24, 1966, p. 3.

Then, among zhongxue, one can subdivide to either chu ߱ ‘beginning’ or

gao 催 ‘high’ levels (i.e., ju nior or senior high), but again chuzhongxue and
gaozhongxue are both three syllables, so again are shortened. Chuzhong
becomes the word for “junior high school” and gaozhong the word for “se-
nior high school.” But the original word xuexiao, which now is dropped
completely, is clearly understood. Similarly, for “hold a meeting” Chinese
uses either kai hui ᓔӮ in two syllables or juxing huiyi В㸠Ӯ䆂 in four,
split 2 + 2. A three-syllable kai huiyi feels a bit awkward, and juxing hui
sounds pretty awful. In four-syllable phrases, the overwhelming tendency is
to break the syllables 2 + 2, and this preference for 2 + 2 holds across differ-
ent kinds of grammatical structure.56 For example yuyan yanjiu 䇁㿔ⷨお
‘the study of language’ is 2 + 2, and so is yanjiu yuyan ⷨお䇁㿔 ‘study lan-
guage’. (There are rare cases in which four-syllable phrases can be broken
1 + 1 + 1 + 1, or 3 + 1, or 1 + 3, or even 1 + 2 + 1, and we will visit some of
these odd cases later; their oddness only serves to underscore the prefer-
ence for balance when a balanced alternative is available.)
Stress patterns become more complicated— and interesting—in cases
where three syllables more or less have to be involved. In three-syllable
strings, it is rare that each syllable is equally weighted, although examples
like Alabo 䰓ᢝԃ ‘Arabia’ and jiguanqiang ᴎ݇ᵾ ‘machine gun’ come
close. Nearly always, the three syllables cluster as either 2 + 1 pingguo da
㣍ᵰ໻ ‘the apple is big’ or 1 + 2 da pingguo ໻㣍ᵰ ‘big apple’. The break in
such phrases often is not a pause so much as a minor slowing down lead-
ing to slightly more stress on the syllable that follows. A “pause” exists
only by comparison to transitions that pause less. For example, to say that
there is a break between guo and da in pingguo da means only that there is
less of a break (or none at all) between ping and guo. This kind of a subtle
stress pattern, because it depends on comparison, cannot obtain when
only two syllables are involved. Er.duo ‘ear’ has an obvious stress pattern,
but it is not the same as—indeed is usually more prominent than—stress
patterns of the kind I am analyzing here, the kind that result from three-
or four-syllable clustering.
56. See Feng Shengli, “Prosodic Structure and Its Implications in Teaching Chi-
nese as a Second Language,” unpublished, April 2003, pp. 12–13.
Rhythm 51
In three-syllable phrases, grammar and meaning usually determine

whether the clustering is 2 + 1 or 1 + 2. Examples where the break comes
between subject and predicate are trivially easy to cite: women yao ៥Ӏ㽕
‘we want’ is 2 + 1, and ta buxin ཌྷϡֵ ‘she doesn’t believe’ is 1 + 2. The
breaks in modifier-plus-noun phrases are often determined by the gram-
mar and meaning of the component words. A fujingli ࡃ㒣⧚ ‘deputy man-
ager’ is always a fu + jingli, never a fujing + li, and doufutang 䈚㜤∸ ‘bean-
curd soup’ has to be doufu + tang, never dou + futang.
Other questions of where breaks occur are more subtle. An interesting
subfield in Chinese prosody, recently pioneered by Duanmu San, Feng
Shengli, and others, has studied what happens in stress patterns when
phrases are shortened. The results vary with grammatical structure. In
phrases where a verb is followed by a second verb, called a “complement”
(the relation of verb and complement in Chinese is similar to scared stiff or
tickled pink in English), shortenings from 2 + 2 syllables to either 1 + 1 or
1 + 2 are both all right, but 2 + 1 normally is not. For example, baifang
zhengqi ᨚᬒᭈ唤 ‘arrange into neat order’ can shorten to either fang qi or
fang zhengqi, but not baifang qi.57 Similarly jiangjie qingchu 䆆㾷⏙Ἦ ‘ex-
plain clearly’ can become either jiang qing or jiang qingchu, but not jiangjie
qing.58 In short, rhythmic factors not only induce certain patterns but can
rule out patterns as well.
Verb-object constructions normally observe the same rule, for exam-
ple, yuedu baozhi 䯙䇏᡹㒌 ‘read a newspaper’ can be either du bao or du
baozhi but not yuedu bao; similarly, zhongzhi shumu ⾡ỡᷥ᳼ ‘plant trees’
can be zhong shu or zhong shumu but not zhongzhi shu.59 Exceptions like
taoyan gou 䅼ॠ⢫ ‘fi nd dogs disgusting’ can be explained by the fact that
gou ‘dog’ does not have an easy two-syllable alternative. The rule is, as it
were, “soft,” in the sense that it describes a defi nite preference but allows
exceptions for good reason. Feng Shengli points out another interesting
exception: 2 + 1 for verb-objects is all right when the verb includes a neutral
57. The example is from Feng Shengli, “Facts of Prosodic Syntax in Chinese,” un-
published paper, p. 34.
58. The example is from Lu and Duanmu, “Rhythm and Syntax in Chinese,” p. 126.
59. Examples from ibid., p. 123.
tone, as in xi.huan qian ୰⃶䪅 ‘likes money’ or xia.hu ren ১଀Ҏ ‘frighten
people’.60
In cases where both modifier and modified are two syllables, a differ-
ence emerges depending on whether the modifier is a noun or an adjec-
tive. When a noun modifies a noun, the preference in shortening four
syllables to three is 2 + 1, not 1 + 2. Thus meitan shangdian ✸⚁ଚᑫ ‘coal
store’ shortens to meidian or to meitan dian, but not to mei shangdian.61 On
the other hand, when the two-syllable modifier is an adjective, 1 + 2 is
preferred: nüxing gongren ཇᗻᎹҎ ‘female worker’ shortens to nügong or
to nü gongren but not to nüxing gong.62 In addition to this puzzling differ-
ence, another difference between adjective modifiers and noun modifiers
is that only adjective modifiers can take the particle de. One can say nü-
xing de gongren for “female worker,” but not shoubiao de gongchang. (If we did
say the latter, its meaning would switch, and be odd. It would mean some-
thing like “a factory belonging to wristwatches.”)
Most of these patterns can be explained by what Duanmu San and oth-
ers call a “nonhead stress rule.” (The “head” of a phrase is the top of the
tree-structure in a phrase diagram.) In a verb-object or verb-complement
structure, the verb is the head; in a modification structure, the modified
noun is the head (unless the modifier is an adjective, in which case the
implied particle de is the head). In all cases, the component of a phrase
that is not the head receives stress. Duanmu discusses apparent exceptions
to the nonhead stress rule and on the whole defends the rule well.63
The rule applies to phrases longer than the four-syllable strings we
have been considering. In verb-object constructions, where the nonhead
object needs more stress, a two-syllable verb cannot go with a one-syllable
object: hence du bao ‘read the newspaper’ is all right, but yuedu bao sounds
wrong. Similarly, a three-syllable verb sounds wrong with a two-syllable
object. We can say fuze bingfang 䋳䋷⮙᠓ ‘take responsibility for the
60. Feng Shengli, “Facts of Prosodic Syntax in Chinese,” p. 5.

61. Lu and Duanmu, “Rhythm and Syntax in Chinese,” p. 123.
62. Ibid., p. 128.
63. Ibid., pp. 126–133. See also San Duanmu, “Stress and the Development of Disyl-
labic Words in Chinese,” Diachronica 16 (1999), pp. 15–16, and The Phonology of Stan-
dard Chinese (Oxford: Oxford University Press, 2000), ch. 6.
Rhythm 53
[hospital] ward’ but not fuzeren bingfang 䋳䋷ӏ⮙᠓ (of the same meaning).
Feng Shengli points out that here it is the nonhead stress problem, not
the 3 + 2 clustering of syllables, that rules out fuzeren bingfang. Dui bing-
fang fuze ᇍ⮙᠓䋳䋷 ‘be responsible for the ward’ is also three syllables
plus two, and is quite fi ne, because now dui bingfang is the nonhead.64
While the nonhead stress rule and related grammar rules can, as we have
just seen, be “strong” enough to rule out certain phrases (i.e., make them
what grammarians call “unsayable” in normal usage), it is also true that
they are sometimes insufficient to determine how a string of syllables should
be said, or set to rhythm, or understood. In such cases meaning also plays
a role. This point is clear if we look at strings of nouns in which the pos-
sibilities of modification are ambiguous. Feng Shengli uses the example pi
Ⲃ ‘leather,’ xie 䵟 ‘shoes,’ and chang ॖ ‘factory’ to illustrate. Each of these
syllables by itself is a noun, and the three start, as it were, on equal foot-
ing. If we string them together as pixiechang, it is possible (if we imagine a
certain context) to understand the phrase as three nouns meaning “leather,
shoes, and factories” (1 + 1+ 1). A person could say, for example: pi xie chang,
women Lanzhou dou you Ⲃ䵟ॖ៥Ӏ݄Ꮂ䛑᳝ ‘leather, shoes, and factories—
we’ve got them all in Lanzhou’. But when we consider the more likely
case in which the nouns can modify one another, the question of cluster-
ing arises. In writing, pixiechang could be seen as either 2 + 1 pixie chang
‘leather-shoe factory’ or 1 + 2 pi xiechang ‘shoe factory made of leather’ (or,
more intriguing, ‘naughty shoe-factory’). In deciding between these two
(and the 1 + 1 + 1 alternative), meaning makes the difference in our choice,
and the “right answer” is fairly obvious. There is no grammatical reason to
prefer pixie chang over pi xiechang, but we realize that (in the normal world,
anyway) a leather-shoe factory is a more plausible thing than a shoe-
factory made of leather. This recognition guides our interpretation of the
grammar and also tells us where the pause and stress should go if we read
the phrase aloud.
In a startling footnote, Duanmu San points out that meaning is some-
times strong enough not only to decide among grammatical possibilities
within a string of syllables but also, like grammar rules themselves, to
64. Feng Shengli, “Facts of Prosodic Syntax in Chinese,” pp. 36–37.

render certain things “unsayable.”65 In certain pairs of statements of the

forms ‘X is P’ and ‘X is not-P’, only one of the two can normally be said.
(The point here is not the logical point that ‘X is P’ and ‘X is not-P’ cannot
simultaneously be true; the point is that only one of the two forms is nor-
mally “sayable,” regardless of whether it is true.) A two-syllable adjective
with a positive meaning like qingchu ⏙Ἦ ‘clear’ or anquan ᅝܼ ‘safe’ can
be preceeded by hen bu ᕜϡ ‘very not’: hen buqingchu ‘very unclear’, hen
buanquan ‘very unsafe’, and so on. But two-syllable adjectives of negative
meaning, like mohu ῵㊞ ‘muddled’ or weixian ॅ䰽 ‘dangerous’ do not
similarly follow hen bu. No one (except in odd contexts) says hen bu weixian
‘very not-dangerous’. To compound this peculiarity, the point exactly re-
verses when one makes the minor change of talking about one-syllable
adjectives instead of two-syllable adjectives. Now adjectives with negative
meanings take hen bu (hen bushao ᕜϡᇥ ‘quite a few’; hen buruo ᕜϡᔅ
‘very not weak—pretty strong’), but adjectives of positive meaning do not.
Normally no one says hen buqiang ᕜϡᔎ ‘very not-strong’. The mystery
gets even deeper if one reverses hen bu to buhen. Now only adjectives with
positive meanings—regardless of whether they are one syllable or two—
can be accommodated. One can say buhen qingchu ‘not very clear’ or buhen
da ‘not very big’, but not buhen mohu ‘not very muddled’ or buhen shao ‘not
very few’. Duanmu presents these data but cannot explain them, and nei-
ther can I. My purpose here is simply to illustrate further the point that
sometimes the meanings of phrases—not just their grammar— go into
determining whether native speakers find them “sayable.”
“External” Rhythms: Dominant and Recessive
We have seen how grammar can affect the rhythms of phrases and how,
when phrases in written form are ambiguous, meaning can make the dif-
ference in determining grammatical relations and associated rhythms.
But there are also rhythms in language that originate outside of grammar
65. Lu and Duanmu, “Rhythm and Syntax in Chinese,” p. 133, n. 7. The authors
credit Lü Shuxiang, “Hen bu . . .” [Very not . . . ], Zhongguo Yuwen [Chinese language]
5 (1965), for the original insights on hen bu.
Rhythm 55
or meaning. I will use the term “external rhythm” to refer to these. These
are cultural artifacts that are used and understood independently of— and
sometimes despite—what grammar requires.
Some cases of external rhythm are hard to miss. In any human culture,
poetry, songwriting, and things like chants and marches include meter
that has patterns that are external to the requirements of grammar. I will
call these kinds of external rhythms “dominant.” By “dominant” I mean
that the people who put them to use— and very often the people who re-
ceive them as well—are consciously aware of what is going on. When a poet
or songwriter composes a line in a dominant pattern, the rhythm is “there
first,” as it were, and the poet or songwriter’s task is to select and arrange
words to fit it. The pattern is “strong” enough that it can be acceptable even
for a poet or songwriter to overrule normal grammar in order to accom-
modate a rhythm.
Normally, for example, the English phrase “the button he pressed” would
be understood as a noun phrase meaning “the button that he pressed.” In a
limerick, though, it can mean “he pressed the button.” This inversion of verb
and object is justified—and does not confuse a cooperative reader—because
the demands of the dominant rhythm (and in the following case, rhyme,
too) need to be met:
Endeavored a boy of No. Dak.

To picture a bear with a Kodak.
The button he pressed,
And the bear did the rest—
The boy stopped running in So. Dak.66
The first line above, like the third, reverses the normal subject-predicate
order, but in neither case do we mind. In a limerick, even such fi xed as-
pects of language as spelling and pronunciation can bow to rhythm and
rhyme: No. Dak. and So. Dak. are made to rhyme with Kodak. Limericks
are, to be sure, extreme cases. They are an especially mischievous form to
which we give especially broad license. But for that very reason they are
66. Cyril Bibby, The Art of the Limerick (Hamden, Conn.: Archon Books, 1978),
p. 193. I have taken the liberty of substituting “boy” for “lady” in Bibby’s version.
good examples of what I want to call “external” rhythm of the “dominant”

kind. The rhythm in a limerick is unapologetic about its formal demands,
makes both writer and reader quite aware of what these demands are, and
can, if necessary, push around other facets of language.
But there is another kind of cultural rhythm, which is still “external” in
the sense that it is not required by grammar, but is not as explicit or nearly
as pushy as the dominant kind. I will call this kind “recessive.” In the In-
troduction, I cited the example of the qiyan pattern that underlies the fi rst
line of Lu Xun’s story “Kong Yiji”: Luzhen jiudian de geju, shi he biechu bu-
tong de 元䦂䜦ᑫⱘḐሔ, ᰃ੠߹㰩ϡৠⱘ ‘the layout of the wineshops in
Lu-town was different from those in other places’. Few ordinary readers
will be consciously aware that qiyan rhythm undergirds this line, but few,
too, will not sense the grace it adds. Ian McEwan has noted a similar way
iambic pentameter underlies John Updike’s line “the rooms are quadrants
of one rustling heart,” which, McEwan writes, produces a “sweetly pitched”
description of a house in which lovers are obliged, by the presence of others
in a house, to enjoy each other quietly.67
Some recessive rhythms are standard parts of a language, even when
they are not consciously noticed. An example from contemporary Ameri-
can English is the pattern (which employs pitch as well as rhythm) that
goes “low-low-low-HIGH, low-low-low-HIGH, low-low-low-HIGH . . .”
(The number of “low” syllables can vary.) For example, if I had a bad day
at the mall I might say, “First it rained, then I couldn’t fi nd parking, then
I lost my credit card . . .” and so on. This pattern usually tells the listener
three things: (1) I am giving you a list of related items; (2) This list has no
certain end; and (often, but not always) (3) I feel exasperation about the
items on the list.
Unlike their “dominant” cousins, “recessive” rhythms defer to other
facets of language. They have ways that they work within the rules of
grammar but still make important contributions to the way phrases “feel”
in the minds of speakers and hearers. Speakers and hearers, while remain-
ing largely if not entirely aware of the rhythms, use them as they use most
grammar and pronunciation rules— automatically, as it were. They do not
67. Ian McEwan, “On John Updike,” New York Review of Books 56, no. 4 (March 12,
2009), p. 6.
Rhythm 57
give them the conscious notice that they normally give to word selection
(at least for the crucial words in an utterance) or, when they are present, to
“dominant” rhythms, either. But recessive rhythms, even though used in
this less than fully conscious way, are still cultural in the sense that they
are the artifacts of certain cultures and not necessarily of others.
In distinguishing “dominant” and “recessive” rhythms, I am not claim-
ing that every example of a rhythm has to be one or the other kind. Lim-
ericks, to be sure, are sufficiently distinctive that we might say they are
always dominant. (Can we imagine a person producing a limerick, or
passing one along, inadvertently?) There are also complex poetic forms,
such as the qinyuanchun ≕೦᯹ ‘spring in Qin’s garden’ form, which is so
dominant that even Mao Zedong obeyed it and so ornate that anyone who
seeks to use it needs to be quite aware of exactly what he or she is doing.
But the great majority of rhythms that I discuss in this book, including
wuyan, qiyan, and other examples in both Chinese and English, are not
essentially either dominant or recessive but can be, and very often are, either
one. Rigorously speaking, what I am distinguishing here are not two sets
of examples of rhythms but two ways rhythms can function.
Enough prologue. What are the “external rhythms” of Chinese? Can
we uncover them? The “dominant” ones have been well catalogued by
scholars of Chinese poetry. This is a vast and rich field, in which there is
always more to learn, but the importance of rhythms in it hardly needs to
be “uncovered.” The “recessive” patterns, though, are another matter.
They have not been studied, and it is tricky to figure out how to uncover
them. The problem is not just that people are unaware of them and there-
fore not able to identify them. The problem is also that the rhythms, be-
ing “recessive,” defer to other aspects of language structure. They are
camouflaged, as were, within syntax. Consider again the example (from
the Introduction) of the Red Guards in Tiananmen Square chanting
women yao jian Mao zhuxi! ‘we want to see Chairman Mao’ in a qiyan pat-
tern of 2–2–3 syllabic structure. Can we say, from this evidence, that
2–2–3 is an “external” pattern of the “recessive” kind? No, we cannot,
because grammar easily explains the 2–2–3 pattern as well. The first two
syllables of women yao jian Mao zhuxi are its subject, the next two are its
verb, and the last three its object. The 2–2–3 pattern comes right out of
the grammar, and there is no need to postulate an “external” rhythm. To
be sure, we might speculate that the Red Guards hit on this particular
combination of subject, verb, and object because they started, somewhere
in their quasi-consciousness, with a sense that qiyan rhythm “sounds good.”
That would be plausible, but hard to prove.
Other examples I have discussed are equally frustrating from this point
of view. In yi kan, er man, san tongguo ‘first look, then go slowly, then cross’,
qiyan rhythm coincides with grammar in a similar way. So does wuyan
rhythm in zhanglang, si guangguang! ‘cockroaches— dead to the last one!’
An interesting question thus emerges. Are there intrinsic preferences for
certain kinds of recessive rhythms in Chinese, where by “intrinsic” I mean
preferences that exist independently of the pushes and pulls of grammar?
If so, is there any way we can identify these preferred rhythms? If we want
to say that recessive rhythms are cultural items in their own right, that
speakers and writers are expressing preferences when they decide to use
them or not, and that the rhythms contribute some kind of sense or feel-
ing to a phrase (all of which are claims that I want to make in this chapter),
then we will need, somehow, to separate rhythms from grammar, to see
what rhythmic preferences might obtain when grammar does not domi-
nate. But how?
In his article “On the ‘Natural Foot’ in Chinese,”68 Feng Shengli has
come up with an ingenious method for doing just what we need—isolating
recessive rhythms. (Feng’s term “natural foot” is synonymous with what I
am calling “recessive rhythm.”) Feng reasons that if we can find strings of
syllables that are exactly parallel grammatically, then grammar will not be
the explanation for any rhythms that might appear within them. If rhythms
do appear, and if they fall into consistent patterns (that is, are not just ran-
dom from case to case), then the patterns must be intrinsic preferences of
Chinese culture.
Feng does his experiment using two kinds of parallel syllables.
One kind are the syllables that are used for sound only in the translitera-
tion of foreign words. For example, just as “Joey Davis” has an internal
stress pattern in English, so does “Mayakovsky” in Russian. But when
68. “Lun Hanyu de ‘ziran yinbu’ ” 䆎∝䇁ⱘ“㞾✊䷇ℹ” [On the ‘Natural Foot’ in
Chinese], Zhongguo yuwen Ё೑䇁᭛, no. 1 (January 10, 1998).
Rhythm 59
“Mayakovsky” is transliterated into Chinese Ma-ya-ke-fu-si-ji ⨾䲙ৃ໿ᮃ෎,

the lilts of Russian are lost. The syllables undergo a radical homogeniza-
tion, reducing to something like tat-tat-tat-tat-tat-tat and, as it were,
“starting over” rhythmically. If any patterns in the new string happen to
emerge, they will come from Chinese cultural preferences, not Russian.
And no grammar, either Chinese or Russian, will get in the way.
Feng’s second method of testing uses items in series. In referring ge-
nerically to kitchen utensils, for example, the Chinese language some-
times uses the set phrase guo wanr piao penr 䤟⹫⪶Ⲛ ‘woks, bowls, ladles,
and pots’. For kitchen staples, there is chai mi you yan jiang cu cha ᷈㉇⊍呑
䞀䝟㤊 ‘kindling, rice, oil, salt, soy sauce, vinegar, and tea’. I will call such
examples “item lists.” The items that they contain are often nouns, but do
not have to be. Dissolute behavior can be summarized in the verbs chi he
piao du ৗୱႪ䌠 ‘eat, drink, brothel, gamble’, and the basic colors are hong
huang lan bai hei 㑶咘㪱ⱑ咥 ‘red, yellow, blue, white, black’. The only re-
quirement of item lists is that the items be parallel.
Item lists usually involve things with much more cultural content than
we encounter in transliterations like Ma-ya-ke-fu-si-ji. The choice of which
kinds of items to put into such lists is a cultural choice, and the order in
which they appear is also fixed by cultural convention. (If, for example, you
accused someone of he du piao chi instead of chi he piao du, people might
worry about you as much as the accused.)69 But even if item lists do open
the door to small cultural influences, Feng Shengli is right to include item-
list examples in his experiment because the cultural elements are not gram-
matical. There is no grammatical reason to insert either stresses or pauses
inside a string like hong huang lan bai hei. Feng is right that if a pattern does
appear, then a “recessive” rhythm must be at work.
Let us see what happens when we do Feng Shengli’s experiment on
syllable strings of various lengths, using both transliterations and item
lists.
69. I say transliterations “usually” have less cultural content because, as marketers
of Western products in China have discovered, a transliteration can be made to sug-
gest a meaning as well. Baishi kele ⱒџৃФ transliterates “Pepsi- Cola” but also means,
to the Chinese ear, “everything’s felicitous.”
Recessive Rhythms of Favor
Two-syllable strings often give equal stress to each syllable: plunk,

plunk. Examples are Aiji ඗ঞ ‘Egypt’, a transliteration, and yinyang 䰈䰇
‘yin and yang’, an item list. Since clustering is impossible with only two
syllables, variation in the length of pauses cannot be a factor. Many
two-syllable words that use “neutral tones” in modern Mandarin put a
stress on the first syllable: di.di ᓳᓳ ‘younger brother’, ma.fan 咏⚺
‘trouble(some)’, and so on.70 But these imbalances do not happen (or any-
way are extremely rare) in transliterations or item lists. Dong.xi ϰ㽓
‘(concrete) thing’ stresses the first syllable, but the same characters dongxi
meaning ‘east and west’ are an item list and give equal stress to each. I
believe it is possible—barely—to hold that two-syllable transliterations
that end with a syllable pronounced in a fourth tone, such as ka-te व⡍
‘Carter’, tend to stress the fourth-tone syllable a bit more than its prede-
cessor. Fourth tones, because they fall sharply and relatively far, can natu-
rally seem more emphatic. But this effect, to the extent that it exists,
probably owes something to the fact that the fourth-tone syllable is the
second of the two in ka-te. In a word like Beining 䋱ᅕ ‘Benin’, the effect
disappears.
With transliterations or item lists that contain three syllables, a “plunk,
plunk, plunk” pattern of equal stresses without clustering (i.e., 1 + 1 + 1) is
often a norm. Luoshanji ⋯ᴝⷊ ‘Los Angeles’ is said this way, as is the
item list tian di ren ໽ഄҎ ‘heaven, earth, human being (the cosmos, in
sum)’. It would be awkward to insert major pauses or stresses into such
phrases. Feng Shengli notes, however, that sometimes, in some translitera-
tions, a subtle 2 + 1 clustering does emerge. Xiyatu 㽓䲙೒ ‘Seattle’, Jia’nada
ࡴᣓ໻ ‘Canada’, and Moxige ๼㽓હ ‘Mexico’ can all be said 1 + 1 + 1, es-
pecially if said slowly. But when spoken quickly and fluently, a subtle
tendency toward 2 + 1 is detectable. The very slight clustering creates not
70. Duanmu San notes that disyllabic words, whether including neutral tones or
not, became the norm in Chinese during the twentieth century. An extensive study by
the Chinese Language Reform Committee in 1959 showed that 69.8 percent of the
three thousand most commonly used words were two syllables. “Stress and the Devel-
opment of Disyllabic Words in Chinese,” pp. 4–5.
Rhythm 61
even a pause, but only a brief slowing-down that results in slightly more
emphasis on the syllable that follows. I will make a slash (/) stand for this
brief slowing-down. If we listen carefully, with Feng Shengli, we can hear
that Xiyatu is Xiya/tu; Jia’nada is Jia’na/da; and Moxige is Moxi/ge. The
subtle difference emerges more clearly if we compare 2 + 1 to 1 + 2. Try (if
you speak Chinese) saying Xi/yatu or Jia/nada. These just don’t work as
well as Xiya/tu or Jia’na/da. The same seems to hold for item lists. Pro-
nounced slowly, shulihua ᭄⧚࣪ ‘math, physics, and chemistry’ and fu-
lushou ⽣⽘ᇓ ‘wealth, position, and longevity’ are both 1 + 1 + 1. But said
fluently, shuli/hua sounds better than shu/lihua and fulu/shou better than fu/
lushou. I believe that even in such tiny differences as these we can begin to
feel the “meaning” of Chinese rhythms. I think a Chinese speaker would
feel a bit more comfortable with a promise of fulu/shou than of fu/lushou.
And I would guess that if you interviewed would-be Chinese tourists
about tour packages to Seattle, they might be slightly more attracted to
Xiya/tu than to Xi/yatu.
Recessive rhythms begin to emerge more clearly when we look at syllable-
strings longer than three. In four-syllable examples, a preference for 2 + 2
is quite clear. Feng Shengli notes that Sri Lanka is normally said as Sili/
Lanka ᮃ䞠݄व, Pakistan is Baji/sitan Ꮘ෎ᮃഺ, and Tanzania is Tansang/
niya ഺḥሐѮ, all with a slight slowing-down between the second and
third syllables; similarly for item lists, Chinese speakers tend to say dongxi/
nanbei ϰ㽓फ࣫ ‘east, west, south, north’, chaimi/youyan ᷈㉇⊍Ⲥ ‘fire-
wood, rice, oil, salt’, and jiajian/chengchu ࡴ‫ޣ‬Ь䰸 ‘addition, subtraction,
multiplication, division’.71 The preference for 2 + 2 is confirmed in cases
where a matching line is also 2 + 2. Chi he piao du ‘eat, drink, brothel,
gamble’, for example, is commonly followed by wu e bu zuo ᮴ᙊϡ԰ ‘no
evil not done’. In wu e bu zuo grammar intervenes to reinforce the 2 + 2
pattern, but this intervention only confi rms that 2 + 2 is the right way to
say the first line.
Rhythmic preferences become considerably more obvious in cases of
five-character and seven-character strings. This is not surprising, in view
of the prevalence of wuyan and qiyan patterns in Chinese culture that I
have noted in a variety of examples. Five-syllable transliterations tend to
71. Feng Shengli, “Natural Foot,” p. 42.

split 2 + 3 in standard wuyan form. Aerbaniya 䰓ᇨᏈሐѮ ‘Albania’ tends

to be Aer/baniya, and Bolshevik is Buer/shiweike ᏗᇨҔ㓈‫ܟ‬. In item
lists, the five elements tend to be pronounced as jin mu/shui huo tu 䞥᳼∈
☿ೳ ‘metal, wood, water, fire, earth’, and the five metals as jin yin/tong tie
xi 䞥䫊䪰䪕䫵 ‘gold, silver, copper, iron, tin’.72 In seven-syllable examples,
Buyinuosiailisi Ꮧᅰ䇎ᮃ㡒߽ᮃ ‘Buenos Aires’ tends to be Buyi/nuosi/
ailisi, and the list of every item one standardly needs in the kitchen,
chaimiyouyanjiangcucha ᷈㉇⊍Ⲥ䝅䝟㤊 ‘kindling, rice, oil, salt, soy sauce,
vinegar, and tea’, tends to be chaimi/youyan/jiangcucha.
The breaks in all these phrases—which are 2/3 for wuyan and 2/2/3 for
qiyan— are the same as in classical Chinese poetry (in the study of which
they are referred to as “caesuras”). In many cases a minicaesura among
the final three syllables in a qiyan line, that is, 2/1, is also detectable. (The
last three syllables of Buyinuosiailisi, for example, are subtly closer to aili/
si than to ai/lisi.)
The breaks that emerge in transliterations and item lists are normally
much briefer than those involved in the oral reading of poetry. When
classical poems are read aloud, the pauses can be emphasized— drawn out
and savored—to marvelous effect. To do the same with Albania, or salt
and soy, would be comical.
Sometimes, too, classical poets alter the standard pattern of caesuras
for deliberate artistic effect. Matthew Chen cites two seven-character
lines by Ouyang Xiu (d. 1072 ce) in which a “mismatch” between syntax
and 2 + 2 + 3 rhythm creates “metrical tension” that seems clearly to be
part of the poet’s art:73
jing ai zhu shi lai ye si 䴰ᛯネᰖ՚䞢ᇎ

Du xun chun ou guo xi qiao ⤼ᇟ᯹‫ي‬䘢⑾‟
A quiet lover of bamboo, I come often to the rustic temple

Alone in search of spring, I happen across the bridge over the brook
72. Ibid., p. 41.

73. Matthew Y. Chen, “Metrical Structure: Evidence from Chinese Poetry,” Lin-
guistic Inquiry 10, no. 3 (1979), pp. 406, 412.
Rhythm 63
Here the grammar clearly asks for a 3 + 2 + 2 rhythm, generating “mis-

match” with the standard 2 + 2 + 3 pattern. The poet is playing with us.
“Handle the mismatch,” he seems to be saying, “and feel its exquisite-
ness.” The 3 + 2 + 2 mismatch appears in more popular forms as well, for
example in the title of a story by Feng Menglong (1574–1646): “Maiyou-
lang duzhan huakui” 䊷⊍䚢⤼ऴ㢅儕 (The oil vendor wins the courte-
san). But 3 + 2 + 2 does not, as far as I can tell, appear in transliterations or
item lists, where the standard patterns of 2 + 3 for wuyan and 2 + 2 + 3 for
qiyan are, even if articulated faintly, overwhelmingly preferred.
But just now, in considering five-syllable and seven-syllable strings to-
gether, I skipped over the case of six-syllable strings. What about them?
And what about strings of eight, nine, or more syllables? Are cultural
preferences in rhythm still involved? Six-syllable rhythmic patterns do
exist in Chinese. Classical pianwen 侶᭛, with roots as old as the Han pe-
riod, observes a pattern of four plus six syllables. Pianwen was a refined
art, and its rhythms were normally “dominant” in the sense in which I am
using the term. They were consciously employed, but also restricted, it
seems, to a cultural elite; there is no evidence that they ever had the kind
of general use that wuyan and qiyan have today.
Whether indirectly related to pianwen or not, modern phrases of six
syllables can be found, although they are not nearly as common as wuyan
and qiyan. A prominent Cultural Revolution slogan was zhua geming, cu
shengchan ᡧ䴽ੑ, ֗⫳ѻ ‘grasp revolution, promote production’.74 A 2004
advertisement on CCTV said pifu hao, yong dabao Ⲃ᳡ད, ⫼໻ᅱ ‘for good
skin, use Great Treasure Cream’. In these cases, the six syllables are bal-
anced 3 + 3, but there are also cases of 2 + 2 + 2. Lesson 15 of a textbook
aimed at spreading literacy among farmers, published in Sichuan in 1979,
opens with zhengyue lichun yushui, eryue jingzhe chunfen ℷ᳜ゟ᯹䲼∈,
Ѡ᳜᚞㳄᯹ߚ ‘in the first month spring arrives and the rains come, in the
second month the insects awaken and spring begins’ and continues with
twelve lines of six characters apiece, all in a 2 + 2 + 2 pattern.75 (This
74. Renmin ribao Ҏ⇥᮹᡹ [People’s daily], September 7, 1966, p. 1.

75. Mianyang diqu wenjiaoju 㓉䰇ഄऎ᭛ᬭሔ, Nongmin shizi keben ‫⇥ݰ‬䆚ᄫ䇒ᴀ
(A textbook for teaching literacy to farmers) (n.p., 1979), p. 22.
clearly is a case of “dominant” pattern use.) A 2007 family-planning slo-

gan also used a 2 + 2 + 2 arrangement: kongzhi renkou shuliang, guan’ai diqiu
muqin ᥻ࠊҎষ᭄䞣, ݇⠅ഄ⧗↡҆ ‘control population numbers, care for
Mother Earth’.76
But why are six syllables in a rhythmic string far less common than five or
seven? There seems to be a cultural preference for five or seven that steers
usage in those directions. For example, the lesson text in the farmers’ text-
book that I just cited, near the bottom of the page, abandons its six-syllable
pattern and uses seven-syllable lines. When this happens a tone shift seems
to occur; the seven-syllable lines are weightier; they sum up the truth of the
lesson. The six-syllable lines bring us through all the farming routines of
the twelve months of the year, and these lines somehow almost suggest in-
completeness: with each there is a next step to come. But with the seven-
syllable lines at the end, the tone suggests summary and wisdom:
Meiyue liang jie bubiangeng ↣᳜ϸ㡖ϡব᳈

Zuiduo xiangcha yi liang tian ᳔໮ⳌᏂϔϸ໽
Buwu nongshi yao zhuajin ϡ䇃‫ݰ‬ᯊ㽕ᡧ㋻
Kexue zhongtian duo gaochan ⾥ᄺ⾡⬄༎催ѻ
Two periods in each month without exception

Never differing by more than one or two days
We must watch closely to miss no farming times
And reap high production through scientific planting.
If there is indeed a cultural preference for five syllables or seven, as opposed

to six, we should expect it to show up on the neutral turf of transliterations
and item lists. Does it? Is there a tendency (where a choice is possible) to
choose five or seven syllables in a transliteration, instead of six? Or to make
item lists five or seven items long, rather than six (or, perhaps, eight)?
Yes, there does seem to be such a tendency, and we can test the case by
adding or subtracting syllables and observing what happens.
For item lists, we should begin by noting that some lists are not arbi-
trarily expandable or contractible. “North, south, east, west,” for example,
76. From “Jihua shengyu xin biaoyu chutai neiqing.”

Rhythm 65
is 2 + 2 in Chinese, and either dongnan/xibei or dongxi/nanbei, but we can-

not make the number of directions three or five just by wanting to. (We
could include the intermediate directions dongnan ϰफ ‘southeast’, xibei
㽓࣫ ‘northwest’, etc., but then our list would automatically jump to
eight, making it impossible to test for the appeal of five or seven.) Luckily,
for our purposes, most item lists in Chinese are arbitrary. They could,
in theory, easily be more or fewer in number, so a fact that they tend to
settle at five or seven (or whatever number) can be considered a cultural
preference.
We saw such preferences in the examples jin yin / tong tie xi ‘gold, silver,
copper, iron, tin’ for metals and chai mi / you yan / jiang cu cha ‘kindling,
rice, oil, salt, soy sauce, vinegar, and tea’ for kitchen supplies. Now let’s do
a thought experiment. What if someone said, of the metals list, “We
should add aluminum. In the modern world we no longer have tin cans—
they’re mostly iron or aluminum. If tin deserves to be on the list, alumi-
num certainly does, too.” So we add lü 䪱, and what happens? It doesn’t
sound very good. Jin yin tong tie xi lü seems awkward, and it doesn’t help
much to break it up as jin yin / tong tie / xi lü or as jin yin tong / tie xi lü. But
if we could come up with one more metal, to make seven, then we would
be in good shape again. Let’s add zinc xin 䫠. Jin yin / tong tie / xi lü xin now
feels much better. The seven-syllable phrase is not standard in the lan-
guage, and my point is not that it is or that it should be. My point is that if
Chinese were to go beyond five items on the metals list, it is likely that
seven would be the pattern—whether or not aluminum and zinc were the
two metals that were added.
A thought experiment on the kitchen staples yields a similar result.
Among chai mi you yan jiang cu cha ‘kindling, rice, oil, salt, soy sauce, vin-
egar, and tea,’ we might ask “Why is kindling there?” People don’t ingest
kindling, as they do the other items, so maybe it should not be on the list.
But does mi you yan jiang cu cha— six syllables— sound as good? Hardly. So
let’s put the kindling back. But then what about garlic suan 㩰? What is
Chinese food without garlic? It would seem as important as the others, so
what happens if we add it and go to eight items? Chai mi you yan jiang cu
cha suan is not too bad, actually. Eight syllables, if not as elegant as seven,
seems better than six. (Some scholars have held that the seven syllables of
qiyan are actually eight with a musical rest at the end; this seems right, and
makes it easier to understand why the expansion from seven syllables to

eight seems natural.) But what if we add ginger jiang ྰ, to make nine?
Chai mi you yan jiang cu cha suan jiang is cumbersome, and not very good.
Nine syllables or more may begin to be too long to feel like a single
phrase, and it may be that limits of this sort inhere in the human brain.77
In addition to thought experiments, there are real-world examples to
suggest a tendency to prefer item lists of five or seven rather than other
nearby numbers. In the early stages of the Cultural Revolution, for ex-
ample, Mao Zedong named “five black categories” of people: landlords,
rich peasants, counterrevolutionaries, bad elements, and rightists. Each was
summarized in a one-syllable abbreviation: di fu fan huai you ഄᆠডണে.
One might say, of course, that “five” was not Mao’s preference; maybe it
just naturally happened that there were five kinds of bad people in China,
just as there are naturally four directions. But later, as further problems
popped up for Mao, other categories bad people had to be labeled. There
were capitalist-roaders, rightist-tendency elements, reactionary scholarly
authorities, and others. At different times and places, the five black cate-
gories expanded to higher numbers. But they didn’t seem to stop at either
six or eight. Eventually two versions, the “five” and the “seven” black cat-
egories, became standard.78
Syllable strings of nine or more, if they do have set rhythms, are likely to
be built from shorter phrases. Take, for example, the pattern that I called
“3–3–7” earlier, a pattern that feels natural in a variety of contexts, from
cheerleading to singing a national anthem. The “3–3” part of it, by itself,
can feel awkward. We have noted that an expansion of the five metals to
six does not sound very good: jin yin tong tie xi lü 䞥䫊䪰䪕䫵䪱 is awk-
ward, whether read as 2 + 2 + 2 or as 3 + 3. The metals sound better in ei-
ther five or seven syllables. But watch: six metals sound just fine if you
follow them with a list of seven more to make 3–3–7: jin yin tong, tie xi lü,
77. The psychologist George Miller, in his seminal article “The Magical Number
Seven, Plus or Minus Two,” Psychological Review 63 (1956), pp. 81– 97, summarizes
studies that show that the human ear and brain normally can handle discriminations
of pitch and of loudness up to a limit of seven or so, after which confusions happen
more easily than before.
78. See Guo Jian, Yongyi Song, and Yuan Zhou, Historical Dictionary of the Chinese
Cultural Revolution (Lanham, Md.: Scarecrow Press, 2006), pp. 13–14.
Rhythm 67
qian bo you gang lei nie wu 䞥䫊䪰ˈ䪕䫵䪱ˈ䪙䪖䪔䩶䭁䬡䩼 ‘gold, silver,

bronze; iron, tin, aluminum; lead, platinum, uranium, steel, radium,
nickel, tungsten’. The whole phrase sounds so nice that a person might
want to go out and buy all thirteen metals.
In any case, I think we can conclude that cultural preferences in rhythm
are strong enough to help shape the length and internal structure of item
lists. There are plenty of examples. It thus becomes interesting to note
that, by comparison, rhythms seem to have less power in shaping translit-
erations. As noted above, Feng Shengli seems right to observe a subtle
preference for Xiya/tu over Xi/yatu when saying ‘Seattle’; but in lengthier
transliterations the cultural shaping seems to lose power. For example,
Chinese uses six syllables to transliterate Uzbekistan as Wucibiekesitan Р
᜜߿‫ܟ‬ᮃഺ, which is normally said as 2 + 2 + 2. The Uzbek language uses
four syllables for this word, and English also uses four. Why does Chinese
use six? Apparently, this is an effort to account for all of the Uzbek conso-
nants. The desirability of covering all the consonants seems to outweigh
the question of syllables, even when an awkward “six” is the result. There
are other cases where “six” might be avoided but is not. For example,
Czechoslovak ia was called Jiekesiluofake ᥋‫ܟ‬ᮃ⋯Ӥ‫ —ܟ‬six syllables, and
again 2 + 2 + 2. English adds a “ia” to the end of the country’s name, and
there is no reason Chinese could not have done so as well. If it had, then
we would have the seven-syllable Jiekesiluofakeya ᥋‫ܟ‬ᮃ⋯Ӥ‫ܟ‬䲙—
easily, and pleasantly, pronounced as 2 + 2 + 3. Yet this did not happen. It
felt all right to keep the syllables at six.
Why might this be? Why should item lists tend to observe the wuyan
and qiyan preferences more than transliterations do? I wonder if the dif-
ference has to do with the cultural familiarity of item lists as opposed to
the alien flavor of transliterations. Names like Czechoslovak ia are, after
all, the creations of foreigners. Chinese speakers may feel, whether con-
sciously or not, that if such names don’t sound quite mellifluous in Chi-
nese, well, there is something appropriate about that. We Chinese should
respect the clunkiness of the Czech language; it’s not our business to be
making adjustments for others. But when we refer to oil, soy, and tea, we are
on home turf, as it were; here our intuitive sense of the “right feel” is indeed
our own business, and we deal with it.
How Recessive Rhythms Affect Structure
I have noted how the rhythms of phrases can make them seem more
graceful or authoritative and will address these “meanings” of rhythms in
a separate section below. Here I want to address the question of how
rhythms can make a difference in the structure of phrases. The point is
obvious for “dominant” rhythms (as in the limerick about No. Dak). But
for recessive rhythms, I have said earlier that they “defer to other aspects
of language” and “have ways that they work within the rules of grammar.”
Still, recessive rhythm in Chinese does affect structure in several ways.
It can, first of all, cause the addition or subtraction of syllables from a
phrase. We have seen how zhong xuexiao automatically shortens to zhongxue
‘high school’ and chu zhongxue to chuzhong ‘ju nior high’. This dropping of
syllables suits a rhythmic preference for two-syllable balance. On the
other hand, Duanmu San has shown how syllables can be added to pro-
duce balance.79 When nonheads are stressed, they must have at least as
many syllables as the head. Zhong suan ⾡㩰 ‘plant garlic’ is all right, but
zhongzhi suan ⾡ỡ㩰 is not. If zhongzhi becomes two syllables, then suan
has to become dasuan ໻㩰. The da ‘big’ is there for rhythmic balance only
and is semantically empty (i.e., the garlic does not have to be big). In modi-
fication phrases, where the nonhead comes first, the pattern reverses. Mei
dian ✸ᑫ ‘coal store’ cannot be mei shangdian ✸ଚᑫ. If you want to use the
two-syllable shangdian, then you have to say meitan ✸⚁ for ‘coal’, even
though tan (literally ‘charcoal’) does not do anything to change the mean-
ing of mei ‘coal’. Duanmu notes that “the extra syllable in the disyllabic
form is semantically redundant or vacuous.” 80
In some compounds of modern Chinese, a redundant second syllable is
added even when “nonhead stress rules” do not require it. For example, in
79. San Duanmu, “Stress and the Development of Disyllabic Words in Chinese,”
pp. 9–10, 24–25. See also Lü Shuxiang, “Xiandai Hanyu dan shuang yinjie wenti chu
tan” ⦄ҷ∝䇁ऩঠ䷇㡖䯂乬߱᥶ [A preliminary study of the problem of mono- and
bisyllabic expressions in modern Chinese], Zhongguo Yuwen [Chinese language] 1
(1963), pp. 11–23; Duanmu, Phonology of Standard Chinese, pp. 140–142; and Feng
Shengli, “Facts of Prosodic Syntax in Chinese,” unpublished paper, pp. 3–7.
80. San Duanmu, “Stress and Development of Disyllabic Words in Chinese,” p. 3.
Rhythm 69
classical Chinese both yi 㸷 ‘upper clothing’ and shang 㻇 ‘lower clothing’

could be used independently, but in modern Chinese shang cannot stand
alone. It is useful only as a second syllable to tack onto an yi in order to
make yishang ‘clothes’.81 Similarly, hu 㰢 ‘tiger’ is conventionally prefixed
with lao 㗕 ‘old’ to make laohu ‘tiger’, even though lao has nothing to do
with the age of the tiger. A baby tiger is a xiao laohu ᇣ㗕㰢 (literally “young
old tiger” but actually just “young tiger”).
The addition of redundant or vacuous syllables to accommodate cultur-
ally favored rhythms such as wuyan and qiyan seems also to happen. We
saw the example rexin xian shehui, zhenqing ai renjian ‘dedicate yourself to
society with ardor, sincerely love humanity’. The final renjian ‘humanity’
could as easily have been the single syllable ren if terseness had been the
main concern. But obviously wuyan rhythm was more important. In zhang-
lang, si guangguang! ‘cockroaches— dead to the last one!’, why is the second
guang there? Isn’t si guang ‘dead with none remaining’ dead enough? Why
make it si guangguang? Might the second guang be there to underscore the
thoroughness of the annihilation? Perhaps. Is it there to make the rhyme
with zhanglang better (completing a pleasant AABAA) pattern?82 Perhaps.
But it can also be viewed as a redundant fifth syllable put there in order to
complete a wuyan pattern.
In some cases it seems that more than one syllable is added in order to
complete a rhythmic pattern. In 1972, when U.S. journalists were allowed
into China with the Nixon entourage, one Western reporter is said to
have asked a group of Chinese on a Beijing street where the mysterious Lin
Biao had gone, to which a young boy reportedly called out: Gerpi zhaoliang
dahaitang! றሕⴔ‫ޝ‬໻⍋ූ ‘Croaked with a belch and a fart’, that is, died.
Gerpi zhaoliang is standard slang roughly equivalent to “croak” or “kick
the bucket” in American English. One interpretation of zhaoliang, liter-
ally “catch cold,” is that it evolved from chaoliang (ᳱṕ ‘face the raf ters’),
describing the attitude of a corpse lying supine. I cannot vouch for this
view. What I want to point out here are the last three syllables, dahaitang,
literally “great sea hall.” They seem to make no semantic contribution to
81. Ibid., p. 30.

82. I am indebted to James R. Pusey for this insight. Correspondence with author.
the phrase at all. Why are they there? Somehow they contribute panache
to the phrase, and at least some of that panache, I believe, derives from
the completeness of the qiyan rhythm.
Another way rhythm can impinge on structure involves no change in
words or word order. The role of rhythm can be simply to induce the lis-
tener or reader to interpret a string of syllables in one way instead of an-
other. Consider, for example, the five syllables ni qu wo ye qu Դএ៥гএ,
literally “you go I also go.” In one rhythm this means “You are going and
so am I.” In another it means “If you go, I’ll go, too.” Implied conditionals
that have no markers except stress pattern and context are much more
common in Chinese than in English. Here is another example, picked
from a myriad of other options: Zhu tao le zenme ban? ⣾䗗њᗢМࡲ? can
mean either “the pig has escaped; what should we do?” or “what should we
do if the pig escapes?”
First-year students of Chinese know that lian 䖲 and dou 䛑 combine to
express the notion of “even”: wo lian bao dou bukan ៥䖲᡹䛑ϡⳟ ‘I don’t
even read newspapers’; ta lian yidian dou buxiang chi ཌྷ䖲ϔ⚍䛑ϡᛇৗ ‘She
doesn’t want to eat even a bit [of it]’. It is common, and widely observed,
that lian can drop out of such phrases: ta yidian dou buxiang chi works just as
well as the sentence just cited. Less widely noticed is the fact that lian and
dou can sometimes both be omitted, while the notion of “even” continues to
be implied: yi dian guanxi meiyou means the same as lian yidian guanxi dou
meiyou 䖲ϔ⚍݇㋏䛑≵᳝ ‘it doesn’t matter in the slightest’. The writer
Wang Shuo has a novel entitled yidian zhengjing meiyou ϔ⚍ℷ㒣≵᳝ (Ut-
terly lacking in decency)83—which means, clearly, lian yidian zhengjing dou
meiyou, although a cool writer like Wang Shuo would never do anything
so uncool as to spell things out for dullards.
So how does a reader or listener know when to infer lian . . . dou if nei-
ther is actually used? Context clearly helps, but another reason, I believe,
is that a stress pattern distinctive to lian . . . dou usage can help. A phrase
like yi dian guanxi meiyou has a distinctive lilt. A phrase like yi wen bu zhi
ϔ᭛ϡؐ ‘not worth a penny—worthless’, condensed from lian yi wen dou
buzhi, is readily understood because it is a cliché, although theoretically
83. Wang Shuo ⥟᳨, Yidian zhengjing meiyou ϔ⚍ℷ㒣≵᳝ (Beijing: Zhongguo dian-
ying chubanshe, 2004).
Rhythm 71
one could understand it as a subject plus transitive verb (awaiting an im-

plied object): “a penny is not worth [something else].”
An imaginary example might show the point most clearly. For a sen-
tence like ta yidian mingqi meiyou le Ҫϔ⚍ৡ⇨≵᳝њ, speakers of Chinese
will naturally supply the lian . . . dou idea, apply its associated rhythm, and
understand it to mean “he completely lost his fame.” But without the lian . . .
dou idea, we can take the same eight syllables in another way. Imagine a
classroom teacher who is angry because he thinks some of his students
have been cutting class. The students say they are innocent and challenge
him to call the roll. When he calls roll and finds that, yes, everyone in-
deed is present, his anger completely subsides, that is, ta yi dianming, qi
mei you le ‘as soon as he calls roll, anger is no longer there’. This second
interpretation is actually more parsimonious than the first, because it re-
quires no importation of understood items. It also has a very different
rhythm from the first. In spoken language, there would be no ambiguity at
all, because rhythm alone—without even the help of context—would de-
cide the matter.
In rare cases, three or even four different interpretations of a phrase
can result from rhythmic variation. In a short novel by Gao Yubao enti-
tled Gao Yubao, which was famous in China in the late 1950s, a man shouts
for others to get out of bed and go to work because dawn has arrived.84 He
presses his point by saying ji dou jiao le 叵䛑িњ! ‘it’s cock-crowing time
already!’ and everyone knows what he means. The same four syllables
could, though, be read in different rhythms to mean, variously: (1) All the
chickens have crowed; (2) Even the chickens have crowed; (3) The chicken
even squawked (because, for example, something hurt it).
At their most “pushy,” recessive rhythms sometimes override consider-
ations of grammar or meaning. In this they resemble what I have called
dominant rhythms, although the crucial difference that remains between
recessive and dominant rhythms is that speakers and listeners are usually
aware of dominant rhythms, while recessive rhythms almost always
work their effects unnoticed. A simple example is how the preference in a
four-syllable string for 2 + 2 balance can intrude in a phrase. A gongsi jingli
݀ৌ㒣⧚ ‘company manager’ is 2 + 2, both grammatically and rhythmically.
84. Gao Yubao, Gao Yubao (Beijing: Zhongguo qingnian chubanshe, 1955) p. 110.
A fu zongjingli ࡃᘏ㒣⧚ ‘deputy to the general manager’ is 1 + 3 gram-

matically (or, at a finer level, 1 + [1 + 2]). We might expect the phrase to be
pronounced 1 + 3, but normally it is not. What people actually say for fu
zongjingli is fuzong/jingli, not fu/zongjingli.85 It just feels better. By stretch-
ing logic, one might conceive the phrase as 2 + 2, that is, fuzong ‘deputy
general’ + jingli ‘manager’. But even the wildest logic cannot explain a
case that Y. R. Chao gives to show how the 2 + 2 rhythmic preference can
superimpose itself on other structures. Chao’s example is wufeibingniu
⛵㚎⮙⠯, which reasonably clusters as either wu feibingniu ‘there are no
cattle with pneumonia’ or wufeibing niu ‘cattle that are free of pneumonia’.
To Chao’s astute ear, however, people still tend to say 2 + 2, wufei bingniu,
as if they mean “sick cattle without lungs.”86
In some cases a recessive rhythm can become assertive enough that it
alters the number or order of syllables in a phrase. Examples can be found
among phrases of a type sometimes used in Chinese that we might call
“false verb-object compounds.” These phrases are distinctive and delight-
ful even though they seem to have no formal name. I mean ones like majie
偖㸫 ‘scold the street’, that is, shout abuse in public, or yangbing ‫⮙ݏ‬
‘nourish illness’, that is, nourish oneself back from illness. In such phrases
the noun in the object position does not name the receiver of the action
but something else. I suspect that these interesting phrases come about in
part because of the allure of 1 + 1 or 2 + 2 rhythm. Majie must have evolved
from something more unwieldy, like zai jieshang ma ೼㸫Ϟ偖—but 1 + 1
was neater and caught on. In yangbing, what one nourishes is not the ill-
ness but more nearly the opposite, yet a phrase like yang kangbing yinsu
‫ݏ‬ᡫ⮙಴㋴ ‘nourish factors that resist illness’ is awkward and hard to put
into pleasant rhythm. Chang dagu ଅ໻哧, literally “sing the big drum,”
omits the important word shu к ‘stories’ (dashu ໻к ‘big stories’—meaning
stories about major figures in history— are what actually are getting sung)
and does not make clear that da here does not modify gu (the drum in fact
85. Feng Shengli, “Natural Foot,” p. 46.

86. “Zhongwenli yinjie gen ticai de guanxi” Ё᭛䞠䷇㡖䎳ԧ㺕ⱘ݇㋏ [How sylla-
bles affect meaning in Chinese], in Lishi yuyan yanjiusuo jikan ग़৆䇁㿔ⷨお᠔䲚ߞ
[ Journal of the research institute for historical language] (Taipei: Taiwan zhongyang
yanjiuyuan, 1968), vol. 40, part 1.
Rhythm 73
is small). Instead, all of the complexities behind the phrase bow to a 1 + 2

rhythm that conforms with the nonhead stress rule. Modern terms follow
rhythmic conventions in similar ways. Dasao weisheng ᠧᠿि⫳ ‘sweep
sanitation— clean up’ would need many more syllables to spell out properly
what is involved. But the verb-object compound, with 2 + 2 balance, feels
good.
Wuyan and qiyan patterns, used “recessively,” can influence syllable ar-
rangement as well. Duanmu San offers the example xiangxin mixin,
xiangxin gui Ⳍֵ䗋ֵⳌֵ儐 ‘believing in superstition and believing in
ghosts’ and notes that the phrase does not work nearly as well the other
way around: xiangxin gui, xiangxin mixin is awkward, and a preference for
the 2 + 2 + 3 rhythmic pattern is what makes the difference.87 In this case
it is easy to switch mixin and gui, so there is no cost, as it were, for accom-
modating the rhythm. But in other cases such costs do exist, and this fact
shows how subtle the power of a recessive rhythm can be. For example,
the political slogan reai renmin reai dang ⛁⠅Ҏ⇥⛁⠅‫ ܮ‬obeys the
2 + 2 + 3 structure of qiyan and sounds good for that reason. Normally,
however, there is a strong preference in the official language of Commu-
nist China to put dang before renmin (as in, e.g., Shaanbei dang yu renmin
datuanjie 䰱࣫咼㟛Ҏ⇥໻೬㌤ ‘the great unity of the Party and the people
in Shaanbei’). In placing renmin before dang, the phrase reai renmin reai
dang has flouted a political rule. The advantage of preserving 2 + 2 + 3
rhythm has outweighed the importance of putting the Party before the
people. (Qiyan, one might say, has been able to achieve what Chinese demo-
crats have not.)
Recessive qiyan can alter even the grammar of a phrase. When Chinese
Premier Zhou Enlai died in January 1976, headlines in many Chinese news-
papers read renmin zongli renmin ai Ҏ⇥ᘏ⧚Ҏ⇥⠅. The straightforward
way to say “the people loved the people’s premier” would, of course, be
the other way around, renmin ai renmin de zongli, but the latter just doesn’t
sound as good. For headline writers, the importance of recruiting a
2 + 2 + 3 lilt seems to have outweighed the importance of standard subject-
verb-object word order. Even Chinese-language teachers have been known
to compromise grammar for the sake of rhythm. In the 1990s an American
87. Lu and Duanmu, “Rhythm and Syntax in Chinese,” p. 127.

university announced a study program in China under the slogan xuexi

Hanyu zai Beijing ᄺд∝䇁೼࣫Ҁ ‘study Chinese in Beijing’. Textbooks in
the program taught the standard pattern “zai [place word] + verb +
object”—for example, zai guanzili chifan ೼佚ᄤ䞠ৗ佁 ‘eat in a restaurant’.
And to follow this pattern students should have said zai Beijing xue Hanyu.
Why the shift to xuexi Hanyu zai Beijing? To reach out to English-speaking
students by imitating English word order? Perhaps. More likely, the genie
of qiyan rhythm was again at work.
How Universal Are the Preferred Rhythms of Chinese?
If all human languages have rhythms, and if the rhythms of languages fall
into more or less standard patterns, then it becomes interesting to ask
how much those patterns tend to be the distinctive products of a given
culture and how much they might derive from human experience that is
common to many or all cultures. There are prima facie reasons to take
both sides of this question.
To argue the side of cultural uniqueness, one could note that the pat-
terns I am calling wuyan and qiyan arose in China and spread elsewhere in
East Asia; or that the limerick arose in Britain and likewise spread else-
where. (By the way, it spread even to China. The Chinese poet Xu Zhimo
[1897–1931] borrowed the limerick form for his poem “Ouran” ‫[ ✊ي‬By
chance].)88 The very idea of “borrowing” such forms shows that we regard
them as distinctive cultural inventions. On the other hand, to argue for
universality, one could easily note that a tendency toward “plunk-plunk,
plunk-plunk”, or “4:4 time,” in things like chants and marches, seems uni-
versal. Occasionally, much odder similarities pop up as well. Is it just
chance, for example, that the ten-syllable lilt of ticket-sellers on Beijing
buses (xia yi zhan, nongkeyuan, mei piao mai-piao! ‘next stop is Academy of
Agricultural Sciences, get a ticket if you don’t have one!’) is the same as
the ten-syllable lilt in U.S. telephone numbers? Here there is no question
of cultural borrowing, so we might ask if there is something about the
88. See Julia Lin, Modern Chinese Poetry: An Introduction (Seattle: University of
Washington Press, 1972), pp. 103–104.
Rhythm 75
human brain that just “likes” this particular rhythm. Or are such resem-
blances mere coincidence?
If there are commonalities in human preference for rhythms, where do
they come from? Historians of music have wondered whether 4:4 time
derives somehow from the physiology of the human body, such as the beat
of the heart or the two-legged walk. If so, there would be a good basis for
explaining commonality across human experience. It seems dubious that
the heartbeat could directly induce a penchant for 4:4 rhythm, since the
heart does not beat a regular two beats. But perhaps it creates a tendency
toward universal preference for some other rhythm. The two-legged walk
theory, on the other hand, may indeed provide a plausible explanation for
a preference for 4:4 among humans (and might raise, incidentally, the
unfathomable question of whether horses or insects would fi nd 4:4 time
congenial).
Whatever the explanation, the following regularity concerning four-
beat rhythm does seem to hold, at least for Chinese and Eu ropean cul-
tures: the more popular a form, the more likely it is to use four-beat
rhythm. Chaucer, Shakespeare, and other elite English poets often used
five-beat meter, while ditties, hymns, and other popular forms in English
commonly use four beats. In China, seven-syllable qiyan lines tend be
read 1–2, 1–2, 1–2–3 as elite poetry but turn into 4:4 time when popular-
ized. For example, lines of Shandong “fast tales” (rhythmical storytelling
accompanied by the percussion of wooden or metal clappers) appear on
paper as qiyan:89
Kan zhe / guniangr / danr zhen da! ⳟ䗭ྥ࿬㞑ⳳ໻

Pao lai / huche / ba wo ma! 䎥՚㚵ᡃᡞ៥㕉
Look at the gall of this girl! Running up to berate me with her

nonsense!
In performance, though, accompanied by the domineering beat of a clap-

per, qiyan gives way to a swinging four-beat line:
89. I am indebted to Wu Xiaoling for this and the following example.

KAN zhe GUniangr DANr zhen DA!

PAO lai HU che(zhe) BA wo MA!
(Zhe ⴔ here is a syncopated beat added in performance.)
Similarly in Henan ballad-singing (zhuizi ഴᄤ):
Gonggong / mingzi / Yan Bairui ݀݀ৡᄫಈⱒ⨲

Nüxu / ming jiao / Yan Jing’an ཇၓৡিಈ᱃ᅝ
The father-in-law’s name was Yan Bairui; the son-in-law’s name was
Yan Jing’an
Turns into:
GONGgong MINGzi YAN BaiRUI

NÜxu MING jiao YAN Jing’AN
Similarly, the example of Red Guards at Tiananmen chanting women yao

jian Mao zhuxi ‘we want to see Chairman Mao’ is formally qiyan, but when
the chanting became feverish it resolved into four-beat mode: WOmen
YAO jian MAO zhuXI. A satire of official language that circulated by text
messaging in 2006 began:
Huiyi meiyou bu longzhong de, bimu meiyou bushengli de;

Jiang hua meiyou bu zhongyao de, guzhang meiyou bu relie de;
Jueyi meiyou bu tongguo de, renxin meiyou bu guwu de . . .
Ӯ䆂≵᳝ϡ䱚䞡ⱘ, 䯁ᐩ≵᳝ϡ㚰߽ⱘ;
䆆䆱≵᳝ϡ䞡㽕ⱘ, 哧ᥠ≵᳝ϡ⛁⚜ⱘ;
‫އ‬䆂≵᳝ϡ䗮䖛ⱘ, Ҏᖗ≵᳝ϡ哧㟲ⱘ . . .
No meeting is un-magnificent, no closing ceremony un-victorious;

No address is un-major, no applause un-thunderous;
No resolution is un-passed, no heart is un-moved . . .
The piece has a clear four-beat rhythm and goes on—mesmerizingly—for

sixty lines.
Rhythm 77
In his essay “The Neural Lyre,” Frederick Turner speculates on

whether commonalties in poetic meter across cultures might have to do
not just with legs or heartbeats but with the structure of the human
brain.90 Turner fi nds that the poetic line, worldwide, varies from seven to
seventeen syllables in length (four to twenty at the extremes) and takes
between 2.5 and 3.5 seconds to pronounce (2 to 4 seconds at the ex-
tremes). He fi nds that even a mime, performing wordlessly, tends to use
a three-second “phrase.” 91 He then makes the bold claim that “the simi-
larities between metered verse in different cultures are real and . . . indicate
a shared biological underpinning.”92 Turner is a professor of English,
so we might want, ideally, to fi nd confi rming evidence in the fields of
psychology and neuroscience. In his classic 1957 article “The Magical
Number Seven, Plus or Minus Two,” psychologist George Miller showed
that the human ear and brain normally can handle discriminations of
pitch and of loudness up to a limit of seven or so, after which confusions
happen more easily than before.93 Miller fi nds, in addition, that short-
term memory seems to break down after roughly seven items, although
he argues that the underlying processes of perception and memory, and
thus the reasons for the “roughly seven” limit in the two cases, are fun-
damentally different. It is hard to say whether or not the brain’s natural
limits, as Miller views them, are causally related to the consistency of
poetic line length that Turner observes. But Turner and others feel that
they could be.
The plausibility of claims that the human brain might have inborn af-
finities for rhythm has been strengthened by studies of infants. In their
very first months of life, before they can have learned very much of arbi-
trary cultural patterns, infants have been shown to recognize rhythmical
features in human speech. Peter W. Jusczyk, in summarizing the results
of studies on infants and speech rhythms, goes beyond saying that infants
can “recognize” prosodic differences to say they are “well attuned to”
90. Frederick Turner, “The Neural Lyre,” in Natural Classicism (New York: Para-
gon House, 1985), pp. 61–108.
91. Ibid., pp. 74–75.
92. Ibid., p. 80.
93. Psychological Review (1956), 63, pp. 81– 97.
them from birth.94 Here Jusczyk’s word “attuned” reflects his claim that
there are natural tendencies in the brain at birth, and the studies he cites
seem to support this claim.
Exactly how much is inborn and how much learned in the formation of
rhythmic preferences is a complex and difficult question that is beyond
our scope here. I do wish, however, to consider one rhythmic pattern that
has appeared in such different times and places that it makes the question
of what gives rise to such patterns especially interesting. This is the syl-
labic pattern 1–2–3, 1–2–3, 1–2–3–4–5–6–7, for which I have used the ab-
breviation of “3–3–7.” It appears in shunkouliu like the following example,
from early 2007, that satirizes the way local officials in China pass false
reports upward in a bureaucracy and only pay lip ser vice to instructions
that come back down:95
Cun pian xiang, xiang pian xian, yizhi pian dao guowuyuan;
Guowuyuan, xia wenjian, yiceng yiceng wang xia nian;
Nian wan wenjian jin fandian, wenjian genben bu duixian.
ᴥ偫е, е偫ও, ϔⳈ偫ࠄ೑ࡵ䰶

೑ࡵ䰶, ϟ᭛ӊ, ϔሖϔሖᕔϟᗉ
ᗉᅠ᭛ӊ䖯佁ᑫ, ᭛ӊḍᴀϡ‫⦄ܥ‬
The village fools the township, the township fools the county, and
the fooling goes on up, right to the State Council;
The State Council, then, sends its orders back down, where, level by
level, they are dutifully read;
When the readings are over there’s a banquet for all, while the orders
are just set aside.
In this example, two 3–3–7 lines are followed by two qiyan lines. This
combination of 3–3–7 and qiyan, which is common, invites the suggestion
94. Peter W. Jusczyk, The Discovery of Spoken Language (Cambridge, Mass.: MIT
Press, 1997) p. 54.
95. “Rang Wen Jiabao nankan de shunkouliu” 䅽⏽ᆊᅱ䲒෾ⱘ乎ষ⑰ (Shunkouliu
to make Wen Jiabao uneasy) Pingguo ribao [Apple Daily] (Hong Kong), March 12,
2007.
Rhythm 79
that 3–3–7 at a deeper level is a version of 4–4–8, with a “rest” after each 3
and after the 7. Whether or not this is so, 3–3–7 remains distinctive not
only because of the distinctive lilt with which it is pronounced but because
the whole thing, the two 3’s and the 7, need to be present in order for the
pattern to feel satisfying. (We saw this fact in sensing how an item list of
six metals can sound a bit awkward while a list of thirteen can sound fine.)
The pattern 3–3–7 followed by a qiyan line occurs regularly in the chil-
dren’s nursery rhyme ni pai yi, wo pai yi, yige xiaohair kai feiji; ni pai er, wo
pai er, liang ge xiaohair diu shoujuanr, and so on. Remarkably, this pattern is
the very same that is found in an English counterpart: This old man, he plays
one, he plays knick-knack on my thumb . . . this old man, he plays two, he plays
knick-knack on my shoe, and so on.96 The striking similarity seems enough
to suggest that China might have borrowed the rhythm from the modern
West, just as it borrowed the rhythm of “Frère Jacques” to make the rhyme
“San zhi Laohu” ϝ䲏㗕㰢 (Three tigers) or “Happy Birthday to You” to
make “Zhu ni shengri kuailuo” ⼱Դ⫳᮹ᖿῖ. The presence of feiji 亯″
‘airplane’ in the Chinese rhyme shows it to be of twentieth-century vin-
tage, which seems to add plausibility to the hypothesis that some kind of
borrowing happened.
But 3–3–7 is too deep and pervasive in Chinese culture to be explained
by modern borrowing alone. As noted above, it appears in things as varied
as popular riddles, cheerleaders’ chants, and the Cultural Revolution an-
them “The East Is Red.” 97 It seems to have roots in rural culture as well as
urban. On a public wall in a mountain village in Sichuan in 2007 someone
had scrawled this protest against forcible family planning:98
Yitai sheng, ertai za ϔ㚢⫳, Ѡ㚢ᠢ

Santai, sitai— gua! gua! gua! ϝ㚢ಯ㚢—ࠂ! ࠂ! ࠂ!
Fetus one leads to birth, fetus two to tying tubes

With fetus three and fetus four— scrape! scrape! scrape!
96. A “dotted” variation occurs in another English nursery rhyme, “A-tisket, a-tas-
ket, a green and yellow basket . . .”
97. See the Introduction and earlier in this chapter.
98. I am grateful to Mao Sheng for this example.
Examples of 3–3–7 also popped up during the Great Leap Forward of the
late 1950s, when Chinese farmers were encouraged to “produce more” of
virtually everything, including poems. Here are two:99
Dazibao! Dazibao! ໻ᄫ᡹! ໻ᄫ᡹!

You xiang xingxing you xiang pao! জ‫ڣ‬᯳᯳জ‫!⚂ڣ‬
Big-character posters! Big-character posters!

They are like stars and also like cannon!
Jinhuang di, jinhuang shan, 䞥咗ഄ, 䞥咗ቅ

Jinguang shanshan ran hong tian 䞥‫ܝ‬䮾䮾ᶧ㑶໽
Golden earth, golden mountains,

Golden light glimmers through the red sky.
At the height of the Cultural Revolution, Red Guards swore fealty to Mao
Zedong with this chant:
Tou he duan, xue ke liu ༈ৃᮁ, 㸔ৃ⌕

Geming jingshen bu ke diu! 䴽ੑ㊒⼲ϡৃ϶!
Cut off my head, and let my blood flow,

But my revolutionary spirit will not go!
Evidence that 3–3–7 has premodern roots in Chinese popular culture

comes from Pekinese Rhymes, a little book published in 1896 by Guido
Vitale, an Italian diplomat.100 Its 274 pages contain children’s rhymes and
other popular verse that, Vitale writes, were “composed by illiterate people
who have no knowledge of the written language.” Vitale’s examples show
no sign of borrowing from the West, and not even much preference for
standard wuyan and qiyan patterns. There are many three-syllable lines,
99. Quoted in Xiao Yuying 㭁㚆◯, “Ping 1958 nian xin min’ge yundong” 䀩ϔбѨ
ܿᑈᮄ⇥℠䘟ࢩ [Evaluating the new folksongs of 1958], unpublished conference pa-
per, 1979.
100. Chinese Folklore: Pekinese Rhymes (Peking: Pei-t’ang Press, 1896; reprint, Hong
Kong: Vetch and Lee, 1972).
Rhythm 81
and the mix of line lengths and irregular patterns does seem to suggest the
kind of unschooled earthiness that Vitale claims. It is interesting, therefore,
that even in this context 3–3–7 pops up fairly often.101
The possibility that 3–3–7 in China is the result of modern borrowing is
definitively ruled out by the simple observation that many examples appear
as early as the Northern Dynasties (fifth and sixth centuries). A folksong
from that era (generally recognized as being borrowed from a northern,
non-Sinitic language) goes:102
Tian cangcang, ye mangmang ໽㪐㪐, 䞢㣿㣿

Feng chui cao di xian niu yang 乼਍㤝Ԣ㽟⠯㕞
Deep blue sky, boundless prairie

The wind blows, the grass bows, sheep and cattle appear
Examples of classic poetry from the Tang period and later use 3–3–7 as
well.103 Li Bai (701–762) uses it in his poem “Jiang jin jiu” ᇛ䘆䜦:
Wu hua ma, qian jin qiu Ѩ㢅侀, ग䞥㺬

Hu er jiang chu huan mei jiu ੐‫ܦ‬ᇛߎᤶ㕢䜦
Magnificent steed, resplendent fur

I call a boy to bring some good wine
As does Du Fu (712–770) in “Bing ju xing” ݉䒞㸠:
Ju linlin, ma xiaoxiao 䒞䔨䔨, 侀㭁㭁

Xingren gongjian ge zai yao 㸠Ҏᓧㆁ৘೼㝄
Chariots rumbling, horses neighing

Soldiers marching, bows readied at the waist
101. Ibid., pp. 31, 95, 150, 154–155, 159, and elsewhere.

102. Cited in Ogawa Tamaki ᇣᎱ⪄‍, “Chokuroku no uta” ᬩࢦȃ℠ “Song of
Chile River,” in Chosakushu 㨫԰䲚 [Collected works] (Tokyo: Chikuma Sh,b,),
1997), vol. 1, pp. 323–337; p. 324.
103. I am grateful to Wang Wei for pointing out the following two examples, and
many others.
Where, then, did 3–3–7 originate? The eminent Japanese Sinologist

Ogawa Tamaki has suggested that it might have spread to China in pre-
Tang times from origins in Central Asia or the Near East.104 If this is
right, then we still might count our observed similarities as a kind of
“borrowing.” We might even ask— although I have no evidence for the
hypothesis—whether 3–3–7 in the ancient Near East might explain 3–3–7
in the English nursery rhyme “This Old Man.” England was not the only
European place to come up with 3–3–7 subsequent to the era Ogawa has
studied. A 3–3–7 chant that was used in Soviet-dominated Poland (about
the Union of Polish Youth, Zwiazek Mlodziezy Polskiej, or ZMP) went
this way: My ZMP, my ZMP; reakcji nie boimy sie ‘We ZMP, we ZMP, we
fear no reactionary’.105
It is possible, of course—perhaps even likely—that 3–3–7 has had ori-
gins in different places independently. Especially when we remember
that, as noted, the addition of “rests” easily turns 3–3–7 into 4–4–8, it may
be that human bipeds in several places discovered the pattern without
needing to borrow anything. In any case, the short answer to the question
“Are rhythms culturally determined or common to human beings every-
where?” seems inevitably to be “Both.”
Do Rhythms Have Meanings?
The question of whether rhythms can have meanings raises a prior ques-
tion, itself extremely large and complex, of what we mean by “meaning.”
An adequate review of this larger question in either Eastern or Western
philosophical tradition is well beyond our scope here. But since we need at
least an implicit answer to it for the case of the “meanings” of rhythms, I
would like to address the topic here briefly. I will begin by asking what the
uses (or functions) of linguistic rhythms are. This question is easier to get
a handle on than “meaning”; once we identify some functions, it might
then be easier to estimate whether they involve “meaning”— and if so, of
104. Ogawa Tamaki ᇣᎱ⪄‍, “Chokuroku no uta” ᬩࢦȃ℠.

105. Anna Wierzbicka, “Antitotalitarian Language in Poland: Some Mechanisms of
Linguistic Self-Defense,” Language in Society 19 (1990), p. 12.
Rhythm 83
what kind and how much. (The approach seems not bad, since many mod-
ern philosophers, following Ludwig Wittgenstein, J. L. Austin, and others,
see “meaning” as pretty much amounting to “use” in any case.)
One function of rhythm in human language— so broad that it almost
need not be mentioned—is to signal that “this is human language.” The
variations of pitch, timing, and loudness that constitute “rhythm” are not
things that mechanical voice simulators can easily imitate. Despite ad-
vances in early twenty-first-century technology, simulated human voices
on telephone recordings remain easy to distinguish from human voices.
A more obvious use of stress patterns is in making distinctions. I have
noted how “Joey Davis” might change in stress pattern from 2–4–1–3 to
1–3–4–4 if Joey’s purpose becomes one of distinguishing himself from his
brother Louie. (Already, here, we see the beginning of what J. L. Austin
would call “meaning,” because the “speech acts” of (1) introducing oneself
and (2) distinguishing oneself from another are different acts with differ-
ent “meanings.”) Y. R. Chao notes that the riddle “Why do birds fly
south?—Because it’s too far to walk” would lose its point if stress were
applied to fly.106 Native speakers of English (or of Chinese, if the riddle
is translated) are easily and automatically aware of the difference the
stress makes. Chao also notes the difference between wo de mingzi shi Yue-
han ៥ⱘৡᄫᰃ㋘㗄 ‘my name is John’ (i.e., not Bill or Pete) and Yuehan
shi wo de mingzi ‘John is my name’ (i.e., not title or nickname). Here, Chao
says, stress is “a marker of the logical predicate,” which most native speak-
ers handle with no problem even though they do not notice that they are
handling anything.107 Sometimes, as I have noted, stress can help to de-
termine which of more than one grammatical structure should apply to a
string of syllables. Ni bu shuo, meiyou ren shuo Դϡ䇈≵᳝Ҏ䇈 can be, with
different stress, either “you didn’t say it; nobody said it” or “if you don’t
say it no one will.”
A considerable range of the conscious use of rhythm in human life has
to do with helping people to coordinate their activity. A single person can
sing, dance, march, chant, or haul loads, but when two, a dozen, or a hun-
dred want to do these things in unison, rhythm becomes an invaluable
106. Chao, Grammar of Spoken Chinese, pp. 79– 80.

107. Ibid., p. 79.
tool. Things would go haywire rather quickly without it, and excellence
or efficiency can depend on how tightly a rhythm is applied. A crew team
could not go as fast if a coxswain did not use rhythm. The laborers who
chant pile-driving songs outside the hotel window at the beginning of act 2
of Cao Yu’s play Sunrise are doing it, at least in part, in order to apply
more thud to the earth per unit of expended energy.108
Rhythms are also useful in aiding memory. Most people remember
rhythmic phrases much more easily than nonrhythmic ones. To confirm
this simple fact, try saying your own telephone number to yourself. If you
are an American, your phone number is ten digits long, and you probably
use the standard 1–2–3, 1–2–3, 1–2–3–4 pattern to say it. Now trying saying
the same number in some other pattern, like 1–2, 1–2–3–4, 1–2–3, 1. Just
from memory, you probably will not be able to do this. In order to use this
other rhythm, you will probably have to write the digits down and then
read them off. This shows that you remember your telephone number
only in a certain rhythm. Similarly, you probably know your nine-digit
Social Security number in a 3–2–4 rhythmic pattern, but could not easily
say it in a different pattern. Telephone and ID numbers in other countries
observe other patterns; but the fact that people use patterns to memorize
numbers appears to be the same everywhere.
Rhythm is an aid not only to personal memory but to community and
cultural memory. Because rhythms make memorization easier, they are
useful in any verbal art that is passed around orally— songs, nursery rhymes,
shunkouliu, and other such things. The rhythmic lines in these oral com-
positions are able to remain consistent over wide spans of time and place
without necessarily being written down. (Examples do admit variation,
of course, but the degree to which they remain consistent is remarkable.)
This ability to achieve consistency without recourse to writing has an
additional use in repressive societies: without paper trails, it is much
harder for authorities to fi nd and punish authors. On the Underground
Railway, the clandestine network that helped fugitive slaves escape the
American South, people passed messages by song. In China today, rhyth-
mic shunkouliu and text messages thrive in part because authors cannot
be traced.
108. Cao Yu ᳍⾎, “Richu” ᮹ߎ [Sunrise] (Singapore: Youth Book, 1966), p. 85.
Rhythm 85
But to say that rhythms are useful in remembering phrases that have
meanings is not the same as saying that the rhythms themselves— on
their own—“mean” something. Do they? Consider a hawker who uses a
distinctive rhythm to call out something like shaubing youtiao! ⚻佐⊍ᴵ
‘sesame cakes and deep-fried dough-sticks!’ In such a case, it is not just
the four syllables he chooses but the distinctive voice, including the lilt,
that does the work of communication for everyone within earshot. The
purpose of the hawker’s cry is less to inform people of a denotative mean-
ing (“I sell sesame cakes and dough-sticks”) than to remind them of a
daily routine (“Here I am again!” or “Now’s your chance to buy again!”).
Would this level of hawker’s meaning get through if he used a very differ-
ent lilt? Probably not— at least not until his customers got used to the new
lilt and associated it with him. In this sense the lilt seems to have its own
“meaning.”
It would be, though, a meaning that applies only in particular circum-
stances. The hawker’s call signals “it’s me again!” only for that hawker or
for others in a group of hawkers who have decided to employ the same
oral tag. It is a somewhat different question to ask whether a rhythm can
“mean something” regardless of who (within a much larger culture) uses
it. Are there such examples? To address this question requires, whether
we like it or not, that we have a definition of what we mean by “mean,” so
let me offer this one: for my purposes in this book, a rhythm has “mean-
ing” if and only if native speakers of a language that customarily uses the
rhythm take from a phrase that employs it a different understanding or
feeling, however slight, from what they would have taken from the same
phrase pronounced without the rhythm. Further, in order to qualify as
“meaning” in this sense, the production of an “understanding or feeling”
must be reliable, that is, it cannot be idiosyncratic (e.g., as a particular
hawker’s call) but has to be similar among the majority of speakers within
a culture or subculture. Theoreticians of poetic meter seem generally to
agree that meter itself does have “meaning” of this kind, although it is not
easy to specify what it is.109
109. Frederick Turner, for example, writes that meter “has significance,” although
its “ ‘message’ . . . is rather mysterious.” “Neural Lyre,” spp. 81– 82.
I have noted the pattern in American English that uses “low-low-low-

HIGH, low-low-low-HIGH, low-low-low-HIGH . . .” to express (1) I am
giving you a list; (2) the list is indefi nitely long; and, often, (3) it involves
some exasperation. Perhaps we can test whether the rhythm itself “means”
these things by trying it out, for example, on items that do not make sense
as a list. If I use the low-low-low-HIGH pattern to say “The cat is on the
mat, two plus two is four, the cow jumped over the moon,” it takes consid-
erable imagination to conjure a situation in which the rhythm would
“make sense.” A similar problem would arise if my list were fi nite and lis-
teners knew it. If I said, for example, that I have three children, and (in
the low-low-low-HIGH pattern), “one is named Jane, one is named Joe,
and one is named Jill,” people would fi nd me odd— and, if they thought
about it, a bit self-contradictory. The “meaning” of my rhythmic pattern
would be suggesting indefinitely many children, even though I had said up
front that there were exactly three.
Rhythms that suggest humor, like that of the limerick, are perhaps an
even better example. The words in limericks are often witty, but I want to
set that point aside. The point I wish to make is that the rhythm itself sug-
gests whimsy. We can test for meaning in the “rhythm itself” by combin-
ing the limerick form with somber content and seeing whether a sense of
incongruity emerges. For example, Macbeth’s soliloquy (“Is this a dagger
which I see before me, the handle toward my hand? . . .”) in limerick form
might be:
That dagger just hung in the air

Where it gave to Macbeth a great scare:
“What specter can such be?
Come, let me clutch thee,
To hang there like that isn’t fair!”
Incongruous? Quite. The whimsicality and the seriousness do not match.

The key question is where the whimsicality comes from. Certainly not
from the words, so it must come from the rhythm. But if that is so, then
the rhythm “means something” in the sense I have defined above. It brings
“funny,” or at least “lighthearted,” into the mix. Whether this funniness
is set by culture or, perhaps, is rooted in human nature is an interesting
Rhythm 87
but largely unfathomable question. Could there be a culture in which the

limerick rhythm suggested seriousness? Or is the bouncy lilt something
that would naturally tend to strike the human mind—any mind, any-
where— as whimsical?110 I lean toward the latter view but wonder if I am
culture-bound.
The question can be pondered for Chinese as well, where certain
rhythmic patterns, like the limerick, suggest “this is a bit funny”—
although do not do it quite so brazenly as the limerick does. The rhyth-
mic patterns in luan qi ba zao хϗܿ㊳ ‘chaotically messy’ and hu li hutu
㊞䞠㊞⍖ ‘muddle-befuddled’ are examples.111 A normal stress pattern for
four-syllable strings, apparent in transliterations like Gelunbiya હӺ↨Ѯ
‘Columbia’ or item lists like dong nan xi bei ϰफ㽓࣫ ‘east, south, west,
north’, is 1–3–2–4. Luan qi ba zao and hu li hutu, however, are closer to
3–1–2–4. Luoli luosuo ୄ䞠ୄ૚ ‘long-winded’, which also is whimsical, is
close to 3–1–3–3, which also differs from the normal 1–3–2–4. In these
phrases, the lilt itself seems to suggest that something is out of control in
a mildly comical way. If you refer to Columbia or the four directions with
this kind of comical lilt, they seem to lose some of their dignity.
There are other stress patterns in Chinese that seem to suggest “mean-
ing” in the broad sense I am using. In addressing or referring to people,
for example, a 2 + 2 syllabic balance seems to imply affectionate respect. If
I address my distinguished colleague Yu Ying-shih as Yu xiansheng ԭ‫⫳ܜ‬,
I sound a bit formal. If I address him as Yingshi xiansheng 㣅ᰖ‫⫳ܜ‬, I
sound just as respectful but closer and more affectionate. In part, of
course, this is because I am using his given name instead of his family
name. But the effect also has to do with the pleasant four-syllable rhyth-
mic balance of 2 + 2, Yingshi xiansheng. The enhanced pleasantness of 2 + 2
(compared to 1 + 2) can be seen in the fact that when a person has a one-
syllable given name then both surname and given name are used in order
to join with the two syllables of xiansheng to make a pleasant 2 + 2 balance.
110. Xu Zhimo’s use of the limerick form in poems like “Ouran” (see note 88 above)
is not a good test of the question. While “Ouran” imitates the limerick in line length
and rhyme scheme, it does not capture the limerick lilt. And if it did, it would likely
produce the same “incongruity” effect that my imaginary example of Macbeth in
limerick produces, because Xu puts serious romance into his poem.
111. I owe this insight to Feng Shengli, email message to author, October 17, 2002.
With a name like Li Rui, for example, one would say Li Rui xiansheng
ᴢ䫤‫⫳ܜ‬, not Rui xiansheng. The convention extends even to such un-
usual contexts as the top of the Communist Party of China. In Zhongguo
liusi zhenxiang Ё೟݁ಯⳳⳌ (The true story of June Fourth),112 when the
Communist elders and Politburo members address or refer to one another
informally as “Comrade So-and-so,” they invariably do it in balanced
four-syllable phrases: Xiaoping tongzhi ᇣᑇৠᖫ, Ziyang tongzhi ㋿䰇ৠᖫ,
Yibo tongzhi ϔ⊶ৠᖫ, and so on. But when the given name is only one syl-
lable, as in Li Peng or Qiao Shi, no one says Peng tongzhi or Shi tongzhi. To
do so would sound funny, even a bit disrespectful. The comrades pre-
ferred Li Peng tongzhi ᴢ吣ৠᖫ or Qiao Shi tongzhi Ш⷇ৠᖫ. In other
words, if one is to preserve the sense of affectionate respect, the principle
of rhythmic balance turns out to be more important than the question of
whether one includes the family name.
The pursuit of 2 + 2 balance in name-plus-title can run into problems
when names are unusual. Duanmu San, the name of the talented linguist
whom I have quoted several times, is an example. One-syllable given
names, like San, are fairly common, although not nearly as common as
two-syllable given names; two-syllable surnames, like Duanmu, are not
very common; and the combination of a two-syllable family name and a
one-syllable given name is highly uncommon. What happens if we want
to address Duanmu San with affectionate respect, as we could by saying
Yingshi xiansheng? San xiansheng would be three syllables and hence call
for a fourth; a fourth would normally be borrowed by adding a one-syllable
family name. But if we add the two-syllable Duanmu, we now have five
syllables—Duanmu San xiansheng. We have “jumped over” the rhythmic
2 + 2 and have something that sounds too formal to convey affection. This
is just his bad luck. Duanmu San deserves both respect and affection, but
there is no way we can use this par ticular rhythmic mechanism to express
them. A similar problem, which Duanmu himself points out,113 occurs
112. Zhang Liang, Zhongguo liusi zhenxiang Ё೟݁ಯⳳⳌ (The true story of Chi-
na’s June Fourth) (New York: Mingjing chubanshe ᯢ䦵ߎ⠜⼒, 2001), abridged and
edited by Andrew Nathan and Perry Link as The Tiananmen Papers: The Chinese Lead-
ership’s Decision to Use Force against Their Own People (New York: Public Affairs Press,
2001).
113. San Duanmu, “Development of Disyllabic Words in Chinese,” pp. 29–30.
Rhythm 89
when the familiar (and often affectionate) prefi xes lao 㗕 ‘old’ and xiao ᇣ
‘young’ are used before surnames: lao Zhang 㗕ᔉ ‘old Zhang’, xiao Li ᇣᴢ
‘young Li’, and so on are perfectly natural as long as two-syllable balance
applies. People surnamed Ouyang ℤ䱑, Duanmu ッ᳼, or Situ ৌᕦ by
(rhythmic) custom are denied the privilege of having the friendly lao or
xiao prefi xed to their names.
The question of meaning for longer and more complex rhythms— such
as wuyan, qiyan, or 3–3–7—is harder to pin down. Regardless of whether
the rhythms are dominant or recessive, what they “mean” can be as vague
as a fog. But that hardly means they are evanescent or unimportant. They
sometimes make a considerable difference, even if the difference is impos-
sible to specify. In this sense, they resemble the meaning of music. In
what follows I am going to use words like authority, naturalness, exaltation,
memorability, and finality to discuss them. None of these words is ideal; yet
none is useless, either, and that is the interesting point.
What is the difference between a line in qiyan like yi kan, er man, san
tong guo (‘first look, then go slowly, then cross’— cited in the Introduc-
tion) and an equivalent in plain language like dajia xiaoxin guo jie ໻ᆊᇣᖗ
䖛㸫 ‘everybody be careful crossing the street’? The qiyan line is smoother
and more aesthetically pleasing, to be sure; but there is something more.
With qiyan the message seems, somehow, to come from a more formal or
authoritative source. This is true even if the source of the qiyan message
(the author of the road-crossing sign) is not physically present, as the
crossing guard is. The crossing guard also possesses a definite authority,
but this derives from his or her social role, and perhaps from a uniform,
not from the rhythm of the phrase yi kan, er man, san tong guo. Frederick
Turner, in writing about what meter “means” in world traditions gener-
ally, says it gives “a sense of power combined with effortlessness.”114
In the early 1950s, Chinese demonstrators who marched in support of
China in the Korean War could enhance their tone of authority by set-
ting aside random shouts in favor of the qiyan slogan kang Mei yuan Chao
gan dao di! ᡫ㕢ᧈᳱᑍࠄᑩ ‘resist America and aid Korea to the end!’ In
this case the slogan chanters were in plain view, on the streets; but qiyan
can convey a sense of authority even when the “speaking authority” is
114. Turner, “Neural Lyre,” pp. 92– 93.

elusive or abstract, indeed as abstract as something like “nature itself” or

“just the way things are.” Who, for example, is the implied authority for
Guilin shanshui jia tianxia ‘the scenery at Guilin beats any in the world’?
Here, I would argue, the qiyan pattern still conjures a sense of authority—
more authority than would be present if the same idea were put in ordi-
nary language—but what authority is it, other than “the natural order of
things”? Turner refers to “a pleasing sense of ‘fit’ and inevitability,” which
seems to impart “mysterious wisdom.”115 Similarly, Edward Slingerland,
in studying the Book of Odes, finds that morally powerful words and actions
are those that show “effortless accordance with what is ‘proper’ or what
‘fits.’ ”116
In a modern context, this sense that words are “fitting” often amounts
to a claim that “what is expressed here is true.” This apparent claim of
truth can be laid bare if we consider rhythmic phrases that imply such a
claim but, if properly analyzed, are not so obviously true. Take, for ex-
ample, the qiyan proverb that warns that Sichuan is a rebellious place:
tianxia wei luan, Shu xian luan ໽ϟ᳾і㳔‫ܜ‬і ‘While all under heaven is
still quiet, Sichuan leads the way to chaos’. Spoken in plain language, the
claim would surely raise eyebrows. How does one know such a thing? Has
anyone actually counted the rebellions that started in Sichuan, and then
compared that number to the numbers of rebellions that arose in other
provinces? Probably not. But when qiyan takes charge, it is as if we agree to
suspend our critical faculties. Something that sounds “proper” does not get
proper scrutiny. Wuyan has a similar power. To say that little girls change
in many ways as they grow up, the Chinese proverb says nü da shibabian ཇ
໻कܿ䅞 ‘a growing girl changes eighteen times’. Eighteen? Not more,
not less? And little boys are different? None of these questions occur—
although they certainly would if someone were to make such an arbitrary
claim in ordinary language.
In his brilliant article “Symbols, Song, Dance, and Features of Articu-
lation,” Maurice Bloch has argued that formalized language—in religious
ritual, political discourse, song, and elsewhere—includes set features (one
115. Ibid., p. 77, 73.

116. Edward Slingerland, Effortless Action: Wu- wei as Conceptual Metaphor and Spiri-
tual Ideal in Early China (Oxford: Oxford University Press, 2003), p. 40.
Rhythm 91
of which is rhythm) that make the claims of what is articulated seem more
authoritative.117 “The notion of true or false or better or worse,” Bloch
writes, “has been eliminated by the way the proposition has been put.”118 It
is ironic that this sense of authority accompanies—indeed depends on—
what Bloch calls the “impoverishment” of language. Impoverishment is his
name for the fact that formal language cuts off alternate possibilities of
expression; it is in that very cutting off that its own authority arises. The
stricter the formal features are, the stronger the authority sounds. At the
extreme, as Bloch rightly observes, “you cannot argue with a song.”119
Should we go so far as to say that rhythm, in a certain sense, tells lies?
Frederick Turner almost says so. He observes that metrical poetry, by
“presenting an experience [that] gives a false impression of reality and
separates one from the harsh world” raises the question of whether “po-
etry deceives.”120 The examples I have cited in which wuyan and qiyan have
been used in advertisements of things like cosmetics, tonics, and gasoline
do suggest a deliberate manipulation in which the “meanings” of rhythms
can add up to something close to lying. But “deception” in a broader sense
(and Turner is clear on this) is not necessarily dishonorable. It includes
inadvertent deception and several kinds of self-deception in which the
deceived are happy to participate. After all, “separation from the harsh
world” can lead to transcendence, sublimity, and religious experience—
not just squalid ignorance. Scholars in several fields have pointed out that
poetry, ritual, chants, song, and other forms in which rhythm is involved
can impart a sense of rising above the world, approaching the sacred, and
even entering a trance-like state.121 Such “rising above” does not need so-
phisticated art. It requires only that someone, in fact, be moved. The quip
that “all bad poetry springs from genuine feeling,” widely attributed to
Oscar Wilde, makes a fair point.
117. Maurice Bloch, “Symbols, Song, Dance and Features of Articulation or Is Reli-
gion an Extreme Form of Traditional Authority?,” Archives Européennes de Sociologie 15,
no. 1 (1974), pp. 55– 81.
118. Ibid., p. 66.
119. Ibid., p. 71.
120. Turner, “Neural Lyre,” p. 99.
121. A classic statement of this effect is in Bloch, “Symbols, Song, Dance, and Fea-
tures of Articulation.”
When a rhythm helps to produce a sense of “rising above,” what exactly

is it that rises? The speaker? The hearer? The object referred to? In the
case of songs and rituals, we would probably think first of the hearer as
the one who is elevated. The singer of a song or the producer of a ritual
might also rise, but not necessarily, because one can imagine a priest who
is simply going through some motions, and a song, if it is recorded, can
move people long after the singer has departed. Sometimes the thing that
is elevated is the object referred to— or, more precisely, the status of that
object in the estimation of the speaker or hearer. This apparently is the
point of using wuyan rhythm to sell China Petrol: jiayou zhong shi hua,
fangxin pao tianxia ‘fill up with China Petrol and go wherever in the world
you want’ (cited above). The hearer (i.e., the potential buyer of China Pet-
rol) might or might not feel personally elevated at the thought of buying a
spiffy brand of gasoline, but the main point, surely, is to exalt the petrol
itself. In some cases the hearer (or other receiver) of the message becomes
quite irrelevant, and only speaker and object are importantly elevated.
When Red Guards chanted women yao jian Mao zhuxi ‘we want to see
Chairman Mao’ at Tiananmen, they were certainly exalting Mao, and
their own spirits just as certainly were being made high. Were there third
parties, listening and thereby also elevated? Perhaps, but they would have
been incidental, and the chanting would have gone on whether or not they
were there.
Rhythm can also contribute to making a line memorable. It can, as we
have seen, help us to remember things like telephone numbers and Social
Security numbers, but that is different from what I mean here by memo-
rability. By “memorable” here I mean worthy of being remembered—
because of the special wisdom or the historical importance of the phrase,
but also for its aesthetic value of harboring a rhythm. For example, Mao
Zedong’s phrase Zhongguo renmin zhanqilai le Ё೟Ҏ⇥キ䍋՚њ ‘the
Chinese people have stood up’ is 2–2–3 qiyan plus a particle. Franklin
Roosevelt’s famous line “the only thing we have to fear is fear itself” has
twelve syllables that divide pleasantly into clusters of 4 + 4 + 4. It is easy to
find many other examples in poetry. In order to understand the aesthetic
value of this kind of rhythm, it is no longer necessary to be concerned
with the difference between speaker and hearer (unless, perhaps, we
want to distinguish between “me now” and “my memory of then”). The
Rhythm 93
“meaning” of rhythm—be it exaltation, authority, the invocation of natu-

ralness, or whatever— can now be something that happens entirely within
a single mind.
Both the mnemonic value of rhythm and its aesthetic value contribute,
I believe, to the telling fact that we often can remember the rhythm of a
phrase while forgetting some of its words. This is what might lead us to
sing, in English, “I come from Alabama with a banjo something, some-
thing” when we cannot remember “on my knee” or, in Chinese, to say yue
luo, shenme, shenme, shuang man tian ᳜㨑⫮咐⫮咐䳰ⓓ໽ ‘the moon falls,
whatever, whatever, frost fills the sky’ when we forget wu ti ⚣୐ ‘raven
cries’ in the middle.122
Along with authority and memorability, it seems to me that sometimes
rhythm can convey a sense of fi nality or completeness— although this as-
pect of the “meaning” of rhythm is one I feel least confident about. Let’s
consider one more time Guilin shanshui jia tianxia. Doesn’t the phrase
convey the sense of ‘the scenery at Guilin beats any in the world (and
that’s the last word)’? And isn’t the 2–2–3 rhythm part of what gives it that
sense of “the last word”? Or take renren zuo hao ren, riri zuo haoshi. To me
it feels like ‘everyone be a good person, every day do good deeds (and that
will leave the world in fine shape)’. Or consider again the Great Leap slo-
gan loushang louxia diandeng dianhua ‘electric lights and telephones up-
stairs and down (and with that you’re all set)’. This kind of “fi nality” or
“completeness” message might be especially clear, I feel, with 3–3–7. I will
leave the reader to experiment with various examples of 3–3–7, but of the
ones I have cited, the cheerleaders’ chant seems to be an especially good
example of letting a rhythm make a completeness-claim:
Yi, er, san; san, er, yi ϔѠϝ, ϝѠϔ,

Yi, er, san, si, wu, liu, qi ϔѠϝಯѨ݁ϗ
Here nothing but numbers is involved, so the test of whether rhythm can
suggest “completeness” is especially pure. If the sense is there, it has to be
the rhythm, not the numbers, that delivers the result. There is no reason
122. From the famous (eighth-century) poem “Fengqiao yebo” ἧ‟໰⊞ [Night
mooring at maple bridge], by Zhang Ji ᔉ㑐.
in arithmetic for qi ‘seven’ to feel “fi nal,” but here qi comes down with a
thud that seems to say “I am the last item you would ever want to ask
about.” It is 3–3–7 that makes this happen.
What Other Formal Features Contribute to Meaning?
If rhythm alone can create differences in meaning (where “meaning,” re-

member, is taken in the broad sense I am using of making at least some
kind of difference in a received message), are there other formal fea-
tures of language—having to do, for example, with things like pitch or
parallelism—that can do this as well? There are. It is beyond our scope to
examine each in detail, but they are worth noting as we pass by, because
the ways they operate in the mind—gently effective although largely
unnoticed— are similar to the ways rhythm operates.
Tones. Here I do not mean literary tone. I mean the categories of pitch
variation in which syllables of Chinese are pronounced and of which
Mandarin Chinese is conventionally said to have four. Tones are “phone-
mic” in the sense that they change the reference of a syllable in the same
way vowels and consonants do—the same way, for example, that b and g
make the difference between bet and get. Like b and g, tones do not “mean”
anything on their own.
But I want to suggest that sometimes, subtly, they might. In choosing
names for their children, Chinese parents sometimes worry over whether
the sound of the name itself is “smooth” (especially for a girl) or “strong”
(especially for a boy), and part of the feeling of smoothness or strength
derives from patterns of tones.
Sometimes (by no means always, but sometimes) the “fourth tone” in
Mandarin, in which the voice falls rapidly, suggests authority or fi nality.
This effect is roughly similar to the “finality” effect of certain rhythms.
Consider again the example renren zuo hao ren, riri zuo haoshi ‘everyone be
a good person, every day do good deeds’. I claim above that the phrase
ends with a sense of fi nality because of its rhythm. But does the effect
perhaps also derive from the fact that shi џ ‘things, deeds’ is fourth-tone?
Let’s try, for example, reversing the two lines to get riri zuo haoshi, renren
Rhythm 95
zuo hao ren. This does not sound quite as good— or at least, not as “final.”
The two basic ideas “everybody be good” and “do a good deed daily” can
be put in either order with not much difference, and the rhythm is the
same in either case, but the feeling of fi nality or authority is stronger
when the whole thing ends with the fourth-tone shi than when it ends
with the second-tone ren (in which the voice pitch rises).
Do other aphorisms or slogans sound more authoritative if they end
with fourth-tone syllables? A 2007 sex education slogan was waichu wu-
gong yao fang ai, qianwan bie hai xiayidai ໪ߎࡵᎹ㽕䰆㡒, गϛ߹ᆇϟϔҷ
‘when you leave home to do work, watch out for AIDS; you absolutely
must not harm the next generation’.123 Here we have both qiyan as well as
rhyme among ai, hai, and dai, and all of this may contribute to a sense of
rightness or authority. But I suspect that the fact that dai is fourth-tone
also contributes. There seems to be an implied exclamation point at the end,
and it seems to come in part from the falling tone on dai. Would it be there
if a similar slogan referred, say, to “watching out for bugs” ( fang chong 䰆㰿)
and ended with “young children” (ertong ‫ܓ‬ス)? Could a second-tone end-
ing conjure the same sense of authority and finality? On the whole not, I
think.124
Consider a phrase that uses all fourth tones, like zhengque duidai di’erci
shijie dazhan ℷ⹂ᇍᕙ㄀Ѡ⃵Ϫ⬠໻៬ ‘regard World War II correctly’.
This phrase (partly because it is a string of fourth tones?) carries an asser-
tive and authoritative air, and might even seem a bit pompous. But is that
because of the content or because of the tones? What about other phrases

124. If fourth tones at the ends of phrases do indeed have an effect of lending au-
thority, one might expect that, on average, aphorisms would tend to end in fourth
tones just in order to borrow that authority. But this seems not to be the case. A survey
of all the wuyan and qiyan examples in the fi rst fifty pages of a standard modern dic-
tionary of Chinese idioms, Tang Shu ૤ᵶ, ed., Chengyu shuyu cihai ៤䇁❳䇁䕲⍋
[Dictionary of idioms and familiar expressions] (Taipei: Taiwan wunan tushu chuban
gongsi ৄ☷Ѩफ೪᳌ߎ⠜݀ৌ, 2000), shows that 27 percent of the wuyan examples
(among a total of 237) and 19 percent of the qiyan examples (among 488) end with
fourth-tone syllables. This puts fi nal fourth tones in the range of normal distribution,
or even a bit below it. So if we wish to test whether fourth tones do contribute to
“meaning” (broadly conceived), we appear to be left with the subjective task of judg-
ing whether phrases that do happen to use fourth tones tend to “feel different.”
that use only fourth tones but whose meanings are less austere, for exam-
ple, Zhao taitai zhan zai podeng shang kan dianshi 䍉໾໾キ೼⸈߇Ϟⳟ⬉㾚
‘Ms. Zhao is standing on a broken stool watching television’ or daxiang
shou pohai hou daochu luan fang pi ໻䈵ফ䖿ᆇৢࠄ໘хᬒሕ ‘after their mis-
treatment the elephants wandered around farting’? Here we may laugh.
But do we laugh only because the examples are inherently droll, or partly
because— and this is the key question—the hacking assertiveness of the
fourth tones somehow does not match the playful content of the phrases?
Do the tones keep frowning, as it were, saying “listen to me,” while the
content is saying “relax, this is fun!”—and does this incongruity contribute
to the humor? We might test the case by considering an example that uses
second tones only: youyu shichang huilai wanr piqiu 剓剐ᯊᐌಲᴹ⥽‫⧗Ⲃܓ‬
‘the squid often return to play with leather balls’. Here the content again is
playful, so we can ask whether the same mismatch between tones and mes-
sage appears. Do second tones fit playful messages better than fourth tones,
while fourth tones are a better fit with austere messages? Such tendencies,
if they do exist, are subtle and unusual. But even if they explain only why
something like renren zuo hao ren, riri zuo hao shi sounds slightly better
than riri zuo hao shi, renren zuo hao ren, they are still worth consideration.
Vowels and consonants. If tones, whose normal role is phonemic (not se-
mantic), can nevertheless suggest subtle differences of meaning, we should
ask whether other phonemes—vowels and consonants— can do so as well.
In English, for example, is it merely chance that so many verbs ending in
-ash have to do with violent motion: crash, clash, mash, bash, trash, slash,
splash, dash, lash, and so on? Does -ash itself “mean something”? Does
-umble have to do with incompetence: mumble, fumble, bumble, jumble,
tumble? In onomatopoeia, we expect sounds in language to imitate sounds
in nature, but the question here is different: whether clusters of sound can
move from word to word carry ing a bit of “meaning baggage” with them.
In Mandarin Chinese, the mo sound seems to conjure a family of mean-
ings. First-tone mo ᩌ is for rubbing something gently, like tousling a
child’s head, or feeling something lightly, like passing a fi nger over the
blade of a knife to see if it is sharp. Mo ᨽ in second tone is to rub harder
or to scrape, as in scraping the skin off your knee. Mo ⺼, also in second
tone, is rubbing hard enough to polish a marble floor, or to grind, as a
Rhythm 97
knife blade against a sharpening stone (before you gently mo ᩌ to check

on your results). Metaphorically speaking, you can mo shijian ⺼ᯊ䯈
‘grind time’, meaning “while it away,” or moren ⺼Ҏ ‘rub somebody’,
rather as we “rub someone the wrong way” in English. Third-tone mo ᢍ
is rubbing gentler again, and is used for applying cosmetics to the face or
spreading butter across a slice of bread. So does the vowel-consonant clus-
ter have a substratal “meaning” here? I am not sure that - ng at the ends of
Mandarin syllables ever “means something,” but it is an interesting fact that
wuyan and qiyan proverbs and aphorisms end with - ng syllables at a rate
significantly higher than normal.125
A small subfield of cognitive science has studied this question and
seems to have found that human beings do indeed tend naturally to asso-
ciate certain kinds of (nononomatopoetic) sounds with certain kinds of
meanings. A famous early study by Wolfgang Köhler found that people,
when presented with a softly roundish shape and a sharply angular one,
and then are given the nonsense words maluma and takete to apply to
them, prefer to call the round thing “maluma” and the pointy thing
“takete.”126 Another study got the same result using bouba and kiki.127 (You
can guess which sound matched which feeling.) Moreover, such results
seem to hold across language and cultures; similar results were found
among children in England and Tanzania.128 Why might this be? Some have
hypothesized that the tendency to associate roundness with sounds like
maluma and bouba comes from the fact that a human being needs to round
125. In the fi rst fifty pages of Tang Shu, Chengyu shuyu cihai, 41 percent of both
wuyan and qiyan examples end with - ng syllables. This compares with 20 percent in a
randomly selected page from the writing of the eminent contemporary literary critic
Liu Zaifu and 29 percent from a randomly selected page from the fiction of the fa-
mous contemporary novelist Jin Yong, Liu Zaifu sanwenshi heji ߬‫ݡ‬໡ᬷ᭛䆫䲚 [Col-
lected prose poems of Liu Zaifu] (Beijing: Huaxia chubanshe, 1988), p. 134, and Jin
Yong 䞥ᒌ, She diao yingxiong zhuan ᇘ卄㣅䲘‫[ ڇ‬Chronicle of the eagle-shooting he-
roes], ch. 4 (Hong Kong: Yuluo chubanshe, n.d.), 1:67.
126. Wolfgang Köhler, Gestalt Psychology (New York: Liveright, 1929).
127. V. S. Ramachandran and E. M. Hubbard, “Synaesthesia: A Window into Per-
ception, Thought, and Language,” Journal of Consciousness Studies 8, no. 12 (2001),
pp. 3–34.
128. R. Davis, “The Fitness of Names to Drawings: A Cross- Cultural Study in
Tanganyika,” British Journal of Psychology 52 (1961), pp. 259–268.
the lips in order to make the sounds. To say takete or kiki, the lips are
more angular. Others hypothesize that the universal tendencies are
rooted somewhere deeper in the human brain.129 There seems to be no
consensus on this question, but the field does agree that at a primitive
level, sounds themselves do sometimes suggest meanings.
Wondering whether Mandarin Chinese sounds might be guides, how-
ever subtly or marginally, to their meanings, I devised an experiment in
1996. Setting aside cases that are arguably onomatopoeia (like pa! ା for a
slapping sound or wang ∾ for the bark of a dog), I wanted to see whether
sounds might suggest meanings in a broader sense. I made a list of twenty
paired opposites in Chinese: qianhou ࠡৢ ‘front and back’, daxiao ໻ᇣ
‘large and small’, kuaiman ᖿ᜶ ‘fast and slow’, and seventeen others. I
brought the list to classes of seventh-graders at the John Witherspoon
Middle School in Princeton, New Jersey. I offered a lesson in beginning
Chinese, so the students and their teacher might feel all right about me.
Then I handed out lists in English of the twenty pairs of opposites and
read the sounds of the paired opposites in Chinese, asking the students to
guess which Chinese sound “sounded like” which meaning. For example,
the students looked at the words “front” and “back” while I read the
sounds qian and hou; then they used their pencils to mark on their sheets
whether “the first sound” meant “front” and “the second sound” meant
“back,” or vice versa. I flipped a coin twenty times, in advance, to deter-
mine the order in which I would read the paired words in Chinese, so that
customary patterns in the orders of such things (which tend to be the
same in the two languages, e.g., we conventionally say front before back in
English and qian before hou in Chinese) theoretically would be neutral. I
also requested that any student who knew any Chinese at all not partici-
pate in the fun. By chance, there were exactly one hundred students par-
ticipating. The table on the next page gives the results.
An average number of right guesses only slightly higher than 50 per-
cent is not a very strong basis for claiming that sounds themselves have
129. Ramachandran and Hubbard, “Synaesthesia”; Daphne Maurer, Thanujeni

Pathman, and Catherine J. Mondloch, “The Shape of Boubas: Sound-Shape Corre-
spondences in Toddlers and Adults,” Developmental Science 9, no. 3 (2006),
pp. 316–322.
Rhythm 99
Results of my 1996 experiment
Correct Incorrect No guesses

guesses guesses (left blank)
Qianhou ‘front and back’ 61 36 3

Daxiao ‘big and little’ 49 49 2
Kuaiman ‘fast and slow’ 33 67 0
Gaoai ‘tall and short’ 50 49 1
Haohuai ‘good and bad’ 58 41 1
Shangxia ‘up and down’ 54 46 0
Duoshao ‘many and few’ 37 62 1
Qingzhong ‘light and heavy’ 90 10 0
Ruanying ‘soft and hard’ 44 56 0
Yuanjin ‘far and near’ 64 36 0
Maimai ‘buy and sell’ 52 48 0
Baoe ‘full and hungry’ 24 76 0
Baohou ‘thin and thick’ 70 30 0
Xinjiu ‘new and old’ 50 50 0
Fangyuan ‘square and round’ 60 40 0
Meichou ‘beautiful and ugly’ 77 23 0
Zaowan ‘early and late’ 46 53 1
Shenqian ‘deep and shallow’ 14 86 0
Gaodi ‘high and low’ 43 57 0
Qingchu/hunzhuo ‘clear and 86 14 0
muddy’
Average 53.1% 46.4% 0.5%
meanings. I do find some of the particular results interesting, however.

The pair that was guessed most accurately was qingzhong ‘light and heavy’,
and a correct-guess rate as high as nine out of ten strongly suggests that
there might be a reason. Does the i sound in qing 䕏 sound “light”? Or
perhaps the palatal initial q- contributes to this feeling? Might it be that
the fourth (falling) tone on zhong 䞡 contributes to a “heavy” feeling? On
similar grounds, it seems natural that the students guessed well (86 per-
cent) on qingchu ⏙Ἦ for ‘clear’, leaving hunzhuo for ‘muddy’. Some of the
examples on which the guesses were worst seem to raise the question
whether, despite being wrong, the guesses may have made some sense. A
student commented to me afterward, for example, that he thought e 体
‘hungry’, with its falling tone, sounded like being overstuffed and about
to vomit—rather the opposite of hungry. And shen ⏅ ‘deep’ and qian ⌙
‘shallow’ might have misled students because of the tones. Shen has a first
(high) tone and qian a third (low) tone—the opposite of what one would
expect for ‘deep’ and ‘shallow’.
Another sense in which sound itself can have meaning seems to be es-
pecially important in Chinese. This effect happens when one meaning
carries a second meaning with it— only because of a sound. Any human
language has homonyms, and sometimes homonyms create “double mean-
ings.” Puns are a common example, and Chinese has plenty of puns. But I
am not speaking here of puns. In a pun, the whole point is that a single
sound has two different meanings that cannot carry over from one to the
other. What makes a pun funny is precisely the “fraudulent claim,” as it
were, that one thing can be another. For example, when the famous Chi-
nese xiangsheng performer Hou Baolin says that he is a zuojia ԰ᆊ ‘author’
because all day he zuo jia തᆊ ‘sits at home’, no one thinks that sitting at
home really does make him an author. The sound zuojia does not have
this power. But in another function of “double meaning” that is very com-
mon in Chinese culture, sound itself is assumed to carry meaning along
with it. No sense of fraud is involved, and nothing is funny. Quite the
contrary, the topic is often serious. At New Year’s, for example, Canton-
ese eat facai 僂㦰‘hair vegetable’, a stringy black vegetable that looks a bit
like hair but—importantly, at New Year’s—is a near homonym of facai ⱐ
䉵 ‘get rich’ ( fat choi in Cantonese). It is also good to eat yu 剐 ‘fish’ at New
Year’s, because yu 们, also second tone, is ‘bounty’. When my Princeton
colleague James Wei, a dean of engineering and a brilliant man, went to
grade school in Shanghai in the 1930s, his mother put onion (cong 㩅),
garlic (suan 㩰), and chicken hearts ( ji xin 䲲ᖗ) in his lunch box. This was
to be sure that he would be smart (congming 㙄ᯢ), be good at mathemat-
ics (suanshu ㅫ㸧), and have a good memory ( jixing 㿬ᗻ). Terese Bar-
tholomew has published a 350-page book on more than three hundred
examples of this kind of meaning-transfers, which she calls “hidden mean-
ings,” that regularly appear in Chinese art and are meant to help people
achieve happy marriage, wealth, longevity, official career, and passing of
exams.130 The care with which Chinese parents traditionally choose their
children’s names sometimes also includes this kind of calculation. The
130. Terese Tse Bartholomew, Hidden Meanings in Chinese Art (San Francisco: Asian
Art Museum, 2006).
Rhythm 101
Anhui novelist Chen Dengke 䰜ⱏ⾥, for example, has a fi ne name that
suggests “Chen who rises in the examination system.” A family named
Wei 儣, however, would likely avoid such a choice, lest the name suggest
“has not passed the exams” ᳾ⱏ⾥. The exiled Chinese political philoso-
pher Hu Ping 㚵ᑇ has a son named Hu Pan 㚵⬨; when a daughter was
born, he and his spouse were considering Hu Sha 㚵㥢 for her name but
later thought better of it. If the two children were attracted to legal careers,
they would be at a big disadvantage: one hupan 㚵߸ ‘crazily sentencing’
and the other husha 㚵ᴔ ‘crazily killing’.131 Similarly, in moving to an
English-speaking country, a Chinese family might, if they could manage
it, avoid a dentist named Dr. Paine or a surgeon named Dr. Slaughter.
Pitch. We have already seen examples where voice pitch (which combines
with loudness and duration to constitute what I have defined as “stress”) is
useful in making distinctions of meaning. When we emphasize that Joey
Davis is not Louie Davis, pitch is part of what makes the difference. When
we distinguish between birds flying south and birds flying south, pitch again
pitches in. In English, pitch differences can convey skepticism or surprise.
In “Where did you go?” the voice normally falls at the end; but in “You went
where?” it rises. In “Would you like coffee or tea?” if the cof- in “coffee” is
spoken in a relatively high pitch and tea in a relatively low one, the meaning
is ni yao kafei haishi cha? Դ㽕੪ଵ䖬ᰃ㤊? ‘which would you like— coffee
or tea?’. But if the last four syllables are said with relatively high and rising
pitch, the meaning is ni yao buyao kafei huozhe cha? Դ㽕ϡ㽕੪ଵ៪㗙㤊?
‘would you like either coffee or tea [yes or no]?’ In the fall of 2009, after two
sets of tennis doubles, one of my fellow players asked the rest of us if we
would like to play a third set. Two people answered “I’m good” but meant
exactly opposite things by it. One used a pitch contour that resembled the
Mandarin Chinese first tone (high level) followed by a second (rising)
tone; this meant “yes, bring it on.” The other used two falling pitches,
like the fourth tone in Chinese; this meant “no, I’ve had it.” No one was
puzzled by the two answers. Both were perfectly clear in meaning. In fact,
no one (except me, later) seemed to notice that the same two syllables had
been used and that it was pitch alone that had made the difference.
131. Hu Ping, email message to author, November 23, 2007.

Pitch in Chinese (in addition to its normal role in tones) does many
similar things, and we cannot address all of them here. But, for example,
the particle lou that gets added to the ends of Cantonese utterances to
mean something like “of course!” or even “only an idiot would need to be
told this!” is consistently said in a high pitch. In Mandarin, skeptical ques-
tions employ a higher-than-normal pitch, as they do in English. This holds
even when a “falling tone” (fourth tone) appears in a question that wants a
higher-than-normal pitch. For example shi ma? ᰃ৫? ‘is that right?’ is said
in higher-than-normal pitch but still allows, within the elevated pitch
range, for a falling pitch on the fourth-tone syllable shi. In skeptical ques-
tions of greater length, like nandao ta shuo de dui ma? 䲒䘧Ҫ䇈ᕫᇍ৫?
‘Could it be that what he said was right?’ a higher pitch runs through sev-
eral syllables at the end of the phrase, but again allows room for fourth-
tone syllables (dui, in this case) to fall in pitch appropriately. Y. R. Chao has
used the image of ripples riding on waves to describe this phenomenon;
Chao observes that what the voice actually does is an “algebraic sum” of
the two effects.132
George Lakoff and Mark Johnson, in their brilliant little book Meta-
phors We Live By, assert that “tone languages generally do not use intona-
tion to mark questions at all.”133 The error is glaring, yet the notion that
seems to have led to it—that languages very different from English might
follow different customs in intonation—is important. In English the voice
rises at the end when we say “you mean this one?” or simply “this one?”
But by custom in northern Mandarin Chinese, the voice drops in saying
the corresponding phrase zhei g’a? 䖭Ͼ䰓?134 For another example, in En-
glish the voice drops on the last syllable of the phrase “what do you think?”
It would be very easy to use the same pattern in Mandarin, where a ready
match is available in the combination of first tone followed by neutral tone
in a phrase like women ting le ៥Ӏ਀њ ‘we listened’. But, despite its easy
availability, Mandarin does not use this ready-made pattern in phrases that
correspond to “what do you think?” In ni shuo ne? Դ䇈ਸ਼? ‘what do you
say?’ the final ne does not drop but stays high, about where shuo is.
132. Chao, Grammar of Spoken Chinese, p. 39.

133. George Lakoff and Mark Johnson, Metaphors We Live By (Chicago: University
of Chicago Press, 1980˅, p. 138.
134. Yuen Ren Chao, Mandarin Primer (Cambridge, Mass.: Harvard University
Press, 1948), lesson 2, n. 3, pp. 129–130.
Rhythm 103
A feature that is sometimes related to pitch in Chinese intonation is

elongation of syllables. If a ner vous mother in China asks a teenage
daughter if she has ever driven long distance before, the daughter might
answer k-a-a-a-i-i-i-i guo! 䭟——䘢! ‘y-e-e-s-s-s-s-s, I have’, where the
drawing-out of kai emphasizes the point in a way designed both to reas-
sure the mother and to suggest that she needn’t have asked.135
The full picture of intonation in Chinese is a large and complex topic.
The purpose here is only to illustrate it briefly and to note its relation to
the question of rhythm.
Parallelism. Just as rhythm can produce senses of naturalness, authority,

or “fit,” it seems clear that parallelism can do the same. Indeed parallelism
and rhythm— as classical Chinese poetry richly shows— often work to-
gether. As with rhythm, the success of parallelism in producing meaning
in Chinese seems attributable, in part, to the flexibility of the single-
syllable morphemes that compose Chinese.
Consider again the four-syllable Great Leap Forward slogan li gan jian
ying ‘erect a pole and see a shadow—get instant results’. Even in such a
simple example, the grammatical parallelism of verb-object followed by
another verb-object seems (at least to me) to suggest a sense of “rightness.”
I have also noted the Great Leap slogan loushang louxia, diandeng dianhua
‘electric lights and telephones upstairs and down’. Here the double parallel-
ism of lou-A lou-B, dian-X dian-Y is, to my ear, even more effective in sug-
gesting “fittingness.” The terseness of the phrases enhances the power of
their parallelism, and the result is nifty penetration. What could possibly
be better, as it were, than loushang louxia diandeng dianhua?
The power of parallelism to create a sense of fit appears in sharper relief
in examples where the words themselves are relatively empty. Take, for
example, the Cultural Revolution slogan zhua geming, cu shengchan ᡧ䴽ੑ,
֗⫳⫷ ‘seize revolution, stimulate production’. What exactly did it mean,
in daily life, to zhua revolution? Everyone at the time knew that this was a
good thing to do, but what exactly did it mean? (Work harder? Obey
your leader? Pick on your politically tainted neighbor?) “Promote produc-
tion,” although somewhat clearer, is still fairly general and abstract, as an
135. The example is borrowed from Ta-tuan Ch’en et al., Chinese Primer (Princeton,
N.J.: Princeton University Press, 2007), vol. 1, p. 85, and vol. 2, p. 183.
instruction. But note: when the two phrases are put side by side, somehow
the grammatical parallelism (verb-object) and rhythmic parallelism (1–2,
1–2) generates a tight sense of fit. The phrase, if not its meaning, is lean
and neat. And it exudes “rightness.”
In poetry and many kinds of popular phrases, the effects of parallelism
are enhanced when parallel syllables are semantically antithetical. A tra-
ditional shunkouliu says:
Hai ren zhi xin bu ke you ᆇҎПᖗϡৃ᳝

Fang ren zhi xin bu ke wu 䰆ҎПᖗϡৃ⛵
The will to harm others you must not have

The will to resist others you must not lack
Here we have both qiyan and an unusually close parallelism: five middle
syllables are identical, while the framing syllables at either end are gram-
matically parallel and semantically antithetical. In elite poetry, such fea-
tures might seem only the technical features of a genre— or even, if they
are overdone, ostentatious wordplay. In popular culture, though, I believe
the effect has been somewhat different. I believe it creates an aura of dig-
nity and credibility about what is expressed. The message becomes right,
natural, fitting, wise.
This tradition has been strong enough to persist through China’s Com-
munist revolution and into the market economy of today. In the early
years of the revolution, a slogan aimed at preserving forests was fandui
langfei mucai, fandui lanfa linmu ডᇡ⌾䊏᳼ᴤ, ডᇡ▿Ӥᵫ᳼! ‘oppose the
waste of lumber; oppose rampant cutting of forests’, in which the gram-
matical parallelism is reinforced by the repetition of fandui as well as by
the quasi rhyme of langfei and lanfa.136 A campaign against corruption and
waste used a slogan that comprised three “verb plus object” phrases in
parallel: qingli zichan, heding zijin, fandui langfei! ⏙⧚䊛⫶, Ḍᅮ䊛䞥, ডᇡ
⌾䊏 ‘clean up property, check on ownings, oppose waste!’137 By 1966, Mao
Zedong had decided that “old culture” should be thrown out, but he still
136. Renmin ribao Ҏ⇥᮹ฅ [People’s daily], October 23, 1951, p. 2. The unusual six-
syllable rhythm is worth noting here.
137. Ibid., October 28, 1951, p. 2.
Rhythm 105
used parallelism. A photograph in the People’s Daily in fall, 1966 shows Mao
waving in front of a large billboard that reads:
Dapo yiqie boxuejieji de jiu sixiang, jiu wenhua, jiu fengsu, jiu xiguan!
Dali wuchanjieji de xin sixiang, xin wenhua, xin fengsu, xin xiguan!
໻⸈ϔߛ࠹ࠞ䰊㑻ⱘᮻᗱᛇ, ᮻ᭛࣪, ᮻ亢֫, ᮻдᛃ!

໻ゟ᮴ѻ䰊㑻ⱘᮄᗱᛇ, ᮄ᭛࣪, ᮄ亢֫, ᮄдᛃ!
Demolish the old thinking, old culture, old customs, and old habits
of all the exploiting classes!
Establish the new thinking, new culture, new customs, and new hab-
its of the proletariat!138
The two phrases are semantically antithetical in five places and, except for
yiqie ‘all’, are exactly parallel as well. The power of the lines clearly derives
in part from the parallelism, and it seems likely that Mao and his follow-
ers appreciated the power without noticing its old-culture roots. Essentially
the same power of parallelism extends into the commercial culture of the
post-Mao era, for example in an advertisement for White Beauty Soap (baili
xiangzao ⱑБ佭ⱖ) that appeared on Shanghai television in 2000: shi toufa
gengen rouruan, ling pifu cuncun nenhua Փ༈থḍḍᶨ䕃, ҸⲂ㙸ᇌᇌႽ⒥
‘makes each one of your hairs soft, makes every inch of your skin smooth’.
The seven-syllable rhythm here is an unconventional “dotted” 4:4 pattern,
but the parallelism is impeccable from start to finish. Similarly, a warning
to motorists not to drink and drive that was posted on public blackboards in
Beijing and elsewhere in 2007 read hejinqu ji di meijiu, liuchulai wushu xuelei
ୱ䖯এ޴Ⓢ㕢䜦, ⌕ߎᴹ᮴᭄㸔⊾ ‘imbibe a few drops of fine wine, shed
countless more in blood and tears’. My English translation does not fully
capture either the rhythm or the parallelism of the original, whose impact
is to convey a sense something like: “this can’t be wrong.”
Although the “meaning” that rhythm, parallelism, and semantic anti-
theticality can conspire to produce has especially strong cultural roots in
Chinese popular culture, there is obviously something universal about it
as well. It also appears, if perhaps not as frequently, in English-speaking
138. Ibid., October 21, 1966, p. 4.

cultures, for example in the famous lines in John F. Kennedy’s 1961 inau-
gural address “United there is little we cannot do, divided there is little
we can do” (in which the combination of rhythm, parallelism, and anti-
theticality could not be more “Chinese”) or his statement that if a free
society “cannot help the many who are poor, it cannot save the few who are
rich.” Why are such phrases effective? Do the rhetorical features them-
selves not make the messages seem more right, natural, and fitting than
they otherwise would? Whether in Chinese or English, rhyme, rhythm,
parallelism, and antitheticality, separately or in concert, can convey the
sense that “this is the real truth” or “this resonates with nature.” An ade-
quate account of these rhetorical devices, in the history of either Chinese
or Indo-Eurpoean languages, is beyond our scope here.
Chiasmus. A distinctive version of parallelism, sometimes called

“chiasmus,”139 combines rhythmic and grammatical parallelism with se-
mantic inversion. “Eat to live, don’t live to eat” is an example, and it is
easy to find others, in both Chinese and English, from the twentieth and
twenty-first centuries. Kennedy’s “ask not what your country can do for
you; ask what you can do for your country” in his inaugural address is a
famous example, as is Mao Zedong’s Cultural Revolution slogan
Jiefangjun xue quanguo renmin 㾷ᬒ‫ݯ‬ᄺܼ೑Ҏ⇥

Quanguo renmin xue jiefangjun ܼ೑Ҏ⇥ᄺ㾷ᬒ‫ݯ‬
The Liberation Army should learn from all the people of the nation
All the people of the nation should learn from the Liberation Army.
Intellectuals have used chiasmus as well. In his 1918 essay “On Establish-
ing a New Literature,” Hu Shi distilled his main message into the phrase
guoyu de wenxue, wenxue de guoyu ೟䁲ⱘ᭛ᅌ, ᭛ᅌⱘ೟䁲 ‘literature in the
national language and a national language’s literature’.140 More recent
139. See Mardy Grothe, Never Let a Fool Kiss You or a Kiss Fool You (New York: Vi-
king Books, 1999). Grothe also has a website, www.drmardy.com, accessed June 23,
2012.
140. Hu Shi 㚵䘽, “Jianshe xin wenxue lun” ᓎ䀁ᮄ᭛ᅌ䂪 [On building a new litera-
ture], quoted in “Jianshe de geming wenxue lun (I)” ᓎ䀁ⱘ䴽ੑ᭛ᅌ䂪 (I) [On a con-
Rhythm 107
scholars have given their essays titles like “Shehui de xianshi he xianshi de
shehui” ⼒Ӯⱘ⦄ᅲ੠⦄ᅲⱘ⼒Ӯ (Social reality and present-day society),
“Cong wenxue de chengshou dao chengshou de wenxue” Ң᭛ᄺⱘ៤❳ࠄ
៤❳ⱘ᭛ᄺ (From the maturation of literature to literature that is ma-
ture), and “Beiai de zhishifenzi he zhishifenzi de beiai” ᚆઔⱘⶹ䆚ߚᄤ੠
ⶹ䆚ߚᄤⱘᚆઔ (Sorrowful intellectuals and the sorrow of intellectuals).141
A freshman seminar at Princeton University in 1999 was named “De-
forming Codes, Decoding Forms”; around the same time, a bon mot from
Toni Morrison appeared on the wall of the Princeton student center:
“The Place of the Idea, the Idea of the Place.”
The “value added” of chiasmus is a bit mysterious. What is it, and what
does it come from? We might test the question by separating the two
halves of certain examples, thereby destroying whatever additional effect
is caused by the pairing of lines. Each phrase in “ask not what your coun-
try can do for you” and “ask what you can do for your country” can make
sense independently, although the parts somehow seem less than the
whole. Something extra seems to emerge when the two parts are joined.
Another example where each of the two halves makes a good point inde-
pendently is the popular Chinese saying about Communist Party meet-
ings xiao huiyi zuo da jueding, da huiyi zuo xiao jueding ᇣӮ䆂‫خ‬໻‫އ‬ᅮ,
໻Ӯ䆂‫خ‬ᇣ‫އ‬ᅮ ‘small meetings make big decisions and big meetings
make small ones’, that is, important choices are made by a few people in
small rooms, and routine approvals are given in huge conference halls.
Each half of this chiasmus has something important to say, and the fact
that the two comprise the chiasmus form is a kind of bonus. On the other
hand, sometimes the two halves of chiasmus seem redundant (what does
“the sorrow of intellectuals” add to the idea of “sorrowful intellectuals”?).
Nonetheless— and this is the crucial point—juxtaposing the two seems to
make a claim that “something extra,” even something a bit mysterious or
structive revolutionary literature], www.mianfeilunwen.com /Wenxue/Hanyuyan

/31407.html, accessed June 23, 2012.
141. The fi rst two of these titles are from a conference on ᮄᯊᳳⱘЁ೑᭛ᄺ ˄Chi-
nese literature in the new era˅ at Zhongshan University, Guangzhou, 1979. The
third is from the conference “The Course of Contemporary Chinese Intellectuals: An
International Symposium Commemorating the Fiftieth Anniversary of the Anti-
Rightist Movement Campaign,” University of California, Irvine, July 29–30, 2007.
profound, is going on. In some cases, the extra element is so important

that if one takes the chiasmatic juxtaposition away, the two parts do not
stand up at all on their own. By itself, for example, either “The Place of the
Idea” or “the Idea of the Place” would look odd on the wall of a campus cen-
ter. Only when the two are side by side does it suddenly become all right to
put them there; and the result, somehow, even seems to lay claim to a special
wisdom. In complex instances of chiasmus, concrete meaning might not be
as important as the aesthetic satisfaction one feels just from seeing all the
parts coordinate like clockwork. In analyzing the late-Qing novel Niehaihua,
for example, David Wang writes that it “literally politicizes the sport of de-
sire as much as it eroticizes the game of power.”142 The reader does not need
to dig for the ultimate meaning of the phrase in order to enjoy its craft.
The “extra something” that chiasmus lays claim to is hard to pin down.
It seems to me akin to the claims of wuyan and qiyan rhythms: that there
is something right, fitting, or authoritative about the message that the
phrase bears. But in chiasmus, the claim may be even a bit stronger,
reaching toward the realm of “deep” or “mystical.” This relatively strong
claim in chiasmus is most visible when it is overdone, by which I mean
when the rightness or profundity of the message does not measure up to
what the chiasmus form seems to be claiming. For example, if a professor
is proud of his verbal integrity and says “I said what I meant and I meant
what I said,” we accept it as a successful emphatic statement; but if he is
proud of his dress and says, in parallel chiasmatic form, “I wear suits when
I lecture and I lecture when I wear suits,” we would have to say that chias-
mus had seduced him into saying something whose detailed implications
he had not considered. This is an imaginary example, but many actual
examples raise the same question in varying degrees. In 2002, for exam-
ple, U.S. congressman John Dingell summed up the behavior of Enron
executives as “either criminally stupid or stupidly criminal.”143 Exactly
what the difference is between “criminally stupid” and “stupidly criminal”
was not clear, and how one might determine which of the two was actually
the case— as the statement suggests one could do—was also murky. But I
would guess that these questions did not even occur either to Dingell or
142. David Wang, Fin- de- Siècle Splendor: Repressed Modernities in Late- Qing Fiction,
1849–1911 (Stanford: Stanford University Press, 1997), p. 106.
143. ABC News with Peter Jennings, January 24, 2002.
Rhythm 109
to many of his listeners, for whom the charm of chiasmus was likely
enough to make the statement seem not just acceptable but memorably
wise. Similarly, a campaign slogan for healthful sex practices in China in
2007 was chenggongde nanren yao jiankang, jiankang de nanren geng cheng-
gong ៤ࡳⱘ⬋Ҏ㽕‫ع‬ᒋ, ‫ع‬ᒋⱘ⬋Ҏ᳈៤ࡳ ‘successful males want health,
and healthy males are more successful’.144 Here it is not clear what cheng-
gong ‘success’ means because “success at what?” is not specified; the only
real meaning of the phrase is a vague positive valuation: whatever it is,
chenggong is “what you’d want to be.” But if that is the case, then it be-
comes banal to say that chenggong males would want to be healthy, and to
turn the phrase around to say that healthy males are more chenggong is
equally vacuous. Yet despite all these weaknesses, the charm of chiasmus
was enough to lead Chinese slogan writers to feel that the words might
have persuasive power, and they probably were right in that judgment.
Can the Users of Rhythm Be Unaware of Its Effects?
At several points in this chapter I have noted that the meanings of rhythms
(and some related devices) are sometimes not fully noticed by those who
use them. I have employed a broad definition of “meaning” by which
merely “a different understanding or feeling, however slight,” can count
as meaning even if users are not aware of just what is causing the small
difference. In one sense, though, to speak in this way is odd. Normally, we
assume that meaning implies intent. When we “mean” something, how can
we be unaware of it? How can we say that rhythms “mean” things when
people are not meaning them to do so?
But perhaps this is not so strange after all. Compare for a moment
“meaning” as it applies to words. Normally, when A says or writes a word,
like “book,” and B receives and interprets the word, we assume the fol-
lowing conditions:
1. that a meaning has passed between A and B, and

2. that both A and B might be—and often are— conscious of the
word’s role in conveying the meaning.

Neither condition ensures that A and B understand each other perfectly,

of course. A might consciously intend one thing (e.g. in saying “cook the
books!”), and B might consciously infer something else (by placing the
books into a steamer). But even in miscommunication such as this, A and
B are both conscious of a meaning for the word “book.” It is true that in
fluent speech we may not consciously notice every word as it fl ies by, and
also true that in some cases meaningful words emerge from our subcon-
scious minds (in Freudian slips, for example, or, more radically, in talking
in one’s sleep). But still, for most words most of the time, we are aware of
our word choices, and sometimes even think long and hard in choosing
words. For words, conditions 1 and 2 both apply.
For such things as rhythm and parallelism, though, condition 1 holds
while condition 2 almost never does. A poet or trained linguist— or a
reader of this book—might notice the rhythm in a phrase and be able to
interpret its contribution to “meaning,” but ordinary native speakers of a
language do not do such things. They do use rhythms, and using them
does make differences to them and to the people with whom they com-
municate. But they do not notice. Chinese schoolchildren learn characters,
memorize sentences, and are explicitly told that these are vessels of mean-
ing. They are not given tables of rhythms to memorize.
Can we say simply that rhythmic meaning is subconscious and let it go
at that? Cognitive scientists are apparently comfortable with the assump-
tion that, at a minimum, “unconscious thought is 95 percent of all
thought.”145 If that is so, why can’t rhythm just be unconscious meaning?
The main problem with such a stance is that the mind is to some degree
conscious of the various effects that rhythm produces— authority, right-
ness, whimsy (for a limerick), and so on. Even if people are not aware of the
causes of these effects, they are certainly aware of the effects. So it ap-
pears that we must say that while some aspects of what rhythms are doing
impinge on consciousness, others do not.
This situation of “partly conscious and partly not”— aware of effects
but not what causes them— seems in some ways parallel to the ways gram-
mar rules operate in the mind. Few speakers of Chinese, for example, are
aware of the very powerful rule in Chinese that modifiers—whether of
145. George Lakoff and Mark Johnson, Philosophy in the Flesh: The Embodied Mind
and Its Challenge to Western Thought (New York: Basic Books, 1999), p. 13.
Rhythm 111
nouns or verbs, and regardless of length—precede what they modify. Yet

the same adept speakers have no trouble obeying the rule or feeling con-
sciously that something’s wrong when the rule is violated. Moreover, the
force of the rule is strong enough that it can bend apparent exceptions to
its will. For example, the fluffy dried meat known as rou song 㙝ᵒ might
appear to be a noun (rou ‘flesh’) followed by an adjective (song ‘loose’), but
in the Chinese mind the modification relation still works in the standard
way: rou song is “fluff” of the “meat” variety; it is “meat fluff.” Similarly,
putao gan 㨵㧘ђ ‘raisin’, composed of putao ‘grape’ and gan ‘dry’, is not a
grammatical reversal of “dry grape” into “grape dry”. Like rou song, it too,
conceptually, is a modifier preceding a noun. A raisin is a “grape-derived
dry-thing.” The strong grammar rule, applied to the mind subconsciously,
helps to turn this trick.
Another revealing example of subconscious obedience of rules is the
common pattern in Chinese of four-syllable phrases in which two of
the syllables are numbers: yi qing er chu ϔ⏙ѠἮ ‘crystal clear’; bu san bu si
ϡϝϡಯ ‘shady (character)’ dian san dao si 乴ϝ‫צ‬ಯ ‘confused, incoherent’;
shuo san dao si 䇈ϝ䘧ಯ ‘speak irresponsibly’; wu yan liu se Ѩ买݁㡆 ‘var-
ied in color’; luan qi ba zao іϗܿ㊳ ‘utterly messy’; and others. There are
some exceptions, but such phrases tend to observe these two rules:
1. the two numbers are adjacent integers (n and n + 1), and

2. of the two integers, the odd one comes first.
The second rule holds even when the odd number is the larger of the two,
as in san yan liang yu ϝ㿔ܽ䁲 ‘in two or three words’ or when the integers
are not adjacent, as in jiu niu er hu zhi li б⠯Ѡ㰢П࡯ ‘the strength of nine
bulls and two tigers—tremendous effort’. It is not hard to fi nd exceptions
to either rule. For example wu hua ba men Ѩ㢅ܿ䭔 ‘multifarious’ violates
rule 1, and si qin san hao ಯࢸϝད ‘four diligents and three do-wells’, a
slogan that was used to fight SARS in 2003, violates rule 2.146 Big numbers
146. The “four diligents” were diligently wash hands, diligently wash face, dili-
gently drink water, and diligently ventilate; the “three do-wells” were wear face masks
well, regulate moods well, and exercise the body well. I am grateful to David Moser,
who observed signs bearing the slogan in Beijing; email message to author, May 8,
2004.
also ignore the rule, as in qian fang bai ji गᮍⱒ㿜 ‘by every possible
means’.147
But the two rules usually do hold, and moreover seem to influence
choices when Chinese speakers invent new phrases. For example, when
Communist Party leaders decided in 1981 to promote improvement in
public behavior, they came up with the formula wujiang simei Ѩ䆆ಯ㕢
‘five pay-attentions and four beautifuls’. The pay-attentions were attention
to civility, politeness, hygiene, order, and morality. The four beautifuls were
beautiful spirit, beautiful language, beautiful conduct, and a beautiful envi-
ronment. The meanings seemed good, the composite phrase sounded good,
and the phrase spread easily through the nationwide bureaucracies and
publicity system. It obeyed the two rules, and that obedience probably had
to do with why it “sounded right” to everyone. But how many people were
consciously aware of the two rules? Were the creators themselves aware?
Likely not, I would guess. Wujiang simei just sounded better than simei
wujiang.
What happens here is similar to what happens with unnoticed rhythms
because, in both cases, grammatical arrangements happen and results
sound good even if people are not aware of the reasons why. The pattern
2–2–3 just sounds better than 3–2–2, hence people say renmin zongli ren-
min ai, not (as grammar would have it) renmin ai renmin de zongli (‘the
people love the people’s premier’). It is worth noting that shortly after
wujiang simei was promulgated, Party leaders felt that “three loves” (love
of the Communist Party, love of the motherland, and love of socialism)
should be added. So how did they add it? They did not say san’ai ϝ⠅
‘three loves’ but san reai ϝ⛁⠅ ‘three ardent loves’. It is, of course, natu-
ral that they would want love of the Party to be “ardent,” but it is more
likely that re was added for rhythm, in order to make the whole phrase a
nice-sounding qiyan pattern: wujiang simei san reai sounds better than wu-
jiang simei san’ai. And you don’t even have to think about it.
147. I do not consider examples like qian pian yi lü ग㆛ϔᕟ ‘stereotyped’ or bai fa bai
zhong ⱒⱐⱒЁ ‘unfailing accuracy’ to be exceptions, because of their different under-
lying grammar.
2
Metaphor
In 1935 a young American named Graham Peck, fresh from Yale and
seeking adventure, set out for China, where he traveled widely, often by
bicycle, making sketches, taking notes, and picking up some Chinese. In
1940 he went to China again, this time for a longer stay during which he
took a post with the U.S. Office of War Information. Eventually, he be-
came frustrated with a gap he perceived between official U.S. hopes for
China and the abysmal conditions he found on the ground. After the war
he put his perceptions into a book that was excellent in many ways but bore
the unfortunate title Two Kinds of Time.1
Peck explained his title this way: “By one Chinese view of time, the
future is behind you, above you, where you cannot see it. The past is be-
fore you, below you, where you can examine it. Man’s position in time is
1. Graham Peck, Two Kinds of Time, rev. ed. (Boston: Houghton Miffl in, 1967)
(originally published 1950). Information in this paragraph is drawn in part from the
introduction to the revised edition by John K. Fairbank.
that of a person sitting beside a river, facing always downstream as he

watches the water flow past.”2 Peck goes on to tell about Chinese gardens
in which a gentleman meditating next to an artificial stream sits in that
same downstream-facing position while servants, standing behind him,
fill cups with wine and float them down to him. The gentleman drinks
when the cups arrive, but never knows just when they will. Peck’s point is
that Chinese culture looks at what has passed, not at what is coming.
This, for him, is “one kind of time.” The other kind is found in “America
and other Western countries,” where
man faces in the other direction, with his back to the past, which is
sinking away behind him, and his face turned upward toward the
future, which is floating down upon him. Nor can this man be static:
by our ambitious Western convention, he is supposed to be rising
into the future under his own power, perhaps by his own direction.
He is more like a man in a plane than a sitter by a river.3
Apparently aware that his analysis might leave the Chinese side looking
inferior, Peck points out the weakness of the forward-gazing airplane rider:
because he does not look back at the past, he flies blind into the future.
Peck’s characterizations of Chinese and “Western” concepts of time are
oversimplified and, even in the simple terms in which he presents them,
garbled. For neither Chinese nor English is his account complete, and the
claim that the two languages and cultures conceive radically different
“kinds of time,” as we shall see below, is false. Still, his speculation on the
ways time is conceived in terms of space is— as Mark Twain said of Rich-
ard Wagner’s music—not as bad as it sounds. Peck’s rudimentary specula-
tions were headed in an important direction. He anticipated by forty
years a subfield of cognitive science that has shown clearly both that (1) his
notion that conceptual systems are built on the metaphor that “time is
space” is correct and well worth studying, and (2) his particular conclu-
sions about Chinese and English are off the mark.
Late twentieth- and early twenty-first-century studies have found that
the “time is space” metaphor is apparently universal in human languages.
2. Peck, Two Kinds of Time, rev. ed., p. 7.

3. Ibid., p. 8.
Metaphor 115
In a 1994 book, Hoyt Alverson examines conceptions of time in English,

Chinese, Hindi, and Sesotho and fi nds similarities that are strong enough
to lead him to hypothesize that “the experience of time is based on a uni-
versal template of human experience.”4 An anthropologist, Alverson is
committed to reasoning from empirical data, and he reflects this commit-
ment in his book’s title, Semantics and Experience. One might wonder how
a strictly empirical approach can get Alverson from four particular lan-
guages to universal claims for all languages. His use of the word “tem-
plate,” although vague, suggests that he feels his study of four languages
reveals something that might be necessary for, or built into, human per-
ception generally. If this is his claim, it is not a strictly empirical ap-
proach, yet this kind of claim does have a fine pedigree. Immanuel Kant
listed time and space as “synthetic a priori categories of the pure under-
standing,” that is, necessary conditions in the human mind for any possi-
ble experience. And if Noam Chomsky is right that fundamental gram-
matical concepts are built into the human mind, it is at least plausible that
conceptions of time and space— or here, more precisely, time as space—
might be so as well.
In this chapter I will be working from examples in Chinese and En-
glish, noting both similarities and differences, and to this extent will be
working empirically. There are demonstrable and interesting differences
among conceptual metaphors in Chinese and English. There are also
striking similarities. Where I fi nd similarities, I will not be assuming
that the overlap is mere chance. I will at least open the door to the kind of
universal theorizing Kant and Chomsky have done and even an anthro-
pologist like Alverson apparently allows.
How Do Metaphors Work in Ordinary Language?
The study of creative metaphor in the literary arts is a wonderful and

endless field that no writer can encompass, and I will not try to. In this
chapter I will look at the more restricted field of metaphor that pervades
4. Hoyt Alverson, Semantics and Experience: Universal Metaphors of Time in English,

Mandarin, Hindi, and Sesotho (Baltimore: Johns Hopkins University Press, 1994),
p. xii.
daily-life expression. Cognitive scientists sometimes call this “conceptual

metaphor” or “structural metaphor.” When a person is angry, we might
say in English she is “all steamed up” or “hot under the collar”; similarly,
in Chinese she huor le ☿‫ܓ‬њ, literally “is on fire,” or shengqi ⫳⇨, literally
“emits vapor.” Speakers of Chinese or English understand the underlying
metaphor, which in this case is similar in the two languages, that “anger is
heat” or “anger is gas under pressure.” Lakoff and Johnson, who pioneered
this kind of inquiry in 1980 in Metaphors We Live By, explain the general
human need for conceptual metaphor this way:
Because so many of the concepts that are important to us are either

abstract or not clearly delineated in our experience (the emotions,
ideas, time, etc.), we need to get a grasp on them by means of other
concepts that we understand in clearer terms (spatial orientations,
objects, etc.).5
Lakoff and Johnson claim not only that “many” of our concepts are meta-
phorical but that “most of our normal conceptual scheme” is “metaphor-
ically structured; that is, most concepts are partially understood in terms
of other concepts.”6 This “structuring” involves a capture, in only one or
a few abstract words, of highly complex processes whose literal and de-
tailed exposition would require many, many more words. An example is
monetary “inflation,” which is not only a one-word summary of a complex
phenomenon but something that, in our ordinary language, we speak of as
a “thing,” sometimes even a living thing. We say things like “inflation is
killing us,” “inflation shows up at the checkout counter,” and so on. Lakoff
and Johnson call the turning of something like inflation into a “thing” an
“ontological metaphor.”7 They cite examples to show that even respectable
natural scientists commonly use this kind of metaphor to make concepts
easier to pass back and forth. They cite a 1997 issue of Science that contains
these sentences, asking us to take note of the italicized words:

Chicago Press, 1980), p. 115.
6. Ibid., p. 56, emphasis added.
7. Ibid., p. 25–32.
Metaphor 117
Most of the neurons in that very early processing stage merely report
what is happening on the ret ina. To do its job, however, this cortex
must cooperate with connected sensory regions that hold and use the
information for briefer periods of time.8
Could the scientists have described their fi ndings without metaphorically

imputing intentions to neurons? Yes, probably. But it would have taken
much longer and been much more tedious. On the other hand, nothing
seems to be lost by the shorthand of metaphor, so long as both reader and
writer understand it correctly. Lakoff and Johnson argue that ordinary
language has become so immersed in conceptual metaphor that it would
be very hard to do without it even if we were to try. They write: “If we
consciously make the enormous effort to separate out metaphorical from
non-metaphorical thought, we probably can do some very minimal and
unsophisticated nonmetaphorical reasoning. But almost no one ever does
this.”9 The philosopher John Searle goes further, claiming that sometimes
only metaphor can say exactly what a speaker wishes to say. “It is often the
case,” Searle writes, “that we use metaphor precisely because there is no
literal expression that expresses exactly what we mean.”10
How does daily-life metaphor work? Why can we say something like
“the ship plows the sea”— something that, if taken literally, is patently
false—yet have listeners not only understand us but think, “right, I see
what you mean”? A brief detour into the field of metaphor theory in cog-
nitive science will be useful here.
Although scholars vary in their terminology, the field consistently dis-
tinguishes (1) the thing described from (2) the nonliteral way it is de-
scribed. “Sam is a pig,” for example, singles out Sam and then tells us
nonliterally, via “pig,” something about him—that he is filthy, gluttonous,
sloppy, or the like. In the professional literature, “Sam” in such a sentence is
sometimes called the “defi ned concept” and piglike behavior the “defi ning
concept”; alternatively, “Sam” is sometimes called a “topic” and piglike
8. Lakoff and Johnson, Philosophy in the Flesh, p. 217.

9. Ibid., p. 59.
10. John R. Searle, “Metaphor,” in Andrew Ortony, ed., Metaphor and Thought, 2nd
ed. (Cambridge: Cambridge University Press, 1993), p. 111.
behavior a “vehicle.” The topic is sometimes said to fall within a “target

domain” while piglike behavior is in a “source domain.” (It is interesting
that these bits of jargon—“target,” “vehicle,” and “domain”— are them-
selves metaphors, which the theorists themselves build easily and under-
stand without much problem.) The field also uses the term “mapping” as if
it were a metaphor—for example, “mapping” a source domain onto a target
domain—but this usage often is not clear or rigorous.11
Metaphor theorists generally agree on the logical processes that a hearer
or reader uses in order to interpret conceptual metaphor.12 These pro-
cesses happen quickly and often without being noticed by the interpret-
ing mind, so what follows is meant not as a description of conscious
experience but as an answer to the question I just posed: how does con-
ceptual metaphor work? Step 1 is to recognize the literal meaning of a
proposition and see if that makes sense. If I say “Sam is a bowler,” you do
not need—unless I offer you some kind of signal—to start asking yourself
“what does he mean by bowler?” But if I say “Sam is a pig” (when Sam is,
in fact, a human being), then my very statement gives you a signal to start
casting about for metaphorical understandings. This “casting about” is
step 2. Steps 1 and 2 need not come in order; research has found that
people seem to process literal and available nonliteral interpretations
simultaneously.
Step 3 is to choose among the likely possibilities that turn up after the
casting about is done. These possibilities depend on custom and culture.
In modern English, “. . . is a pig” conventionally offers such options as:
messy, gluttonous, filthy, selfish. And how does the hearer choose among
them? The choice depends on context, where “context” includes surmise
11. The function of a map is not analogical in the way that “Sam is a pig” is; literally,
a map stands in relation to the ground it maps by (1) miniaturizing it, (2) marking
certain features (roads, towns, whatever) that it wishes to highlight by using amplifica-
tion, coloring, etc., and (3) eliminating all other features of the ground. This is hardly
what is going on between “target” and “source,” or “topic” and “vehicle,” when meta-
phors are used. In mathematics there is a more rigorous defi nition of “mapping”— by
which the elements in one set are assigned to elements in another— but this is not
what metaphor theory is doing, either.
12. My description here is drawn primarily from Sam Glucksberg, Understanding
Figurative Language: From Metaphors to Idioms (Oxford: Oxford University Press,
2001), pp. 8–11, and from Searle, “Metaphor,” pp. 102–108.
Metaphor 119
about what is in the speaker’s mind. If Sam never washes or makes his bed,
and the speaker is his roommate, then “Sam is a pig” will likely be under-
stood as “Sam is messy” or “Sam is filthy” (or both). If Sam ate the whole
pumpkin pie and the speaker is someone else at the table who did not get
any, then “Sam is a pig” will likely mean “Sam is a glutton” or “Sam is
selfish.” If the common cultural possibilities run out, then the hearer
has to look further into the context and/or do further speculation about
the speaker’s mind. If Sam is, for example, neat, clean, anorexic, and self-
effacing, and if someone nevertheless says, “Sam is a pig,” it might be be-
cause Sam cheers for the Arkansas Razorbacks. It could also be because he
is a Chinese born in 1971 or 1983. The hearer needs to cast around.
Obviously, too, the hearer must have a cooperative attitude toward the
speaker’s utterance. After hearing that “Sam is a pig” one could, if one
chose to be difficult, logically infer any of literally hundreds of statements
about Sam, including “Sam is a vertebrate,” “Sam has bristles,” and many
other propositions that would not be, in most contexts, what the speaker
had in mind. But people seldom do this— and when they do, they gener-
ally know they are mischievously “breaking the rules.” The philosopher
Paul Grice has noted the operation of a “cooperative principle” by which
participants in conversation expect that each will make a “conversational
contribution such as is required, at the stage at which it occurs, by the ac-
cepted purpose or direction of the talk exchange.”13 Grice’s famous prin-
ciple applies to the interpretation of metaphor. When Shakespeare says
“Juliet is like the sun,” no one is confused. We know that he means “Juliet
is bright and warm,” not “Juliet is for the most part gaseous” or “Juliet
will incinerate you if you get too close”— even though the latter two are
well-known facts about the sun.14 Lakoff and Johnson, in noting this
feature of metaphor, point out that the “defi ning concept” in metaphor
always contains more stuff than is carried over to the “defi ned concept.”
The conceptual metaphor “ideas are food,” for example, leads English
speakers to speak of “raw facts” or “half-baked notions”— but, Lakoff
and Johnson observe, this conventional metaphor does not normally
13. H. P. Grice, “Logic and Conversation,” in Peter Cole and Jerry L. Morgan, eds.,
Syntax and Semantics, vol. 3, Speech Acts (New York: Academic Press, 1975), pp. 45.
14. See Searle, “Metaphor,” p. 96.
lead to many other possibilities, such as “sautéed, broiled, or poached

ideas.”15
It is important to note that the implicit claims of metaphor do not have
to be true for metaphor to work properly. “Richard is a gorilla” can
mean— and be correctly understood to mean—that Richard is fierce and
nasty, even if it is false that gorillas are fierce and nasty. If the speaker’s
meaning is that Richard is fierce and nasty, then, in order for the meta-
phor to work properly, it need only be true that both speaker and hearer
are familiar with a cultural convention that views gorillas as fierce and
nasty; the literal proposition does not have to be true, and it does not even
have to be the case that speaker and hearer believe it to be true.16 Consider
the two statements “my daughter is an angel” and “my daughter is no an-
gel.” For any human being with a daughter, the first of these statements is
literally false and the second is literally true. Moreover, in literal terms the
two statements are contradictory, that is, in no possible world could both be
true at the same time. But as metaphor, both statements can be true of the
daughter (at different times), and both can be easily and correctly under-
stood, provided that speaker and hearer are on the same wavelength (“coop-
erating,” as Grice would say) about what “angel” and “no angel” entail.
The fact that a person needs context in order to interpret a metaphor
has led some to suppose that this is what distinguishes metaphor (and
other figurative speech) from literal expression. But the matter is not that
simple. Searle has rightly observed that the interpretation of some kinds
of literal expression also requires context.17 “Mary is tall,” for example,
cannot be interpreted without reference—whether explicit or implicit—to
the group of people with whom the speaker is comparing Mary. Is she a
fourth-grader? A volleyball team member? “Tall” would refer to different
heights in these two cases. Rigorously speaking, “Mary is tall” means
“Mary is taller than most of the people within the reference group I have
in mind.” Searle is also very good at showing how metaphorical expres-
sion is only one example of nonliteral expression that obeys the same
general set of rules that metaphor interpretation obeys. Irony, for example,
15. Lakoff and Johnson, Metaphors We Live By, p. 109.

16. Ibid., Searle, “Metaphor,” pp. 92– 93.
17. Ibid., pp. 85– 86.
Metaphor 121
also asks the hearer to begin by ruling out the literal interpretation of an
utterance; it then asks the hearer to accept an opposite meaning.18 Here
intonation can be a cue, as when you say “Wonderful!” after I have knocked
over and shattered your Ming vase. But often it is context alone that leads
the hearer to interpret an ironic utterance correctly. Sometimes “speech
acts” as well follow the same general rules. When I say “Can you pass the
salt?” my hearer normally rules out the possibility that I mean to ask a ques-
tion about your ability to pass salt. He or she instead just passes the salt—
and in that, of course, reflects a correct understanding of the utterance.
The standard use of questions like “Can you pass the salt?” to mean
“Please pass the salt” or “How are you?” to mean “Hello” has a counter-
part in the often-noted phenomenon of the “dead metaphor.” Most meta-
phor theorists allow that daily-life metaphors can and do die. After death
they often remain useful, however, in much the same way literal words or
phrases are useful. For example, if I say “Interstate 95 goes to Boston,” the
metaphor that suggests that the road itself “goes somewhere” is utterly
dead. It is hardly different from the literal statement that “Interstate 95 is
a road.” The metaphor’s deadness is so well established that any effort to
bring it to life would seem perverse. If you say “Interstate 95 goes to Bos-
ton,” and I say, “Nonsense, Interstate 95 just lies there on the ground,” we
both know that I am breaking rules that are so well established that we
normally do not even notice them. Owen Barfield has pointed out that
etymology, if pressed far enough, shows that nearly every word has roots
that refer either to “a solid, sensible object, or some animal (probably hu-
man) activity,” and that everything else, in a sense, is dead metaphor; for
example elasticity has roots in ‘draw’ and abstract in ‘drag’.19
Dead metaphors, we can assume, usually or always began as live ones.
Lakoff and Johnson have written that conceptual metaphors often had
“physical” or “experiential” bases in their beginnings. For example the ideas
of “up” and “more”— as pure concepts—have no essential connection. Yet
we commonly use the conceptual metaphors “more is up” and “less is down”
to say things like “murders are up,” “foreclosures are down,” and so on.
18. Ibid., pp. 109.

19. Owen Barfield, Poetic Diction: A Study in Meaning (Middletown, Conn.: Wes-
leyan University Press, 1973), p. 64.
Lakoff and Johnson hypothesize that the “physical basis” for this usage is
the simple fact that when you get more of something, and put it in the
same place, a pile forms. Pile tops go up, so “more” becomes “up.”20 This
theory gets support from the fact that many languages use the metaphor.
In China, as elsewhere, things like pile tops and fluid levels go up when
more of something is in one place, or in one container; accordingly, Chi-
nese says things like jiaqian gao Ӌ䪅催 ‘the price is high’ and shuiping di ∈
ᑇԢ ‘the level is low’. Such usages have been influenced by modern West-
ern languages and so might not seem good evidence for the universality of
the “physical basis” for “more is up.” Other examples, though, make the
point without ambiguity. In English we say “red in the face” to mean an-
gry or agitated, and in Chinese say mianhong erchi 䴶㑶㘇䌸 ‘face red ears
red’ in quite the same way. Here there is no sign of borrowing; there is,
though, a plausible common “physical basis” in the fact that for human be-
ings in any language or culture, blood tends to rush to the face when one
is angry or agitated.
An important characteristic of conceptual metaphor is that it is not just
used in one or two phrases but can underlie a variety of related expres-
sions. Lakoff and Johnson cite the example “love is a journey,” which can
underlie a variety of expressions: our relationship “is at a crossroads,” “has
hit a dead-end,” “is spinning its wheels,” or “has entered the fast lane” or
“we are going in different directions,” and so on.21 Lakoff and Johnson
might better have called this metaphor “a romantic relationship is a jour-
ney,” not “love is a journey,” since there are many kinds of love for which
“journey” is harder to apply. (“I love chocolate ice cream,” “God so loved
the world . . .” etc.) But the point itself—that metaphorical expressions
tend to come in families—is important. Not only do the various expres-
sions associated with a conceptual metaphor tend to cohere; sometimes
conceptual metaphors become strong enough that they can shape the way
people absorb new experience as it arises. I have noted Lakoff and John-
son’s identification of the English conceptual metaphor “consciousness is
up and unconsciousness is down” according to which we fall asleep and

Metaphor 123
wake up, sink into a coma, and so on.22 When such metaphors become es-
tablished in a culture, they tend to shape understanding of new experience.
When Freud began to speak of the less-than-conscious mind, for example,
it was natural to speak of a “sub”-conscious.” This kind of fact is what leads
Lakoff and Johnson to claim that conceptual metaphors are “metaphors we
live by.” They also claim that strong, broad metaphors have a power to
govern smaller clusters of others. For example, the broad metaphor “life is
a journey” (e.g., “she got a head start in life,” “he’s never let anyone get in
his way”) can be seen as subsuming, and perhaps giving rise to, conceptual
metaphors of smaller domain such as “love is a journey” (“We’re in the fast
lane,” etc.) and “a career is a journey” (“He’s on the fast track,” etc.).23
Lakoff and Johnson have been criticized for claiming too much struc-
turing power for conceptual metaphors. While it may be true that we of-
ten conceive things in terms of metaphor, can we go further, as Lakoff and
Johnson do, to claim that “it structures the actions we perform”?24 When
I embark on a romantic relationship, do I actually behave differently be-
cause I am thinking in terms of “love is a journey”? A key question here
has been whether established metaphors are “automatically accessed” or
not. If, for example, an English speaker thinks of a romantic relationship,
does “love is a journey” always kick in, or is it optional? Psychologists Sam
Glucksberg and Matthew McGlone devised an experiment in which
American college students were asked what “our love is a voyage to the
bottom of the sea” or “our love is a bumpy roller-coaster ride” might mean.
Here the verbal cues clearly set the stage for conceptions of the “love is a
journey” sort. If conceptual metaphor is accessed automatically, one
would expect the student responses to mention discovery, roads over roll-
ing hills, “uncharted waters,” “where we are headed,” and other phrases
consistent with “love is a journey.” And a few of Glucksberg and McGlone’s
22. Lakoff and Johnson, Metaphors We Live By, p. 15. Noted in the Introduction.
23. George Lakoff, “What Is a Conceptual System?,” in Willis F. Overton and Da-
vid S. Palermo, eds., The Nature and Ontogensis of Meaning (Hillsdale, N.J.: Erlbaum,
1994), p. 62.
24. Lakoff and Johnson, Metaphors We Live By, p. 4. Boaz Keysar, Yeshayahu Shen,
Sam Glucksberg, and William S. Horton explicitly argue against this claim by Lakoff
and Johnson in “Conventional Language: How Metaphorical Is It?,” Journal of Mem-
ory and Language 43 (2000), p. 578.
responses were indeed of this type. But, significantly, most were not.
Most students thought the given phrases meant “we don’t talk enough,”
“we’re drowning in each other’s problems,” “our love is exciting but not
very stable,” and other such things.25 Experiments like this show that con-
ceptual metaphors do have influences on the way we conceive things, but
that they are not obligatory. They are more like arrows in a quiver—to be
drawn on when useful, otherwise not.
Work by Boaz Keysar and others has found the interesting result that
very conventional metaphors— or “dead” metaphors—have fewer connec-
tions to underlying concepts than fresh metaphors do. In the metaphor
“happy is up, sad is down,” for example, we have English examples like
“she’s feeling low,” “I’m depressed,” and so on. Similarly, in Chinese we
have gaoxing 催݈ ‘high mood—happy’ and diluo Ԣ㨑 ‘low fall— downcast’.
Keysar and his colleagues argue that a conventional phrase like “I’m de-
pressed” normally is understood lexically, hardly different from how “I’m
a professor” is understood. The hearer does not need to draw on the under-
lying concept of “sad is down” in order to understand. Similarly, a hearer
would probably not need “happy is up” in order to understand “I’m on
cloud nine” or “she’s higher than a kite,” because these familiar clichés are
hardly different from ordinary words. But it is precisely when new expres-
sions come along, Keysar argues, that conceptual metaphors are useful
and perhaps even necessary if correct understanding is to result. In a case
like “I’m feeling lower than a piece of gum stuck on the bottom of your
boots,” a hearer will likely need to draw on “sad is down” in order to make
sense of the phrase. 26 Here again we see conceptual metaphors waiting in
abeyance, as it were—ready to help when needed, but not demanding to
be used.
What happens when two or more metaphors are used together? In lit-
erary matters “mixed metaphor” is supposed to be a mistake, largely for
aesthetic reasons, but what happens when daily-life conceptual metaphors
come together? Do they collide? Muddle the underlying concept? Or
does it not matter?
25. Sam Glucksberg and Matthew S. McGlone, “When Love Is Not a Journey:
What Metaphors Mean,” Journal of Pragmatics 31, no. 12 (1999), p. 1547.
26. Keysar et al., “Conventional Language,” p. 579.
Metaphor 125
Experiments in the psychology of metaphor have shown that hearers

tend to prefer consistent metaphors. For example, to describe someone
speaking and acting in anger, people prefer “she fumed and exploded” to
“she fumed and growled.” Even when different subjects are involved—thus
allowing a bit more “space” in which to accommodate different metaphors,
people still prefer “she fumed; he exploded” to “she fumed; he growled.”
But preferences aside, the same experiments show that the mixing of con-
ceptual metaphors normally does not inhibit understanding or even slow
down reading speed.27 This result is further evidence that dead or frozen
metaphor is processed in the mind more or less as ordinary words are.
“Conceptual clash” appears not to be a problem.
These experiments were done in English. Does the point hold for Chi-
nese as well? Although no experiments are available, I believe the point
does hold. It is easy to find examples of clashing dead metaphors in Chi-
nese, and such clash does not seem to cause any confusion among users.
Ning Yu, an analyst of conceptual metaphor, cites the example gaiguo
shenghuo shuiping bi mouxie fada guojia luohou wushinian 䆹೑⫳⌏∈ᑇ↨ᶤ
ѯথ䖒೑ᆊ㨑ৢѨकᑈ ‘the living standard in this country is fifty years
behind certain developed countries’. 28 Here are clashing metaphors: the
development “water level” (shuiping) of a country “falls behind” (luohou)
those of other countries. “Water level” implies vertical movement, be-
cause water levels go only up or down, not left or right; but “fall behind”
normally implies horizontal movement left or right. (It is true that luo,
literally “fall,” can be conceived vertically, but when it fuses with hou, lit-
erally “behind,” luohou is pretty clearly horizontal.) Do these clashing
metaphors cause puzzlement that slows people down? I doubt it. It is more
likely that the metaphors are simply accepted as lexical items.
The point also seems plain when one considers the ease with which
Chinese speakers use different metaphors for a “next” or “last” time pe-
riod. The day after tomorrow is houtian ᕠ໽ (literally “the behind day”),
where the metaphor hou ‘behind’ implies a horizontal time line. Similarly
27. Glucksberg and McGlone, “When Love Is Not a Journey: What Metaphors
Mean,” pp. 1553–1554.
28. Ning Yu, The Contemporary Theory of Metaphor: A Perspective from Chinese (Am-
sterdam: Benjamins, 1998), p. 120.
qiantian ࠡ໽ ‘the day before yesterday’, qiannian ࠡᑈ ‘the year before

last’ hounian ᕠᑈ ‘the year after next’, and metaphors such as guo qu le 䘢
এњ ‘has passed (away)’ all imply a horizontal time line. On the other
hand, for weeks and months, the implicit time line is vertical: xia xingqi
ϟ᯳ᳳ (literally ‘the week below’), shang ge yue Ϟ‫ן‬᳜ (‘the month above’),
and so on. Do users of Chinese get confused or slowed down by having to
“change conceptual schemes” every time they shift from days to weeks, or
from months to years? The answer is so obviously no that an experiment
almost seems a waste of time. In a sentence like ta bushi shang ge yue lai de,
shi qiannian lai de ཌྷϡᰃϞ‫ן‬᳜՚ⱘ, ᰃࠡᑈ՚ⱘ ‘she came the year before
last, not last month’, the metaphor shift will not even be noticed, let alone
be troublesome. A statement like shangge libai yijing guoqu le Ϟ‫⾂ן‬ᢰᏆ㍧
䘢এњ ‘last week has already passed’ might seem irritating, but only be-
cause it is a worthless tautology, not because of the metaphor clash between
shang (implying a vertical time) and guo qu (implying a horizontal one).
Edward Slingerland, who has used contemporary metaphor theory to
study ancient Chinese thought, has concluded that “literal consistency is
not something that we require of our metaphors.”29 Here Slingerland’s
“we” includes people living in pre-Qin China as well as today. He notes,
for example, that in ancient Chinese—just as in modern English—the self
can be metaphorically divided into two: a self that acts and a self that is
acted on. In modern English we can say “I dragged myself out of bed” or
“he lost control of himself,” and Slingerland notes examples from
Zhuangzi, including zi shi 㞾༅ ‘lose oneself’ and in particular the state-
ment of Zi Qi of Southwall that wu sang wo ਒୾៥, literally “I lose me,”
that seem to illustrate a similar metaphorical split between “subject” (the
actor) and “self” (the acted-on).30 Verbs in pre-Qin texts that metaphor-
ically describe self-cultivation in terms of “following” nature (either one’s
own nature or a transcendent nature)—for example cong ᕲ, yin ಴, yi ձ,
and shun ䷚— all can be seen as implicitly assuming this distinction be-
tween subject and self. But another family of metaphors that conceive the
self, which Slingerland calls the “at ease” group (an ᅝ, shu 㟦, jing 䴰, xi
ᙃ, xiu ӥ, you ␌, and others) implicitly assume a unitary self that admits
29. Edward Slingerland, Effortless Action: Wu- wei as Conceptual Metaphor and Spirit
Ideal in Early China (Oxford: Oxford University Press, 2003), p. 36.
30. Ibid., p. 31.
Metaphor 127
no split between actor and acted-on.31 How can we reconcile the two sets
of metaphors? Slingerland argues that we do not have to, and indeed that
it would be a mistake to demand that we try. It is a mistake because under-
lying human nature (in ancient China no less than today) is complex
enough to be difficult to capture from a single point of view. Differing
metaphors illuminate different facets of it, but none is or can be compre-
hensive, hence there is no requirement that every useful metaphor should
fit with every other.
We are left with a theory of conceptual metaphor in which meta-
phorical expressions are useful to our thinking and communicating but
only as assistants, as it were, in getting at one or another aspect of com-
plex topics. There is no contradiction between using metaphors that clash
and holding that a full view of complex issues can be coherent and inter-
nally consistent. As long as metaphors get their particular jobs done, it is
not terribly important whether they fit together, clash, or are mutually
irrelevant. None by itself can or should be viewed as reflecting a unified
worldview.
Sometimes one conceptual metaphor piggybacks on a second and, as it
were, invades its territory. But here, too, pragmatism seems to trump pu-
rity: if the result works, no one minds. Consider the sportscaster who
comments before a basketball game that “Georgia Tech will have to keep
the game in the sixties; anything north of that and they’re in big trouble.”32
He means that the Georgia Tech players are specialists in defense and
will need to keep the score low in order to exploit their comparative
advantage. He combines the well-established conceptual metaphor “more
is up” with the modern cartographical convention “up is north” to yield
the piggybacked metaphor “more is north.” Another sportscaster—this
one commenting on the U.S. Open tennis tournament in September
2005— observed that “men north of thirty,” who were beyond their prime,
would have a tougher time competing. Hearers of such phrases are not
startled into asking themselves “Hey, how do I put this conceptual clash
into a coherent worldview?” The metaphor serves its purpose, and that is
enough.
31. Ibid., pp. 29–30.

32. Christopher Russo on Mike and the Mad Dog, WFAN radio, April 5, 2004,
7:32 p.m.
But if conceptual metaphor behaves this way, if it can do fl ips and twists
in the ser vice of underlying ideas, are the cognitive scientists who claim
that “conceptual metaphor” sometimes shapes our thinking therefore
wrong? Not necessarily. It can remain true in some cases that our habits
in metaphor guide the ways we think. At a minimum, we can make the
empirical observation that metaphorical expressions in human languages
do tend to occur in internally consistent families such as (for modern En-
glish) “consciousness is up, unconsciousness is down,” “more is up, less is
down,” “a romantic relationship is a journey,” and others. Given that
these family resemblances among metaphors do exist, it is reasonable to
infer that the resemblances are not accidental but must have some cause.
But then we should ask: which way does the causality go? Is there some-
thing about the commonality of human experience— or perhaps the
structure of the human mind—that causes metaphors to bunch in consis-
tent families? Or do the bunchings themselves take over the ways we
think and give rise to their own consistency? (Or both?) There have been
scientific experiments in this area; there are also some unsettled contro-
versies. We turn to some of them next.
Metaphor and Thought
Lakoff and Johnson hold a strong position on the relation between meta-
phor and thought. The back cover of the paperback edition of Metaphors
We Live By claims that “reality itself is defined by metaphor, and as meta-
phors vary from culture to culture, so do the realities they defi ne.” The
authors’ claim for metaphor can be viewed as a particular instance of what
Benjamin Whorf claimed in 1939 for language in general—to paraphrase
the famous “Sapir-Whorf hypothesis,” that the language a person speaks
affects the way he or she thinks. Whorf has drawn much criticism from
later cognitive scientists, in part because his reasoning can seem circular.
As satirized by Steven Pinker, for example, Whorf claims that “[Eskimos]
speak differently so they must think differently. How do we know that
they think differently? Just listen to the way they speak!”33 Pinker holds—
33. Steven Pinker, The Language Instinct (New York: Morrow, 1994), p. 61.
Metaphor 129
and others, including Noam Chomsky and Jerry Fodor, hold similar
positions—that at least some human thinking, including the fundaments
of grammar, is built into the human mind, therefore universal, and there-
fore not subject to cultural variance.34
There are two different kinds of arguments in the field for why “con-
cepts,” including metaphorical concepts, sometimes appear to be universal.
The Pinker-Chomsky-Fodor kind of argument about the structure of the
mind is one; Lakoff, Johnson, and others espouse a different sort of argu-
ment, which is that the physical experience of human beings—which after
all shares a great deal in common, no matter where or in what culture one
lives—naturally induces the same kinds of metaphors in any language. Stud-
ies have shown, for example, that heart rate and body temperature both
tend to rise when people feel anger, or even when they are asked only to
imagine anger or to simulate it in their bodily movements.35 If this is true,
then it is not surprising that “hot with anger” occurs in English and huor le
☿‫ܓ‬њ ‘on fire—angry’ is used in Chinese. Environmental commonalities
can have the same effect. “More is up” is a natural inference to draw in any
context where physical objects and gravitational force are both present.
It is unfortunate that Lakoff and Johnson often use the word embody to
refer to the relation they find between physical experience and conceptual
metaphor, because this use of the word clashes with the normal sense of
it. When Lakoff and Johnson write that “the mind is inherently
embodied”36 they mean that it is “shaped by the body’s experience.” But in
common use (as well as in dictionaries), embody means “put into a body” or
“give a material or concrete character or form to.”37 Lakoff and Johnson’s
34. Noam Chomsky, Reflections on Language (New York: Pantheon, 1975); Jerry
Fodor, The Language of Thought (Cambridge, Mass.: Harvard University Press, 1975).
35. Paul Ekman, Robert W. Levenson, and Wallace V. Friesen, “Autonomic Ner-
vous System Activity Distinguishes among Emotions,” Science 221, no. 4616 (Septem-
ber 16, 1983), pp. 1208–1210.
37. Defi nitions from The Shorter Oxford Dictionary on Historical Principles (Oxford:
Clarendon Press, 1933), 1:598. The phrase “the embodied mind,” in terms of these
dictionary defi nitions, would have to mean something physical, like “the brain,” or at
least something metaphor ically substantive, like Immanuel Kant’s Critique of Pure
Reason. But a reader who conscientiously follows Lakoff and Johnson’s usages in this
direction will fi nd that the pursuit is futile.
idiosyncratic usage thus might cause confusion where there needn’t have
been any, but this should not obscure the value of their basic claim, which
is that one important reason why human beings sometimes think and
speak similarly is that they all have bodies.
Lakoff and Johnson’s kind of reason for conceptual commonality dif-
fers from that of Pinker, Chomsky, and others, in that it is based in expe-
rience. This means that at least in theory, concepts that are held in com-
mon might vary from context to context. Gravity varies from place to place
only by minuscule amounts, and bodily responses to anger might be the
same everywhere; but in other aspects of living, cultural as well as physi-
cal, there can be significant differences, and metaphors might therefore
be different. When Lakoff and Johnson postulate that both metaphors
and thought patterns vary from culture to culture, they implicitly raise
the “Whorf” question: Do different languages induce different ways of
thinking?
Psychological experiments in the early twenty-first century have lent
new credibility to at least some parts of the Whorf hypothesis. Lera Boro-
ditsky, who sets aside the “strong Whorfian view” that “thought [is] en-
tirely determined by language,” nevertheless defends a weaker version
according to which “frequently invoked mappings may become habits of
thought.”38 For example, Boroditsky notes— correctly—that Chinese uses
a “future is down” metaphor more often than English does; she then tests
in several ways the hypotheses that Chinese speakers tend to think verti-
cally about time more easily than English speakers do and that English
speakers tend to think horizontally about time more easily than Chinese
speakers do. She does this by presenting vertical and horizontal cues (a
fish apparently in motion is a horizontal cue, for example) and then tests
to see how quickly native speakers of the two languages can confi rm that
March precedes April (where “precedes” moves either horizontally or
vertically, depending on language habits). As hypothesized, Chinese speak-
38. Lera Boroditsky, “Does Language Shape Thought? Mandarin and English
Speakers’ Conceptions of Time,” Cognitive Psychology 43 (2001), pp. 2 and 7. Lakoff,
the senior figure in the group that holds this view, sometimes carries it so far that he
undermines his credibility. See Lakoff, “Staying the Course Right over a Cliff,” op-
ed, New York Times, October 27, 2006, p. A19.
Metaphor 131
ers in her experiments performed better after seeing a vertical display,

and English speakers did better after seeing a horizontal display. The re-
sults suggest that in matters of time, Chinese speakers were “ready to
think” vertically more easily than English speakers, who were more “ready
to think” horizontally.39 In a related experiment, Boroditsky found that
English speakers did better with vertical thinking about time after practic-
ing imaginary English examples of it— such as “Nixon was president above
Clinton.”40 She concludes that “language is a powerful tool in shaping
thought about abstract domains.”41
Boroditsky has been challenged by Jenn-yeu Chen, who did not get the
same results from the same experiments. Chen also correctly points out
that the differences between Chinese and English in the use of “horizon-
tal” versus “vertical” time lines are only a matter of degree and that, in
fact, horizontally conceived usages in Chinese outnumber vertically con-
ceived usages by about five to three.42 Chen’s challenge to Boroditsky casts
considerable doubt on the claim that Chinese speakers really do have a
penchant for vertical conception of time. But Chen does not undermine,
indeed seems to accept, the larger claim that languages tend to “shape
thought about abstract domains.”
One important objection to the work of both Boroditsky and Chen
could be the one I attributed above to Pinker: that the reasoning seems
circular— one claims that language shapes thought, but the evidence for
the shaped thought is the very language that does the shaping.43 In an
apparent effort to skirt this objection, Boroditsky and nine colleagues
devised some clever ways to test how people think of time in terms of
39. The Chinese speakers had to overcome a flaw in Boroditsky’s experiment de-
sign. Her vertically arrayed fish are moving up, but Chinese time—when it does move
vertically—moves down. Hence for Chinese speakers in the experiment, the prefer-
ence for verticality in general had to be stronger than the preference for direction
within the verticality.
40. Boroditsky, pp. 17–18.
41. Ibid., p. 1.
42. Jenn-yeu Chen, “Do Chinese and English Speakers Think about Time Differ-
ently? Failure of Replicating Boroditsky (2001),” Cognition 104 (2007), pp. 429– 430.
43. Boroditsky’s data are “response times,” which themselves are not verbal, but
they are triggered by verbal cues.
space without using language—by using only spatial images on a computer

monitor.44
These experiments by Boroditsky and her colleagues study speakers of
English, Indonesian, Greek, and Spanish. The researchers begin by show-
ing that all four languages use both length and volume as spatial metaphors
for time, for example, in English, a “I waited a long time” or “the meeting
took much time.” They then show that in daily-life language, English and
Indonesian tend to use the length metaphor much more often than the
volume metaphor, while Greek and Spanish use the volume metaphor
much more often than the length metaphor. They then pose the question
whether these different habits in language use might induce different
habits of thought, including thought that is not expressed in language.
They showed subjects a line that grows in length across a computer moni-
tor; then they showed more such lines that grow at different speeds and
extend to different lengths. Subjects used the click of a mouse (not lan-
guage) to indicate which lines were longer in length and which took longer
in time to complete. No one had much trouble identifying which lines
were longer in length. But for the metaphorical question of which lines
took “longer” in time, the physical length of lines was shown to interfere
with people’s judgment. People thought that a longer line “took longer” in
time to complete than a shorter line, even if the time duration of the two
lines was the same; similarly, they thought that physically shorter lines
lasted less time than lengthier ones even when this was not the case. They
continued to make these mistakes even after being apprised of the tricki-
ness of the stimuli. So here was evidence that a mental habit that appar-
ently originated in metaphor was indeed affecting the way people think,
even before anything was expressed in language.
Speakers of all four of the languages involved seemed to be influenced
by the metaphorical habits of their own languages. When presented with
44. Daniel Casasanto, Lera Boroditsky, Webb Phillips, Jesse Greene, Shima Gos-
wami, Simon Bocanegra-Thiel, Ilia Santiago-Diaz, Olga Fotokopoulu, Ria Pita, and
David Gil, “How Deep Are Effects of Language on Thought? Time Estimation in
Speakers of English, Indonesian, Greek, and Spanish,” Proceedings of the 26th Annual
Conference of the Cognitive Science Society (Hillsdale, N.J.: Erlbaum, 2004),
pp. 575–580.
Metaphor 133
lines across a screen, speakers of English and Indonesian, in which “time

is length” metaphors are common, had much more trouble than did speak-
ers of Greek and Spanish, where “time is length” is less common. The
mental habit of “time is length” was evidently more deeply ingrained for
English and Indonesian speakers. But Boroditsky and her colleagues did
another experiment where, instead of presenting growing lines to repre-
sent “more time,” they presented containers being filled; these “filled up”
to different degrees and at different speeds in a manner that was parallel
to the “growing lines” of the first experiment. Now the results reversed:
Greek and Spanish speakers experienced more “interference” from irrel-
evant cues than did English or Indonesian speakers. For Greek and Span-
ish speakers, the mental habit of “time is volume” appeared to be more
deeply ingrained.
Although this work by Boroditsky and her colleagues is impressive evi-
dence for one side in the debate over Whorf, in one respect it also tends to
confirm the other side, even though they do not say so. By showing that
native speakers of four languages have conceptual habits that correspond
to the “time is space” metaphors that are characteristic of their languages,
the work of Boroditsky and her colleagues tends to confirm Whorf. But
by showing, for all four languages, that spatial metaphors interfere with
judgments about time but not vice versa, the work tends to confirm the
views of Chomsky, Pinker, and others who claim that there are concep-
tual structures in the mind that no language or culture can affect. Boro-
ditsky and her colleagues observe that although their subjects were some-
times confused when physically longer and shorter lines took an equally
“long” time, “line duration did not affect subjects’ distance estimates.” 45
The length of a line was easy to conceive regardless of how much time
elapsed in completing it, but the reverse was not the case: the time a line
took to grow was not easy to conceive as length varied. The authors call this
an “asymmetric relation between space and time,” and it deserves more
careful thought. What exactly is this asymmetry? Might it be something
that is “given” in the human mind? Lakoff and Johnson, Boroditsky, and
many others have observed that in general, human languages conceive
45. Ibid., p. 577.

time using spatial metaphors.46 Why don’t we conceive space using tem-
poral metaphors? Can such a thing even be done? Will the structure of
the mind allow it? I doubt it. Here the “innate-structure” notions of
Pinker, Chomsky, Kant, and others seem vindicated.
One kind of example of “space in terms of time” is easy to observe: we
all say things like “the library is just five minutes from here” or “Philadel-
phia is an hour away.” But these are only ways of measuring distance in
terms of the assumed average speed of certain vehicles, in these two cases,
our legs or our automobiles. Measuring can be done a number of other
ways, for example in money. Farmers in rural Guangdong use money to
measure distance. Hin-gong yat ho mm (Xian’gang yi-hao wu) 㲀ቫϔ↿Ѩ
‘Xian’gang is fifteen cents (in bus fare) from here’; Fat-san gaw-yee (Foshan
ge er) ԯቅϾѠ ‘Foshan is 1.20 yuan away’. But these ways of measuring—
by vehicle speed, vehicle fare, or whatever—are matters quite different from
the question of conceiving space as time. In order to do that conceiving,
we would have to be able to make sense of phrases that somehow do the
inverse of what phrases like “a long time” do. Can we? Let’s try by sub-
stituting “space” for “time” in that phrase, and then try to substitute a
modifier that corresponds to “long” but refers to time. It is hard to think
of such a modifier, and it is worth reflecting on why this is so hard. Will
“ancient” or “eternal” work, perhaps? Can we conceive an “ancient space”
or “eternal space” in a way that delivers the notion of “a very large space”
just as easily as “a long time” delivers the notion of a very large time? We
normally do not do such things, and it is far from clear that we can do
them. What, to ask a parallel question, would be the time-for-space ana-
logue of “a short time”? Can we conceive “a brief space”? To the extent
that our minds do make sense of phrases such as “an ancient space” or “a
brief space,” I still do not feel that we are conceiving space in terms of
time, but are probably doing smaller, more poetic things. “An ancient
space” feels as if it might mean something like “a musty crypt with Dead
Sea scrolls stacked in a corner.” And “eternal space” still summons the idea
46. Lakoff and Johnson, Philosophy in the Flesh, p. 139; Boroditsky, “Does Language
Shape Thought,” p. 4; Ning Yu, Contemporary Theory of Metaphor, p. 84. Ning Yu sug-
gests that the commonality may have an experiential basis simply because objects in
space can be pointed to, whereas objects in time cannot be.
Metaphor 135
of a long-lasting space, not necessarily an endlessly expansive one. It is true

that in real life we do use phrases like “a two-hour space,” but note that
here we are not conceiving space in terms of time but reverting— once
again—to the “time is space” metaphor to which our minds are well ac-
customed. “A two-hour space” likely refers to a block of time on a chart. In
short, it may be no accident that human languages the world over conceive
time in terms of space but not space in terms of time. There is something
about the structure of the human mind that makes this so.
But if that is so, then both sides in the Whorfian debate have part of the
truth. Does one language lead its speakers to think differently from the
ways speakers of other languages think? Yes, for some concepts. But does
the structure of the human mind condition and limit the ways any person
thinks, regardless of language or culture? Also yes. It is not contradictory
to hold both positions.
In the remainder of this chapter, I will not be concerned with trying to
winnow evidence in search of the borderlines between the two basic posi-
tions on Whorf but will look primarily at the side on which Whorf is
right, that is, I will be interested in showing how two different languages,
Chinese and English, correspond to different ways of conceiving things.
The exercise is especially useful for languages that differ considerably, as
do Chinese and English. To compare neighboring languages, like French
and English, is not as mind-stretching. I have noted in the Introduction
that when Lakoff and Johnson’s Metaphors We Live By was translated into
French, most of their examples went into French easily. For example, when
the authors point out the metaphor “causation is emergence” and give ex-
amples like “He shot the mayor out of desperation” and “He dropped
from exhaustion,” these became French easily as “De désespoir, il a tué le
maire” and “It est tombé d’épuisement.”47 But, for complex reasons— some
of which I will address below—the “causation is emergence” metaphor is
not available in Chinese. The translator’s task would be much more diffi-
cult in producing a Chinese version of Lakoff and Johnson’s book. The
work would involve not only a wider search for examples but, more often
47. George Lakoff and Mark Johnson, Les métaphores dans la vie quotidienne (Paris:
Editions de Minuit, 1986), trans. Michel de Fornel Jean-Jacques Lecercle, ch. 14, “La
causalité: En partie émergente et en partie métaphorique.”
than for French, some fundamental rethinking of the theory that gener-
ates them. The Chinese language sometimes structures things very dif-
ferently, and the differences are interesting. On the other hand, similari-
ties can be interesting as well. In a world where cultural differences are
sometimes assumed to be great or even daunting, it can be refreshing to
note cases where human beings do think alike, even when, using the same
basic cranial equipment, they do not need to.
Time
Although it is universal, or at least nearly so, that human languages use

space as a metaphor for time, this metaphor can work in a variety of ways,
even within a single language. Events in time can be ordered either hori-
zontally or vertically (conceivably, of course, they could be ordered on
other axes), can “move” either toward or away from “us” at the conceptual
center. They can unfold either “in front of” you (where your face defines
your front) or behind you. It is noteworthy that in Chinese and English
(and in many other languages) events in time are arranged in one-
dimensional space. The human mind easily conceives three dimensions
of space— and, as we have seen, sometimes uses two dimensions, or three,
in describing amounts of time. But we seem uniformly to adopt only
one— a “line”—when dealing with the order of events in time. We do not
use a sheet or a block for this purpose— and it is not clear that we could do
so even if we wanted. The “time lines” humans invent are not strictly one-
dimensional, rigorously speaking. They are conceptually “fat” enough to
carry an event with them, as when we say “Christmas is coming,” and also
are fat enough to accommodate simultaneity, as when we say fu wu shuang
zhi ⽣⛵䲭㟇 ‘blessings don’t arrive in pairs’. Still, they remain basically
lines.
By the early twenty-first century, scholars had pretty well figured out
the time lines of Chinese and English. Both languages use two different
kinds of horizontal time line plus a vertical time line, and in both languages
these three time lines can work together as “mixed metaphors” without
confusing people. Moreover, the three time lines are conceptually very
similar in the two languages. It is true that modern Chinese has been in-
Metaphor 137
fluenced by Western languages, but the fundamental congruity between

Chinese and English time lines is something that well predates that influ-
ence. Graham Peck’s romantic notion that “two kinds of time” undergird
radically different worldviews has been thoroughly “deconstructed.”
The only significant difference between time lines in Chinese and En-
glish lies not in their conceptualization but in the relative frequency of
their use. Vertical metaphors of “past is up, future is down” are common
in Chinese: shang ge yue Ϟ‫ן‬᳜ ‘last month’, shuo xia qu 䂀ϟএ ‘keep on
talking’, cong tou lai ᕲ丁՚ ‘start from the head [i.e., beginning]’, and so
on. Jenn-yeu Chen, using searches of Yahoo and Google Chinese news as
data for a study published in 2007, found that 36 percent of the time meta-
phors in these sources were vertical.48 Chen did not gather parallel data
for English, but no English speaker will doubt that “past is up, future is
down” metaphors are used much less than 36 percent of the time in En-
glish. Indeed, they are rare enough that one needs to reflect in order to re-
alize that they exist at all. We refer in English to “high antiquity” and to
“descendants” from earlier generations, and we call clothes given to younger
siblings “hand-me-downs.” But there are not many such examples. Amanda
Scott counted eleven instances of shang ‘up’ or xia ‘down’ as temporal meta-
phors within a twenty-five-page 1987 Chinese drama script; she could find
no instances of “past is up, future is down” within a ninety-five-page 1972
English drama script.49
To say that Chinese and English are alike in their metaphorical con-
ceptions of time is not to say that the patterns are easy to see. On the
contrary, they present some difficult puzzles. But it is significant that the
puzzles themselves are very similar in the two languages. In the Introduc-
tion I noted the oddity of a phrase like houdai de qiantu ᕠҷⱘࠡ䗨, which
could be translated as “the future path of future generations” even though
the fi rst “future” in this phrase reflects the word hou ‘behind’ and the
second reflects qian ‘in front of’. The oddity that attends such a phrase is
not just its contradictory metaphors but the fact that the phrase is quite
48. Jenn-yeu Chen, “Do Chinese and English Speakers Think about Time Differ-
ently?,” pp. 429–430.
49. Amanda Scott, “The Vertical Dimension and Time in Mandarin,” Australian
Journal of Linguistics 9 (1989), p. 308.
clear and unremarkable to people who use it. It does not give its users any
sense of contradiction or confusion.
All by itself, the single syllable qian ‘in front of’ can illustrate the prob-
lem of a radical ambiguity that somehow leads to no confusion in under-
standing: qiancheng ࠡ⿟ is “the road ahead,” and qianchen ้ࠡ is “the dust
of the past.” Moreover, chen and cheng are both second-tone, so only that
little -g makes all the difference, orally, between past and future.50 Ning
Yu, whose book The Contemporary Theory of Metaphor: A Perspective from
Chinese provides the most systematic account we have of time metaphors
in modern Chinese, notes a similar puzzle: guoqu de shiwu nian li 䖛এⱘक
Ѩᑈ䞠 ‘in the past fifteen years’ and shiwu nian lai कѨᑈᴹ ‘for the past
fifteen years’ both refer to the same fifteen-year time span, that is, the
span that began fifteen years ago and ends now, even though one expres-
sion says the years are “gone” (qu) and the other suggests that we have (or
someone has) been “coming” (lai) through them.51
In Ning Yu’s analysis, the horizontal conception of time in Chinese has
two fundamentally different modes, which he calls case 1 and case 2. (The
vertical conception of time movement is a third and separate matter.) In
horizontal case 1, events in time form a linear sequence that moves out of
the future and toward “us” (speaker, listener, etc.) and then passes us as it
heads into the past. Yu uses a sketch of a passing railroad train to illustrate
case 1.52 This conception of time in Chinese well preceded the advent of
railroads, so the illustration is anachronistic; but conceptually it is very
apt. Case 1 can explain the phrase guoqu de shiwu nian li, because the fif-
teen years in question are the ones that have just gone by us and are now
heading into the past. The general term guoqu ‘in the past’, without any
time period attached, is also governed by case 1. Terms meaning “the fu-
ture,” like weilai ᳾ᴹ (literally “the not-yet-come”) or jianglai ᇚᴹ (liter-
ally “the will come”) obey case 1 as well, because weilai or jianglai are, as it
were, the cars near the end of the train that are still on their way toward
us. Case 1 can also explain why houtian ৢ໽, literally “the behind day,”
50. More amazingly still, people who live along the Yangzi River, from Jiangsu to
Sichuan, speak dialects that drop even the -g, or sometimes add it when it should not
be there. Yet life goes on.
51. Ning Yu, Contemporary Theory of Metaphor, p. 106.
52. Ibid., p. 107.
Metaphor 139
means “the day after tomorrow”; this is because mingtian ᯢ໽ ‘tomorrow’

has not yet arrived, and houtian is that day waiting “behind” mingtian, one
day further from us in the line of arriving days. Similarly, qiantian ࠡ໽
‘the day before yesterday’ is that day out there “in front of” yesterday in the
train of events that is headed into the past. The general terms yiqian ҹࠡ
‘in the past’ and yihou ҹৢ ‘in the future’ are also governed by case 1.
When we recall things in Chinese, we can zhuixiang 䗑ᛇ or zhuinian
䗑ᗉ, literally “chase after” thoughts or ideas. In reminiscing, we zhuiyi
䗑ᖚ ‘pursue memories’. Ning Yu points out that this mental “pursuit” of
things that have passed us by is also based on a case 1 understanding of
time.53 We look into the past and “catch up” with events that have already
gone past us. The same concept explains why zuxian ⼪‫ܜ‬, literally ‘clan
firsts’, means “ancestors,” and qianbei ࠡ䕜 ‘the in-front generation’ means
the elder generation. These are the people who were out in front on the
train of events.
This case 1 conception of time is used in English as well, where we
speak of “forebears” using the same metaphor as zuxian in Chinese, and
say “Christmas is coming” with the same metaphor of “the future is com-
ing” that weilai uses. Chinese and English also share the notion that arriv-
ing events can have impacts on people. In English we say “Christmas
brought great joy” (or maybe “brought a lot of family squabbles”) just as in
Chinese we can say chuntian dai gei le renmen wuxian meihao de xiwang
᯹໽ᏺ㒭њҎӀ᮴䰤㕢དⱘᏠᳯ ‘spring has brought people unboundedly
beautiful hopes’ (the example is from Ning Yu).54 With this kind of ex-
ample in mind, Ning Yu claims that there is a “time is a changer” metaphor,
but this is a mistake that confuses time with events in time.55 After all, it is
spring, not time, that causes the hopes; and certain events at Christmas, not
53. Ibid., p. 102.

54. Ibid., p. 114.
55. Ibid., p. 117. Yu appears to have misunderstood George Lakoff and Mark John-
son, who observe in their More Than Cool Reason: A Field Guide to Poetic Metaphor
(Chicago: University of Chicago Press, 1989, p. 40) that “because changes occur as
time passes, it is possible to personify time itself as being the agent of change.” We say,
for example, that “the passage of time brought wrinkles to the old man’s face” as a
shortcut for saying that events that occurred during the time that passed are what
brought the wrinkles.
time itself, cause the joys or the headaches. Yu is similarly imprecise when
he labels his case 1 “time as a moving object.”56 The things that move in
case 1 are events in time, not time itself. But this minor imprecision does
not affect the basic soundness of Yu’s analysis of case 1.
Case 2 describes time in a fundamentally different way. Here we are
looking toward the future, as in qiancheng ࠡ⿟ ‘road ahead’, qiantu ࠡ䗨
‘path ahead’, qianjing ࠡ᱃ ‘scene ahead’, or, in more classical style, zhan-
wang ⶏᳯ ‘looking forward’. Past events lie behind our backs, out of view
unless we turn around, which we do when we huigu ಲ乒 ‘turn around and
look’ or huishou ಲ佪 ‘turn [our] heads’. In case 2, we can be stationary, but
need not be. We can zou xiang xin shiji 䍄৥ᮄϪ㑾 ‘march toward the new
century’ and if we really get excited can do something like ben xiang can-
lan de mingtian ༨৥♓⚖ⱘᯢ໽ ‘gallop toward a brilliant tomorrow’.57
Similarly, when we turn around to look at the past (huigu or huishou), we
might not just look but actually “go backward,” wang hui zou ᕔಲ䍄, into
the past.
One key difference between cases 1 and 2 is that in case 1 it does not
matter which way we are facing but in case 2 it does. Qian and hou in case
1 are defined by the order of the train of events that is passing by us, not
by which direction we are looking. We could be mere points, without eyes
or noses at all, and words like qiantian ‘the day before yesterday’ and hou-
tian ‘the day after tomorrow’ would still make perfect sense. But in case 2
it is crucial which way we are facing, that is, where our eyes and nose are
pointing. If we were mere points, a phrase like qianjing ‘scene ahead’ or
xiang qian kan ৥ࠡⳟ ‘look forward’ would be uninterpretable, and turn-
ing around to huigu, look at the past, would be equally so.
There are difficult examples of case 2, in both Chinese and English,
where we speak as if time or something that measures time were moving
with us in the same direction, sometimes as a competitor. We can say in
English that we are “in a race with time” or that we need to “beat the
clock,” or in Chinese that shijian budeng ren ᯊ䯈ϡㄝҎ ‘time does not
wait for people’ or women bixu genshang shidai de bufa ៥Ӏᖙ乏䎳Ϟᯊҷⱘ

57. Example from ibid., p. 119.
Metaphor 141
ℹӤ ‘we must catch up with the pace of the times’.58 It can seem in such
phrases that “time itself” is moving and therefore that there must be some
kind of variant of case 2. But I think not. In such phrases the word “time,”
I believe, is always some kind of metonym (abbreviation) for other more
understandable and temporally bounded things. In shijian budeng ren, for
example, shijian could mean a deadline to accomplish some task, or the
aging process that affects everyone all of the time, or something like that.
Similarly, shidai de bufa does not imply that “time itself” is pacing forward
(what could such a phrase mean?) but that some process within time—the
spread of the Internet, the advance of human freedom, or whatever—is
going forward and needs to be kept up with. Ever since social Darwinism
entered Chinese thinking in the late Qing years, and especially after
Marxism came in a few years later, the notion that history “moves forward”
has tended to imply change for the better, or “advance.” If someone “can-
not keep up with the times” (genbushang shidai 䎳ϡϞᯊҷ), it is clearly the
times—not the things that lag behind them—that are to be preferred.
Women yao shizhong zou zai shidai de qianmian ៥Ӏ㽕ྟ㒜䍄೼ᯊҷⱘࠡ䴶
‘we want always to walk at the forefront of the times’ suggests a quest to
be the best, not merely the newest.59
The case 1 and case 2 analyses of horizontal time lines in Chinese are
clear and consistent. The reasons why puzzles can arise have to do, usu-
ally, with “interference” between the cases. If, following case 2, we look
forward (xiang qian kan) toward the future, it can somehow feel awkward,
or downright wrong, that one of the future days we are looking at is, ac-
cording to case 1, houtian or the “behind day.” But the two concepts them-
selves are clear; only their combination is awkward. And it is worth stress-
ing again that ordinary people in ordinary life are not bothered by such
puzzles; only people who try to think systematically about language are.
Graham Peck was a very intelligent man, and it is understandable that his
thinking about time got tangled. Even Ning Yu, inventor of the terms
“case 1” and “case 2” that I have been using here, occasionally shows slip-
page between his concepts of the two cases. For example, he summarizes
58. These examples are from ibid., pp. 125 and 127.

59. Examples from ibid., p. 128.
case 1 by writing that “times are moving objects, with their fronts toward
the face of the Observer.”60 Here we must understand “times” to mean not
“time itself” (which would make the statement uninterpretable) but “events
in time,” whether large or small— such as Christmas, next Saturday, the
Tang dynasty, or whatever. Yu’s “slippage” occurs because, as we have seen,
case 1 usages are indifferent to the direction the observer faces; it is case 2
that wants the observer facing this way or that. But here, in speaking of
case 1, Yu asks that the fronts of events be “toward the face of the Ob-
server.” In another sign of slippage, Yu speaks of events themselves as
having fronts or backs: “their fronts” point toward the observer, he writes.
But what can that mean? When a human head has a “front,” or a store has
a “front,” it is because of distinctive features on one side of the thing.
What could it mean for Christmas or the Tang dynasty to have a “front”?
For Christmas it might mean the fi rst few hours of the day, and for the
Tang it could mean the first few years or decades of the dynasty. But if
that is all “front” means, then it is tautological to say “the front comes first.”
Moreover, as soon as the “front” passes the observer and goes into the past,
it no longer points toward the observer. Ning Yu’s sentence should have said
something like this: “events in time are objects that move from the future
toward the observer, reach the observer, and then move away from the ob-
server into the past.” At least part of the inaccuracy in Yu’s original sentence
seems attributable to interference from case 2.
If case 1 and case 2 are kept analytically clear and are applied to the
“puzzles” that I originally noted in this discussion, the mysteries about
them pretty much dissolve. In houdai de qiantu ‘the future path of future
generations’, the hou in houdai is a case 1 metaphor, and the qian in qiantu
a case 2 metaphor. Only linguists, not ordinary people, worry about the
effects of putting them into a single phrase. Similarly, qianchen ‘the dust of
the past’ is a case 1 example, and qiancheng ‘the road ahead’ a case 2 ex-
ample. Guoqu de shiwu nian li ‘in the past fifteen years’ is a case 1 example
and shiwu nian lai ‘for fifteen years’ a case 2 example. A cliché like huishou
qianchen ಲ佪ࠡᇬ, literally “turn the head to look back at earlier dust,”
switches from case 2 to case 1 exactly in its middle. Gu qian bugu hou 乒ࠡ
ϡ乒ৢ is an intriguing case. Because qian and hou are standard opposites,
60. Ibid., p. 117.

Metaphor 143
it seems easy to analyze the phrase as “looking one way but not the other,”
but this interpretation does not square with its standard meaning (con-
fi rmed by dictionaries) of “forging ahead without considering the
consequences.”61 In gu qian bugu hou we look toward the future, but those
“consequences” (houguo, here abbreviated to hou) are also out there in the
future. No “looking in different directions” happens. Like huishou qianchen,
the phrase “switches modes” in midstream. Gu qian is case 2, looking for-
ward; bugu hou is case 1, failing to look at the houguo that are coming down
the pike. This kind of combining of case 1 and case 2 time metaphors is
common in English and other languages as well.62
We might ask in passing whether horizontal time lines in Chinese con-
ceive the past on the left and the future on the right or the other way
around. There is nothing intrinsic to terms like qian and hou that would
tilt the balance either way on this question. In theory a horizontal line
could go either way, or indeed on any axis within a horizontal plane. The
drawings in Ning Yu’s book, as well as the time lines in standard Chinese
language textbooks, put the past on the left and the future on the right.63
This direction accords with left-to-right writing across the page in Euro-
pean languages. It seems, too, to accord with a larger left-to-right bias in
modern English-speaking cultures, or at least American culture. Of the
thirty-two teams of the National Football League, twenty have logos with
a left-to-right orientation (i.e., with the panther, the raven, or whatever
61. From Beijing waiguoyudaxue yingyuxi cidianzu, A Chinese-English Dictionary

(Beijing: Foreign Language Teaching and Research Press, 1995), p. 350. I have substi-
tuted “forge ahead” for “drive ahead” in the dictionary’s rendition.
62. Lakoff calls the phenomenon “simultaneous mappings.” George Lakoff, “The
Contemporary Theory of Metaphor,” in Ortony, Metaphor and Thought, pp. 218–219.
Although correctly noting the existence of the phenomenon, Lakoff writes about it
incorrectly. He uses the example of “within the coming weeks” and claims that
“within makes use of the metaphor of time as a stationary landscape which has exten-
sion and bounded regions, whereas coming makes use of a metaphor of times as moving
objects.” (p. 219). But “within” does not imply “stationary,” as Ning Yu’s “case 1” (for
which there are many examples in English as well as Chinese) clearly shows. The
moving “train of events” does have extension and can indeed be bounded, but it is not
stationary.
63. Ning Yu, Contemporary Theory of Metaphor, pp. 99, 105, 106; Ta-tuan Ch’en et al.,
Chinese Primer: Notes and Exercises (Princeton, N.J.: Princeton University Press, 1989),
pp. 64, 74.
looking rightward) while none have right-to-left orientations. (Twelve

have neutral orientations.) On U.S. paper currency, the $10 bill shows
Alexander Hamilton looking leftward, but the gentlemen on the $1, $2,
$5, $20, $50, and $100 bills all look to the right.64 Even professional lin-
guists who work on Chinese, in which the traditional order of writing was
from top to bottom and in columns right to left, adopt the left-to-right
metaphor of European languages when they write, for example, that “di-
syllabic feet are built from left to right.”65 Do speakers of Arabic or He-
brew, where writing goes right to left, conceive time lines in that direc-
tion? In what direction did Chinese conceive time lines before the arrival
of European-language influences? This is hard to say. My point is only
that the left-to-right convention that predominates today is arbitrary, and
could well be otherwise.
Next to Ning Yu’s cases 1 and 2 we should postulate, as a case 3, the
vertical metaphor “past is up, future is down.” Jenn-yeu Chen has shown
that this occurs about 36 percent of the time in Chinese, but it is impor-
tant to note that it does not happen any time you want it to. It requires
certain grammatical contexts, like [verb] xiaqu ϟএ ‘continue [verb]ing’
or certain lexical contexts, like xingqi ᯳ᳳ ‘week’ or yue ᳜ ‘month’ but not
tian ໽ ‘day’ or nian ᑈ ‘year’. (There are dialects in which shang ‘up’ can
work with tian ‘day’, but these patterns are still set, and not up to the speak-
er’s preference.) There is a word xiachang ϟഎ ‘down-field’, meaning “sit-
uation one ends up in,” but no xiajing ϟ᱃ ‘down-scene’, even though this
metaphor would make just as much sense for the same idea. Vertical meta-
phors also cannot work with certain other common metaphors, like daolu
䘧䏃 or lucheng 䏃⿟, that mean “road.” Perhaps because vertical roads are
awkward to picture in the mind, daolu and lucheng keep to horizontal
metaphors like qian, hou, and guo 䖛 ‘cross’.
64. I do not wish to overstate this point. Among major league baseball teams, for
example, twenty-three have orientationally neutral logos, five look left, and only two
look to the right. Of the common U.S. coins, only the penny looks squarely to the right;
the half-dollar, quarter, and dime all look to the right, and until recently the nickel did
as well. On 2006 nickels, like 2007 dollar coins, the gaze is slightly to the right.
65. Bingfu Lu and San Duanmu quoting Feng Shengli, “On the ‘Natural Foot’ in
Chinese,” in Bingfu Lu and San Duanmu, “Rhythm and Syntax in Chinese: A Case
Study,” Journal of the Chinese Language Teachers Association 37, no. 2 (May 2002), p. 127.
Metaphor 145
I have noted that English also uses “past is up, future is down” meta-
phors (as when we say “down through the ages” or speak of a person’s
“descendants”) but uses them less often than Chinese. Television hosts
sometimes refer to “the top of the show” to mean its beginning. Hence it
is a puzzle that in English we also say “upcoming events” or “let’s move
the meeting up a week,” because these usages seem to suggest a “past is
down, future is up” metaphor that goes oppositely from the normal direc-
tion. Lakoff and Johnson try—not quite successfully, in my view—to ex-
plain this usage by observing that, as physical objects come nearer to us,
their images appear larger and the tops of the images are therefore farther
“up.”66 To me it makes more sense to abandon the notion that up has to
have a literal basis in such phrases and just let it be (metaphorically) hori-
zontal. Up plainly does have such horizontal uses, even when used strictly
spatially, as in “up to the water’s edge.” Since water necessarily must be
below the ground at the water’s edge, “up to the water’s edge” cannot be a
vertical conception and therefore must be horizontal. But if “up” can be
horizontal, then upcoming events might simply be events that, in accor-
dance with Ning Yu’s case 1, are headed toward us horizontally. Events
“down the road” might similarly exist within a horizontal conception.
A number of cognitive scientists have speculated on what the “physical
basis” for vertical time metaphors might be, but their results have in gen-
eral been unsatisfying. Ning Yu has argued that “later is down” can be
derived from the horizontal “case 1,” in which later things are “behind” in
the train of events, because, Yu reasons, when one stands up from a lying
position, one’s head, which was in the front-of-the-train or “earlier” posi-
tion, is now up, which therefore becomes the “earlier” position in vertical
terms; meanwhile one’s feet, which began in the back-of-the-train posi-
tion or “later” position horizontally, are now down, in the “later” position
when measured vertically.67 But Yu gives no reason why one metaphor-
ically “lies” with one’s head or feet either this way or that. He also gives
no reason why we should employ horizontal case 1 instead of horizontal
case 2 to interpret the vertical metaphor, and horizontal case 2 would
seem to lead to an opposite conclusion. Hence his effort is valiant but

67. Ning Yu, Contemporary Theory of Metaphor, pp. 111–112.
unpersuasive. Amanda Scott has suggested that vertical time might derive
from the rise and fall of the sun in the sky. Here, to be sure, we are speak-
ing of a natural vertical movement (only an apparent one, of course— as we
know from Copernicus) that corresponds to earlier and later times.68 The
problem with Scott’s explanation is that the sun moves both up and down
in equal measure, and Scott has no explanation for why such movement
could be the basis for a metaphor that, in both Chinese and English,
moves only down.
A letter to the editor of the Wall Street Journal, commenting on an op-
ed essay by Lera Boroditsky, claims that gravity is the basis: “if you drop
something, the past is where it was (above) and the future is where it will
be eventually (below).”69 Other speculations about the origins of vertical
time metaphors for Chinese in particular look not to “physical bases” but
to human creations like the writing system. The plausibility here comes
not just from the fact that the next character one writes is indeed below the
current one (if one’s paper is vertical) or that to continue writing is to xie
xiaqu ‫ݭ‬ϟএ ‘write down’ in a literal sense. It comes as well from a general
drift in temporal thinking that the writing system both reinforces and
obeys. But this “general drift” is also what makes the theory questionable
as a causal theory. Which way does the causality go? Did a predilection in
thought lead to the writing custom, or did the direction of writing create
the habit of thought? (Once the pattern was established, causality no doubt
continued for centuries to go both ways.) Scott speculates that other
metaphorical uses of shang ‘up’ to mean “high (in status), excellent” and xia
‘down’ to mean “low, inferior” might seep into the connotations of the two
words when they mean “past” and “future.”70 But this seems far-fetched. It
is true that the ancient past in China has been regarded as glorious, but to
hold that in modern Chinese shang for the past and xia for the future reli-
ably connote “better” and “worse” would mean, for example, that future
weeks and months (where xia is used) are consistently drearier in the Chi-
nese mind than future days and years (where horizontal metaphors are
used), and there is no evidence for such a difference.
68. Scott, “Vertical Dimension and Time in Mandarin,” p. 307.

69. Letter from Sean Cox, July 31, 2010.
70. Ibid., pp. 312–313.
Metaphor 147
In sum we can say that there are four kinds of metaphorical expression
that use space to conceive events in time in Chinese: two kinds of hori-
zontal time lines (Ning Yu’s cases 1 and 2), a vertical time line, and, as a
fourth category, a variety of mixtures among the fi rst three. English us-
age draws on the same four categories, although in somewhat different
proportions. After fi nishing his systematic study of time metaphor in
Chinese and comparing his findings to Lakoff’s for English, Ning Yu feels
that he has found “strong evidence in favor of certain universals in the
human cognition of time.”71 This claim, which echoes Alverson’s hypoth-
esis of “a universal template” in the human experience of time, is persua-
sive. Are there “universal templates” in human experience of other things,
such as color?
Color
It is easy to note that in any language colors are used metaphorically, as in

English when we have the blues, are green with envy, are yellow-bellied
cowards, get red-faced (or even livid, i.e., white) with anger, compile
blacklists, and so on. Is all of this cultural invention, or are there univer-
sals to be found? The question depends on a prior question: to what extent
are the definitions of colors the same across languages? That is, before we
study how the metaphorical uses of hong 㑶 and red compare, it seems im-
portant to know how much we are comparing the same things: how well
do hong and red (and rouge in French, akai in Japanese, and so on) pick out
the same spans on the natural color spectrum?
This question is not the same as the philosopher’s metaphysical ques-
tion of whether one person’s qualia (phenomena of subjective experience)
are the same as another’s. Your experience when you see red and mine
when I see red might be different, even radically different, even while
our talk using the word “red” proceeds without a hitch. But this philo-
sophical question is untestable and beyond our scope. Here, by “see red” I
mean only “have qualia that allow one to use the word ‘red’ in a way that
accords with how other people use it (whatever their qualia may be)” or,

alternatively, “perceive with eye and brain light waves that fall within a cer-
tain span.”
A cursory comparison of how the Chinese and English languages label
colors lends considerable support to a view that has been called “linguistic
relativity” in color definition. (The view derives from Sapir and Whorf
and was dominant around the middle of the twentieth century.) In Chi-
nese, for example, the term huang 咘 does not correspond well to any En-
glish word. Dictionaries cite “yellow,” but huang covers much more of the
spectrum than yellow does. It begins with yellow, spans all of tan, and
goes pretty far into brown. As noted in the Introduction, China’s Huanghe
咘⊇, Yellow River, is brown, not yellow. A huang gou 咘⢫ in Chinese is a
brown dog, or at the lightest a tan one, hardly a yellow one. I teach classes
wearing huang pixie 咘Ⲃ䵟, ‘brown leather shoes’ and might think twice
about going to class if they actually were yellow. When Western languages
arrived in China with their major distinction between “brown” and “yel-
low,” modern Chinese responded by using other brown-like words— such
as hese 㻤㡆 ‘earth-tone brown’ and zongse ẩ㡆 ‘palm-fiber brown’—to
serve as counterparts of “brown.” But in daily-life Chinese usage, hese and
zongse have never really competed with huang, and, when they do get used,
they refer to much narrower bands on the color spectrum than huang. A
similar problem attends qing 䴦, which can be green as in qingjiao 䴦Ủ
‘green pepper’, blue as in qingtian 䴦໽ ‘blue sky’, or even black as in
qingbu 䴦Ꮧ ‘black cloth’. Examples such as these would seem to lead to the
conclusion that divisions within the color spectrum are arbitrary. Indeed,
we might feel that they must be arbitrary. The spectrum is, after all, a
continuum; no natural line separates blue from green, orange from red,
and so on, so any language of course must be arbitrary in saying where one
color ends and the other begins. Or so it would seem.
In their 1969 book Basic Color Terms, Brent Berlin and Paul Kay pre-
sented research results showing that we should not make this assump-
tion.72 Berlin, Kay, and their students found that some factors that go into
the naming of colors are constant across languages. They presented speak-
ers of twenty different languages with varying shades of red, blue, green,
72. Brent Berlin and Paul Kay, Basic Color Terms: Their Universality and Evolution
(Berkeley: University of California Press, 1969).
Metaphor 149
and so on, and then asked them to identify which was the most standard
or representative example within each family of shades. They found con-
siderable agreement on which were the “best red,” “best blue,” and so on.
This agreement held not only among speakers of the same language but
also— and here is the key point— across speakers of different languages.
The favored examples of red, rouge, hong, akai, and so on tended to be the
same shade regardless of language. Berlin and Kay gave the name “focal
colors” to these favored shades: focal red, focal blue, and so on. They
found that agreement was clearest for what they identified as the “basic
colors” black, white, red, green, yellow, blue, and brown. After comparing
data in secondary material for seventy-eight languages in addition to the
twenty they had studied directly, Berlin and Kay concluded that “the foci
of basic color terms are similar in all languages.”73 The fringes of a color’s
definition—where red blends into orange, and so on—varied considerably
among languages, but the “focal” cases did not.
Berlin and Kay make some further claims that seem doubtful, in part
because their own color terms, which are based in English, lead them to
pose questions that preclude certain answers that other languages might
produce. For example, they argue that there is a natural order in which
labels for the basic colors enter human languages. They find, arguing
from empirical survey, that all languages have black and white, and that,
when a third basic color enters a language, it invariably is red. For lan-
guages that go to four basic colors, the next to be added is either green or
yellow, and when there are five, both yellow and green are on the list. For
six-color languages, blue is added; for seven, brown; and after that purple,
pink, orange, and grey all arrive, although these four observe no particu-
lar order.74 Berlin’s and Kay’s extensive data do seem to confi rm this pat-
tern, but one needs to wonder how much English-language categories
shaped the very data they were working with. When they say that green is
the fourth or fifth color to arrive in languages, a Chinese speaker might
ask “do you mean here qing 䴦 or lü ㍴?” (It is not clear whether there are
“focal” shades of both these Chinese notions of “green.”) Or one might
ask “In what order does huang arrive?” Would it be in fourth or fifth place
73. Ibid., p. 10.

74. Ibid., pp. 2–3.
(where “yellow” is) or seventh (where “brown” is)? These objections do

not seriously impair Berlin’s and Kay’s argument for the existence of “fo-
cal colors”; but they do question some of their related claims.
If “focal colors” exist, it is natural to ask why they exist. Studies in the
late 1960s of the neurophysiology of the eye of the macaque, a primate
whose visual system is close to that of humans, showed that four “re-
sponse cells” in the eye of the macaque had to do with the perception of
red, green, yellow, and blue, while two others had to do with brightness,
which helps to explain the perception of black and white.75 Later cogni-
tive scientists, including Paul Kay and George Lakoff, have claimed that
neurophysiology can explain the phenomenon of the focal colors.76 Lakoff
holds that in addition to neurophysiology, there is a “something else” (his
term) that might account for “different values in different cultures.” But
he does not explain clearly what that “something else” might be. (We are
told only that it “consists of a complex cognitive mechanism incorporating
some of the characteristics of fuzzy set theory union and intersection” and
that it “has a small number of parameters” that make the difference.)77 In
any case, at bottom, the Kay-Lakoff argument depends on the assumption
that a relatively simple physiological source (a certain neuron for “red,”
for example) can explain a natural preference for a certain hue. Combina-
tions of firings for red and yellow produce orange, for blue and red pro-
duce purple, and so on, while subtler shades presumably involve more
complex firings.
Kay and Lakoff’s theory raises thorny issues in what Western philoso-
phers have called the “mind-body problem.” How do neural fi rings in the
(physical) brain relate to perceptions— such as colors—in the mind? Kay
and Lakoff do not address this question but do make an assumption about
its answer when they assume that relatively simple neural firings corre-
spond to perceptions of “focal” colors. That may be so; but it is also
75. R. L. DeValois, I. Abramov, and G. H. Jacobs, “Analysis of Response Patterns of
LGN Cells,” Journal of the Optical Society of America 56 (1966), pp. 966– 977; R. L De-
Valois and G. H. Jacobs, “Primate Color Vision,” Science 162 (1968), pp. 533–540.
76. Paul Kay and Chad McDaniel, “The Linguistic Significance of the Meanings of
Basic Color Terms,” Language 54, no. 3 (1978), pp. 610– 646; Lakoff, Women, Fire, and
Dangerous Things, pp. 26–30.
77. Lakoff, Women, Fire, and Dangerous Things, p. 30.
Metaphor 151
possible that it is not so. Might it not also be that the firings that produce
what a human being chooses to call “best red” are complex? Kay and
Lakoff seem to have an assumption that a comparatively simple or basic
neural experience will lead a person to feel “right” or “primary” (or some-
thing like that) about the corresponding subjective impression. But is this
necessarily so?
Consider, by analogy, the taste sensation. If we gave bits of chocolate of
slightly different flavors to people (professors of Chinese literature do not
have research budgets for this kind of experiment, so this is a thought-
experiment only) and found that the subjects generally agreed that sam-
ple 7 was the “most chocolate-y” of the lot, would we want to assume
that “most chocolate-y” corresponds to relatively simple physical stim-
uli? Or to stimuli created by chocolate of the simplest chemical compo-
sition? I think we would not want to make such assumptions. The addi-
tion of sweetness from an admixture of sugar, for example, might well
enhance a sense of chocolate-ness. And what about wine? Would “focal
chardonnay” be the one that corresponds to simpler physical stimuli
than those of other chardonnays? If so, any connoisseur could tell us
that focal chardonnay is not the best chardonnay. I don’t mean to sati-
rize Kay and Lakoff here; they may still be right. But their argument
does contain a hole. A correspondence between simpler neural activity
and more “focal” subjective experience needs to be demonstrated, not just
assumed.
But even if the connections to neural bases are not well understood, the
empirical finding of focal colors is an important result. It can give us con-
fidence, as we begin to compare metaphorical uses of color-terms across
languages, that we are dealing with the “same thing”— or more precisely,
in the jargon of the field, the same “vehicle” or “defi ning concept.” Thanks
to the discovery of focal colors, we can be fairly certain that hong and red,
hei 咥 and black, and bai ⱑ and white are very close to the same for native
speakers of Chinese and English. Hui ♄ ‘grey’ and zi ㋿ ‘purple’ also
probably correspond well. Special problems involving colors like qing and
huang will require that we be careful before using English to interpret
Chinese metaphors that use such colors.
Now let us turn from colors themselves to the question of their meta-
phorical uses. One of the bases for commonality in color metaphors across
languages is (as noted earlier) what metaphor theorists call the “physical”
or “experiential” bases of metaphor. I have noted the example in which
blood rushing to the face during anger apparently has led to metaphors
like “red in the face” or “seeing red” in English and qi hong le lian ⇨㑶њ㜌
‘angered [oneself] red-faced’ or mian hong er chi 䴶㑶㘇䌸 ‘face and ears
red’ in Chinese. Note also that blood gathering in the facial capillaries
can have causes other than anger and that, correspondingly, this par ticu-
lar “physical basis” can undergird metaphors for conditions other than
anger: embarrassment, for example, as when we say lian hong 㜌㑶 in Chi-
nese or “red-faced” in English. Robust health or cold air can also bring
blood into the facial capillaries, which leads us to say manmian hongguang
⒵䴶㑶‫‘ ܝ‬bright red fills face’ in Chinese or “rosy cheeks” in English.78 In
extreme frustration, so much blood might rush to the head that we say, in
English, one is “blue in the face”; intense anger might also bring blueness
or, as both Chinese and English suggest, drain the blood from the face, so
that one is “livid” in English or, in Chinese, qi de lian hong yi zhen bai yi
zhen ⇨ᕫ㜌㑶ϔ䰉ⱑϔ䰉 ‘so angry that the face alternated between red
and white’.
The commonality by which “green” tends to mean “young” seems to
have an obvious physical basis in the vivid greenness of young plants. In
English we say a neophyte is “green,” and in Chinese qingnian 䴦ᑈ ‘green
years’ refers to young people. The physical basis connecting “green” with
“young” is perhaps most apparent in qingchun 䴦᯹ ‘green spring—youth’,
where the young sprouts of spring embody both greenness and newness.
In Su Xiaokang and Wang Luxiang’s famous 1988 television series He-
shang (River elegy), huang ‘brown’ served as an extended metaphor for
“conservative, hidebound, or inward-looking.” Even this had an indirect
physical basis. Huang is the color of the earth in the inland region of
north China; that is where, for Su and Wang, the earthbound, uncosmo-
politan side of Chinese culture was rooted; it is also where the Commu-
nist Party had its base in Yan’an in the 1940s; a continuing “feudal” men-
tality was still alive in the 1980s, according to the television series, and it
could be symbolized as huang or “earth-colored” culture.
78. For this and a number of other promptings about the metaphorical meanings of
colors, I owe a debt to Ye Minlei.
Metaphor 153
But if these color metaphors can be traced to physical bases, there are a
large number, probably a majority, that cannot be. These other metaphors
apparently have arisen from more accidental causes and are better viewed
as cultural creations in which the connection of ideas to specific colors is
arbitrary. Why, for example, are we “green with envy” in English but have
hongyanbing 㑶ⴐ⮙ ‘red eye disease— envy’ in Chinese? Why are porno-
graphic films “blue movies” in English but huangse dianying 咘㡆⬉ᕅ ‘yel-
low movies’ in Chinese? Why is “blue” in English “sad, depressed,” whereas
qing ‘blue or green’ in Chinese can be upbeat, as in mingchui qingshi ৡൖ䴦৆
‘name suspended through (blue or green) history—go down in history’. Is
it not just arbitrary that “red” in English signifies danger (red flag, red
alert, etc.) and lü mao 㓓ᐑ ‘green hat’ in Chinese is what a cuckold wears?
To account for all such ad hoc color metaphors would be to write a short
encyclopedia.
A single color can have very different connotations even within the
same language. In Chinese, bai ‘white’ is a good thing in jiebai ⋕ⱑ ‘clean,
pure’ or qingbai ⏙ⱑ ‘stainless, immaculate’, but stands for evil in the bai-
lian ⱑ㜌 ‘white face’ of popular opera. Red is especially rich and versatile.
In addition to “angry,” “embarrassed,” “healthful,” and “envious,” noted
above, hong can mean “hot, popular” as in hong mingxing 㑶ᯢ᯳ ‘popular
star’; “favored,” as when person A is a hongren 㑶Ҏ in the view of person B;
“lucky,” as when somebody zou hong 䍄㑶 ‘gets lucky’ or has hongyun 㑶䖤
‘good luck’; “beautiful,” as in hongfen jiaren 㑶㉝ՇҎ ‘red-powder beauty’
or hongyan yilao 㑶买ᯧ㗕 ‘red face ages easily—beauty does not last’; “loyal”
as in chidan zhong xin 䌸㚚ᖴᖗ ‘red gall loyal heart—utter devotion’; and
even “ordinary, pedestrian” as in kanpo hong chen ⳟ⸈㑶ᇬ ‘see through the
red dust [of the ordinary world]’.
Modern borrowings from Western languages have added even more
meanings of red to Chinese. “Red ink” meaning debt has come into mod-
ern Chinese as chizi 䌸ᄫ ‘red characters’. Red to mean “Communist” has
given new meaning to hongqi 㑶᮫ ‘red flag’, hongjun 㑶‫‘ ݯ‬red army’, and
many other terms. Some of the traditional positive connotations of hong—
such as “favored” and “loyal”—have seeped into the new usage of hong as
“Communist.” The hong wu lei 㑶Ѩ㉏ ‘five red categories’ (workers, poor
peasants, revolutionary officials, revolutionary soldiers, and revolution-
ary martyrs) were the bedrock Communist people during the Cultural
Revolution, but also the “favored” and “loyal” ones—making them safe
on both grounds from persecution.
Other obvious modern borrowings of color metaphors include “black”
to mean “secret” or “nefarious,” in terms such as hei shehui 咥⼒Ӯ ‘black
society—underworld’, hei mingdan 咥ৡऩ ‘blacklist’, heishi 咥Ꮦ ‘black-
market’, and others. After a borrowed metaphor takes root, it can take on
a power to structure new concepts (can be a “metaphor people live by,” as
Lakoff and Johnson would say) and produce terms in the borrowing lan-
guage that did not exist in the lending language. For example, Chinese
farmers who live in cities without legal urban residence permits are said to
have heihu 咥᠋ ‘black residence’ in the cities. Huang meaning “porno-
graphic” arose in the twentieth century in China after evolving from the
English term “yellow journalism,” whose roots were in New York of the
1890s. In the United States, though, the term has meant scandal-mongering
and sensationalism, not specifically pornography, as it does in Chinese.
One possibility in explaining similarity in metaphors across languages
is, of course, simply coincidence. This happens, after all, even with ordi-
nary terms, such as low meaning “mean or base” in English and lou 䰟
meaning “mean or base” in Chinese. In both languages, “colorful” can
mean “rich, interesting.” We can have a “colorful life” in English and a
duoziduocai de shenghuo ໮࿓໮ᔽⱘ⫳⌏ ‘multiappearance multicolor life’
in Chinese. Is this coincidence, or might there be a subtle “experiential
basis” at work—in that different colors might imply variety, therefore in-
terest, maybe even travel?
Although color metaphors can seem highly contingent, they also can
have considerable power. In October 1988, at a high-level conference on
Sino-American scholarly exchange, the distinguished American historian
Frederic Wakeman, of the University of California at Berkeley, borrowed
Su Xiaokang’s term “yellow” to describe the inward-looking side of Chi-
nese civilization and contrasted it with the “blue” side (also Su Xiao-
kang’s term) that looked across the ocean (hence “blue”) at the interna-
tional world. Wakeman observed that scholarly exchange might play a
role in China’s transition toward a blue civilization out of a yellow one.
Yan Dongsheng, chief of the Chemistry Division of the Chinese Acad-
emy of Sciences, heard the term “yellow civilization” as “pornographic
civilization” and rose to denounce Wakeman in terms so impassioned that
Metaphor 155
no one at the meeting (including me) could recover in time to explain that
a color metaphor had been misunderstood. Explanations had to wait until
the next day.
Up and Down
We have seen how shang ‘up’ and xia ‘down’ are useful in the vertical concep-
tion of time in Chinese. But shang and xia have a wide variety of other uses
as conceptual metaphors, as do up and down in English. The two languages
sometimes resemble each other in these matters, but sometimes they do
not—and sometimes they present puzzles. Why, in English, after chopping
a tree down, do we chop the wood up? Why, in Chinese, do we memorize
something down (beixialai 㚠ϟᴹ) but remember it up ( jiqilai 䆄䍋ᴹ)?
The conceptual metaphor “more is up, less is down” is well established in
both languages. This fact is not surprising, because, as noted, the “physical
basis” of the metaphor is universal: for speakers of any language, “more” of
something tends to create a pile, and a pile top tends to go up. During the
twentieth century—in Chinese, English, and many languages— abstract
uses of the metaphor, in phrases like tonghuo pengzhang shangqu le 䗮䋻㝼
㚔Ϟএњ ‘inflation has gone up’, became increasingly common. Uses of
terms like gao 催 ‘high’ and di Ԣ ‘low’ to fit the metaphor also seem to
have increased: xinshui hen gao 㭾∈ᕜ催 ‘salary high’ means more money,
wendu jiang le ⏽ᑺ䰡њ ‘temperature fell’ means less heat, and so on. Such
usages are now so common in both Chinese and English that we usually
forget that they are metaphors. But they are: originally, after all, “up” and
“more” are unrelated concepts.
From up meaning “more in quantity” it is but a small shift—which Chi-
nese, English, and many languages have made—to the meaning “more in
quality.” In Chinese, both ancient and modern, a shangce Ϟㄪ is a good
plan, a xiace ϟㄪ an inferior one. Mao Zedong urged students to haohao
xuexi tiantian xiangshang དདᄺд໽໽৥Ϟ ‘study hard and go up [i.e., get
better] every day’.
Such examples usually refer to technical quality, but examples of moral
quality are even more numerous. Up connotes virtue; down, baseness. We
can say in English that someone is “upright” or “high-minded,” and in
Chinese gaoshang 催ᇮ ‘high-esteemed—noble’ or gaoming 催ᯢ ‘high-

bright—enlightened’. In English we “stand up for principle,” and in Chi-
nese tingshen er chu ᤎ䑿㗠ߎ ‘straighten the body and go out— step boldly
forward’. These images vary, but all reflect a connection between “up” and
“morally right.” Similarly, a moral scoundrel in English can be “low-down”
and others might need to “stoop” to his level. In Chinese such behavior is
beixia थϟ ‘base, low’, and perhaps xialiu ϟ⌕ ‘mean, obscene’. There do
exist exceptions to the pattern, however, and these seem perhaps more
common in Chinese than in English. Ning Yu cites examples where Chi-
nese uses the metaphors basically the other way around: wenzhong 〇䞡
‘stable and heavy’ and chenwen ≝〇 ‘weighty and stable’ both mean
“solid and reliable” in a positive sense, whereas qingfu 䕏⍂ ‘light and float-
ing’ and piaofu ⓖ⍂ ‘drifting and floating’ mean “frivolous, flighty” in a
negative sense.79 Similarly, in English “flighty” combines up with a negative
connotation.
Another use of “up” or “high”—in Chinese, English, and many
languages—is to indicate more status or power within an administrative
structure. In English an official can be “high-ranking” and on completion
of a term of ser vice might “step down.” In premodern Chinese, da ໻ ‘big’
and xiao ᇣ ‘little’ were used in this regard: a high official was a daguan ໻ᅬ
‘big official’. But in recent times, the term gaoji guanyuan 催㑻ᅬਬ ‘high-
ranking official’ has become more common.
Desmond Tutu and others who ran the Truth and Reconciliation Com-
mission (TRC) in South Africa after 1994 consciously sought to counter
the “up is good” and “up is powerful” metaphors that were implicit in both
English and South African languages by having everyone at hearings sit
on the same level. “We had to avoid any impression that [the witnesses]
79. Ning Yu, Contemporary Theory of Metaphor, p. 65. The terms gexia 䭷ϟ, literally
‘below the pavilion’ and zuxia 䎇ϟ, literally ‘beneath the feet,’ both of which are ways
to say “you” in premodern polite language, also might appear to be exceptions, be-
cause normally Chinese honorifics elevate the person to whom they refer. But these
are not exceptions. The sense of gexia (or zuxia) is ‘you, the one beneath whose pavil-
ion (or feet) I (stand to) address’. The dictionary Cihai explains that, because it might
seem too audacious to address an august personage directly, gexia is used to make the
approach indirect by addressing the servants and retainers who are arrayed below.
Cihai bianji weiyuan hui, Cihai (Sea of words) (Shanghai: Cishu chubanshe, 1979),
p. 2016.
Metaphor 157
were in the dock,” Tutu has written, “so they sat on the same level as the
TRC panel hearing their testimony.”80
In a metaphor that combines the concepts “up is more power” with “up
is to go on stage—that is, be more visible and responsible,” Chinese says
shangtai Ϟৄ ‘move up onstage’ when an official assumes office and xiatai
ϟৄ ‘move down offstage’ for leaving office. “Up” is sometimes used in
Chinese not only for a person who is administratively higher but for a geo-
graphical place that is. In Wu dialect, for example, shangqu Ϟএ ‘go up’ is
used to mean “go to the city” while xiaqu ϟএ ‘go down’ means “go to the
countryside.”81 It is probably more than just administration that explains
the “up”-ness of an urban area, because the city has more of other things as
well: markets, entertainment, population, and general bustle. But some-
times the political implications are clear. When a political leader makes an
inspection tour of the countryside, it is xiaxiang kaocha ϟе㗗ᆳ ‘go down
to the countryside to inspect’. When youth in the Mao era left the cities
to “learn from the peasantry,” one might have expected the prestige of the
peasantry— a “revolutionary class” in Maoist ideology and the putative
“teachers” in this project—to put them higher than the youngsters com-
ing from the cities. But this did not happen, at least not at the level of the
conceptual metaphors in daily life. The phrases xiaxiang ϟе ‘go down to
the countryside’ and xiafang ϟᬒ ‘transfer down’ remained standard.
Whether people noticed the implication or not, the metaphorical lowness
of the countryside outweighed the political ideal that the peasants should
be teachers.82
Lakoff and Johnson (and many others) have noted “happy is up, sad is
down” as another example of a standard metaphorical meaning of “up” in
English. We say in English that our spirits “rise” or that they get a “boost”;
80. Desmond M. Tutu, No Future without Forgiveness (New York: Doubleday, 1999),
p. 110.
81. I am grateful to Ye Minlei for this example. Email message to author, March 7,
2008.
82. The full four-syllable phrase shangshan xiaxiang Ϟቅϟе only reinforces the
metaphoric significance of xia, because shang here is literal. Going up a mountain is a
matter of altitude; going “down” to the countryside is seldom a matter of altitude. (On
average, cities are at lower altitudes than farms.) So xia in the phrase needs to have a
metaphorical understanding.
but our spirits can “sink,” after which we feel “down,” “low, or “depressed.”83
We can be “on cloud nine”; we also can be “down in the dumps.” Some
theorists have speculated on what possible “physical basis” there might be
for the notion that “happy is up.” Lakoff and Johnson, and Ning Yu, think
that such expressions “arise from the fact that as humans we have upright
bodies.”84 Perhaps so, but the assertion needs further explanation. Is the
point that when our bodies are not upright we are often not fully alert,
as when we are ill or asleep? Can we say in general that we are happier
standing up than lying down? Zoltán Kövecses has a different theory. He
suggests that “up” reminds us of birds—which are above us, and seem
free, and thus seem happy.85 But Kövecses’s explanation seems to risk cir-
cularity. Do birds, who seem happy, cause us to associate up with happy? Or
do we, perhaps, attribute happiness to birds because they are “up there”?
This is hard to say.
Whatever the physical basis for the “happy is up, sad is down” meta-
phor might be, it is a simple matter to show that it occurs in Chinese as
well as in English. In Chinese we have gaoxing 催݈ ‘high spirit—happy’
or xingfen ݈༟ ‘spirit-lift—(pleasantly) excited’ on the one hand and diluo
Ԣ㨑 ‘low fall— downcast’ or chenmen ≝䯋 ‘sink stifled—in low spirits’ on
the other.86 On the whole, though, up and down as metaphors for happy
and unhappy may be less common in Chinese than in English. David
Moser has correctly observed that Chinese tends to use container meta-
phors such as open or unfettered for happy and boxed in, cornered, or faced
with barriers for unhappy.87 Thus kaixin ᓔᖗ ‘open heart’ is “happy” and xi
chu wang wai ୰ߎᳯ໪ ‘joy overflows outwards’ is something like “exuber-
ant,” and nanguo 䲒䖛 ‘hard to cross over (something)’ means “sad.” Kai
also has associations of “free, happy” when used as a verbal complement
in phrases such as jiekai 㾷ᓔ ‘untie open— solve (a puzzle or mystery)’,
or xiangkai ᛇᓔ ‘think open—rise above (something)’— a phrase that can

84. Ning Yu, Contemporary Theory of Metaphor, p. 61; Lakoff and Johnson, Metaphors
We Live By, p. 15
85. Zoltán Kövecses, “Happiness: A Defi nitional Effort,” in Metaphor and Symbolic
Activity, 6:29–46.
87. Letter to the author, October 25, 1993.
Metaphor 159
be used in advising a person how to get over a personal loss. Shunxin 乎ᖗ

‘accord with heart’ indicates a happy contentment, and when things go
wrong a feeling of bushun ϡ乎 ‘nonaccord’ arrives. The word men 䯋 cap-
tures the metaphor well all by itself (even the character— a heart enclosed
by a door— suggests the metaphor). Spoken in first tone, men refers to
stuffy air, muffled sound, keeping quiet (i.e., not talking), or shutting one-
self indoors; it is also what you do to tea leaves when you trap them under
a tight lid and steep them in hot water. In the fourth tone, men means
bored, depressed, or sealed off. These uses of “open” and “closed” in Chi-
nese as metaphors for “happy” and “unhappy” exist alongside “up” and
“down” for the same ideas. As with different directions for time lines, the
mix causes few, if any, problems in getting through daily life.
There is another set of up and down metaphors that seem conceptually
at odds in both Chinese and English. Sometimes up suggests “unknown”
or “unsettled” and down almost the opposite. We say “that’s still up in the
air” and then, when the matter is decided, that “it’s settled.”88 On the other
hand, we sometimes use up to suggest “finished” or “all set,” as when we
“chop up,” “wrap up,” or “dress up.” The meaning of up is elusive here, but
it does seem to suggest that something did in fact get chopped or wrapped
or dressed. The effect seems to survive even when the result is not a pleas-
ant one. If we “mess up” or “screw up,” up still seems to spotlight the end
result—rubbing it in, as it were.89 The fi xity of up and indeterminacy of
down are also illustrated in what Lakoff and Johnson call the “control is
up” metaphor, for which they cite “I’m on top of the situation” as an ex-
ample.90 Here the notion is that things down below, far from “settled,” are
in some kind of confusion that calls for control.
Although parallels with Chinese are not perfect, they are sufficiently
similar to be intriguing. In English when we memorize something we
88. These examples are from Lakoff and Johnson, Metaphors We Live By, pp. 120,
137.
89. I am claiming here only that “fi nished” is a common meaning of up after a verb,
not that it is the only one. A phrase like “put up with idiocy” would likely need a dif-
ferent explanation.
90. Lakoff and Johnson, Philosophy in the Flesh, p. 53. The authors hypothesize that
the “experiential basis” for this metaphor is that “it is easier to control another person
or exert force on an object from above, where you have gravity working with you.”
might say we have it “down pat,” while in Chinese we say it has been bei
xialai le 㚠ϟᴹњ ‘memorized down’. Chinese also uses luoshi 㨑ᅲ, liter-
ally “descend to facts,” when something theoretical, like a government
policy, is put into practice. Of a group of noisy children Chinese can say
they anjing xialai ᅝ䴭ϟᴹ ‘quiet down’, just as English says “quiet down.”
Something can be “up for grabs” in English, and in Chinese can “hang”
(xuan  ) awaiting decision, as in zhe shir buneng lao xuanzhe 䖭џ‫ܓ‬ϡ㛑㗕
 ⴔ ‘this matter can’t keep on hanging there’.91 These several examples,
in both languages, illustrate the conceptual metaphors “up is unknown,
or in flux,” and “down is known or settled.”
What about the opposite case, where up suggests “all set” or “in con-
trol” and down suggests “indeterminate” or “in flux”? Chinese seldom
uses shang Ϟ the way English uses “up” in a phrase like “chop up.” (Ganshang
䍊Ϟ ‘catch up’ is an unusual exception.) But Chinese has other phrases
that fit the metaphor fairly well. We just saw that luoshi can mean “de-
scend to facts”; but Chinese also says luokong 㨑ぎ ‘descend to emptiness’
when an idea or plan falls through or comes to naught. A term like zhenya
䬛य़ ‘repress’ (zhen and ya both mean “press down”) also accords with the
metaphor in that it suggests that control is on top and chaos, which needs
control, reigns below. Qiyi 䍋Н, literally “rise righteousness,” a common
term for rebellion, might seem from the point of view of the rebels to
mean “headed upward, toward the sky.” But from the viewpoint of the
ruling dynasty, the rebellious activity is taking place “down below,” and qi
thus means “moving upward toward us,” which creates the need to zhenya
or “press down.”92 Other examples in which “here, the present level” seems
standard and clear while “the level below” seems murky or questionable
are phrases like shuo bu shanglai 䂀ϡϞ՚ ‘can’t come up with [a word, ex-
planation, etc.]’ or xiang bu qilai ᛇϡ䍋՚ ‘can’t recall’, in which shang and
qi tell us that the movement is upward or rising and lai tells us that it is
91. The example is from Beijing waiguoyudaxue yingyuxi cidianzu A Chinese-

English Dictionary, p. 1151.
92. It is worth noting that another conceptual metaphor in Chinese, which might be
called “control is clenching,” uses the verbal complement zhu ԣ ‘hold in place’, and
has examples like na zhu ᣓԣ ‘hold fast (with hand)’, yao zhu ઀ԣ ‘clench (with teeth)’,
or kongzhi zhu ᥻ࠊԣ ‘bring under control’. In Chinese, this metaphor of “control is
clenching” is far more common than “control is up.”
Metaphor 161
toward the speaker. This combination of notions—“up” and “toward the

speaker”—implies that the loose accumulation of words or memories one
is trying to draw on lies somewhere beneath one. In xia hai ϟ⍋ ‘jump into
the sea’, which by extension, in the late twentieth century, came to mean
“leave the state employment system and jump into the free-market econ-
omy,” the notion of down applies in the literal sense that the sea is downhill
from the land. But it also applies in the sense that the free-market economy,
down “below,” is murky and beyond easy control.
Still other uses of shang that suggest determinacy (or in this case, for-
mality) and of xia that suggest the opposite (flexibility or informality) are
their uses as transitive verbs. Shang ban Ϟ⧁, literally “go up to the group,”
means “go to work” or “be on duty”— that is, be constricted to a certain
role—while xia ban ‘go off duty’ releases one to go do anything else.
Shang ke Ϟ䂆 ‘go to class, hold class’ and xia ke ‘get out of class’ have simi-
lar implications, as do shang (or xia) tai Ϟ(ϟ)ৄ ‘go onstage (or offstage,
literally or figuratively)’. Literal meanings of shang and xia tend to trump
figurative ones; a miner who joins his group to head down into the mines,
for example, is more likely to xia kuang ϟ⻺ ‘go down into the mine’ than
to shang ban ‘go to work’. But underlying conceptual metaphors survive
such temporary trumpings.
Can we make sense of the fact that major conceptual metaphors us-
ing up and down seem to contradict each other? Why is “fi xity” some-
times down and other times up? To revisit an earlier example, why do we
memorize something down (beixialai 㚠ϟᴹ) but remember it up ( jiqilai
䆄䍋ᴹ)? It might seem, for an example like this, that we can unify the
concepts by saying that we store something down in the bank of memory
(beixialai) until we need to draw on it, when we recall it back up ( jiqilai).
But the little word lai ‘come’, which appears in both phrases, is a major
problem for this interpretation. Lai means “move toward the speaker.”
Therefore, beixialai, literally “memorize down toward us,” suggests that
the galaxy of unmemorized language is hanging over our heads before we
begin to memorize. Then, as we memorize, the memorized phrases come
“down” to our level. The puzzle is that jiqilai also uses lai, so here again
the motion is conceived as moving toward the speaker. But if the phrases
that are being recalled rise toward the speaker, as qi suggests, then they
had to have started out beneath the speaker. This contradicts the implicit
claim of beixialai, which tells us that memorized phrases come down from
above.
One might try to rescue the jiqilai example by arguing that lai does not
have to imply “toward the speaker” when used with qi. In other colloquial
uses in standard Chinese, it sometimes does not. For example if you are
lying in bed next to your partner on Saturday morning, arguing about
who is going to get up to feed the dog, you say ni qilai, wo buxiang qilai
Դ䍋՚, ៥ϡᛇ䍋՚ ‘you get up, I don’t feel like it’. You use lai, not qu, even
though the motion of “getting up” is clearly “away from the speaker.” This
“illogicality” is conventional in modern Mandarin, where qiqu 䍋এ ‘rise
go [away from the speaker]’ is extremely rare (although it is more common
in other regional Chinese languages). But even if jiqilai might be unified
with beixialai, the many other examples of clash between “fixity is up” and
“fi xity is down” would still remain. There is no possible way, I believe, to
unify luoshi ‘descend to facts’ and luokong ‘descend to emptiness’. We just
need to accept that both conceptual metaphors are available and that
speakers and listeners draw on them as needed, and without confusion.
This kind of drawing on what you need in order to make your point applies
to proverbs in a language as well as metaphors. In English we say both
“look before you leap” and “he who hesitates is lost”; the two are exactly
contradictory, yet each is regarded as wise. Contradiction of this kind is
not a barrier to efficient communication.
North and South
Yet another intriguing metaphorical use of up and down involves the four
directions. Human cultures everywhere, from ancient times, seem to
have accepted the “fourness” of the basic directions. This may be, as some
have speculated, because of the arc of the sun across the sky. In any case,
the fourness is a human convention, not a fact of nature.
Literally speaking, of course, none of north, south, east, or west is up. In
the real world, each of these directions is perpendicular to up.93 When we
93. Or very nearly so— a caveat we need to add because of the slight curvature of the
earth.
Metaphor 163
take north to be up, as nearly all modern maps do, we are observing a meta-
phorical convention—and, strictly speaking, one that applies only when a
map is vertical, as when it is on a wall. When a map lies horizontal on a table,
two metaphors need to piggyback: “north is up” combines with “up is the
direction away from the viewer” to yield “north is the direction away from
the viewer.”
On a stone-engraved map of Mespotamia that dates from around 2300
bce, in the dynasty of Akkad, the mapmaker engraved “east” at the top
edge of the map. (We know which edge was the “top” because of the ori-
entation of script and symbols on the map.) “West” was engraved at the
bottom, and “north” at the left.94 Since the map in this case is a block of
stone, it is probably anachronistic to imagine that users placed it verti-
cally or thought of it in vertical terms. We need to remember that “top” is
a modern metaphor for “the side away from the viewer,” just as “bottom”
is “the side toward the viewer.”
The earliest maps we have from China are a group of seven dating from
around 300 bce that were unearthed from a tomb in Tianshui ໽∈ city in
Gansu Province in 1986.95 Drawn in black ink on wood, these maps pri-
marily show rivers and their tributaries, although mountains, roads, and
timber sites are shown as well. It is impossible to say that any particular
sides of these maps were considered the correct sides from which to view
them. With few exceptions, the characters that label the streams and riv-
ers follow, from top to bottom, the direction of the water flow. (The di-
rection of flow can be inferred from the tributary structure.) As rivers do
not all flow in the same direction, the characters on the maps do not ei-
ther and are thus of no use in defi ning which side should be “up” for any
of the maps. One map, which scholars have agreed to call “number 2,” has
94. Alan R. Millard, “Cartography in the Ancient Near East,” in J. B. Harley and
David Woodward, eds., The History of Cartography (Chicago: The University of Chi-
cago Press, 1992) vol. 1, bk. 1, pp. 113. Millard does not state how he knows which side
of the stone-engraved map (pictured on p. 114) is up, but the orientations of the script
and map conventions (overlapping semicircles for ranges of hills) seem to make this
clear. I am grateful to Wang Haicheng for assistance with this and other ancient maps.
95. See Cao Wanru ᳍ဝབ, “Youguan Tianshui fangmatan Qinmu chutu ditu de
jige wenti” ᳝݇໽∈ᬒ偀Ⓗ⾺๧ߎೳഄ೒ⱘ޴Ͼ䯂乬, Wenwu (Beijing) 403 (December
1989), pp. 78– 85; also Mei-ling Hsu, “The Qin Maps: A Clue to Later Chinese Carto-
graphic Development” Imago Mundi 45 (1993), pp. 90–100.
the odd feature of including the character shang Ϟ at one of its edges, and
some have argued that this character indicates “top of map.” But there are
two major problems with such a claim.
One problem is that “top” is a metaphor that makes sense only when the
habit of displaying maps vertically has been established. Was there such a
habit in 300 bce? On a more sophisticated map from about 130 years later,
the more understandable labels dong ᵅ ‘east’ and nan फ ‘south’ are written
at two of the map’s edges. There is no sign of shang ‘up’, even though this
later map, which is painted on silk, not wood, would have been more ame-
nable to vertical display.
The other major problem in interpreting shang Ϟ on map number 2 as
indicating “top of map” is that the character points inward, toward the
center of the map. If one turns the map (or oneself) so that the shang is at
the map’s “top” (away from viewer), the character is upside-down. Why
would a mapmaker do that? Mei-ling Hsu sensibly recommends that we
turn the map back around so that shang is right-side up, but this puts the
shang, necessarily, at the “bottom” of the map. Hsu then makes the some-
what far-fetched claim that the character still indicates “top of map” because,
while resting at the bottom, it “points to the top.”96 Hsu further assumes
that the shang character is pointing “north,” as a compass needle does, even
though 300 bce was several centuries before the appearance of the compass
in China and even though, when the compass did appear, it was considered
by convention to be “pointing south” (zhi nan ᣛफ), not north.
My own guess is that shang is on the map for some reason other than to
tell us which side of the map is “up.” I further suspect that the mapmaker(s)
may have felt no need for an “up.” The ground was lying there, and so was
their sketch; who needs an “up”? In any case, if we cannot know which way,
if any, was considered up, there is no way to determine which of the four
directions up might have been.
The “more sophisticated” map I just referred to is one of three dating
from the second century bce that were excavated from a Han tomb in
Changsha, Hunan Province, in the early 1970s.97 One of these three is a
96. Mei-ling Hsu, “Qin Maps,” p. 92.

97. See Mei-ling Hsu, “The Han Maps and Early Chinese Cartography,” Annals of
the Association of American Geographers 68, no. 1 (March 1978), pp. 45– 60, and Kuei-
Metaphor 165
city map that was found in such bad condition that scholars have not been
able to make much of it. The other two were in good condition and were
restored well enough that useful copies of them could be made.98 They
are drawn on silk and include both symbols and written characters that
indicate features both natural (streams and mountains) and human-made
(roads and settlements). One is called a zhujuntu 侤䒡೪ ‘military map’
and the other a dixingtu ഄᔶ೪ ‘topographic map’.
Again, it is not easy to determine which side, if any, the mapmaker(s)
regarded as “up,” but for these two maps there is more to work with than
there was with the Qin maps. On the military map, an especially large dong
ᵅ ‘east’ is written at one edge and another large nan फ ‘south’ at another.
(The other two edges indicate nothing about direction.) Dong and nan are
written with the “tops” of the characters pointing off the edge of the map,
the opposite of the way the shang is written on Qin map number two.
Therefore, viewing dong and nan in their “upright” positions, it becomes
very plausible that either east or south is the “top” (or, without the vertical
metaphor, the “away from viewer side”) of the map. The question then
becomes whether there is a way to decide between east and south. On the
map itself, there are about seventy-five written labels for various things,
and if the characters for these labels all observed a certain orientation,
that would be strong evidence for which way is “up.” But the labels are writ-
ten in various directions, including diagonally. (We can infer “diagonal” if
we take dong and nan as defining vertical and horizontal). My best count of
these labels finds twenty-seven pointing south, twenty-two east, twenty-
two west, and thirteen north. (I say “best count” because one needs to be a
bit arbitrary in deciding which way a slanted column basically is pointing).
This distribution is good evidence that south was preferred to north as the
“up” direction, but only slim evidence for saying that south was preferred to
east (which is the question we were trying to decide).
To further complicate the matter, the labels that point in a given direc-
tion (east, south, whatever) are not clustered in one part of the map; they
sheng Chang, “The Han Maps: New Light on Cartography in Classical China,” Imago
Mundi 31 (1979), pp. 9–17.
98. Mawangdui Hanmu boshu zhengli xiaozu, Guditu (Beijing: Wenwu chubanshe,
1977).
are scattered everywhere on it, mixed among one another. This fact sug-
gests that the mapmakers did not expect viewers to turn the map (or move
themselves around it) in order to study one part of it from one side, then
another part from another, and so on. It seems, rather, that they expected
all of the map to be useful when viewed from any side, even though view-
ers looking southward (in the terms of the map) had an advantage over
northward-lookers in that they needed to read fewer characters upside
down. Whether this small bias in favor of south was conscious design,
unconscious cultural habit, or just accident is an interesting question but
impossible to answer.
However, the maker(s) of the other Han map, the “topographical” one,
did orient the characters in their labels in a consistent direction, thus
making it clear which side they expected people to view it from— or, in
our modern metaphorical terms, which side was “up.” Unfortunately, they
did not write dong, nan, or anything else on the map to indicate which
direction that was. This problem is easy to solve, however, because the
congruity of the ancient map with modern maps of the same area is very
good. The ancient map shows a body of water at its “top” (inferable from
the orientation of the characters on it), and modern maps make it clear
that that water is the Xi Jiang 㽓∳ estuary that leads to the South China
Sea off the coast of Guangdong Province. Hence, for this map, we may be
certain that “south is up.” Whereas the military map “faces south” only
ambiguously and very slightly, on the topographical map the privileging
of south is clear.
In later times, as is well known, Chinese culture has favored south in a
variety of ways. Residences, including those of emperors, faced south.
The compass has been called a zhinanzhen ᣛफ䞱 ‘south-pointing needle’,
as Chinese culture has preferred to see that needle (which after all “points”
both north and south) as pointing south. A reference as early as the first
century ce mentions a sinan ৌफ ‘take charge of south’, which apparently
was a spoon-like object that a lodestone could pull as a way to indicate
south.99 A zhinanche ᣛफ䒞 ‘south-pointing chariot’, said to be invented
99. See Li Shu-hua, “Origine de la Boussole 11: Aimant et Boussole,” Isis 45, no. 2
( July 1954), pp. 175–196.
Metaphor 167
by the legendary Yellow Emperor, bore a wooden rider whose hands were
supposed to point south regardless of which way the chariot turned.
Can there be any reason for favoring either north or south—any “physical
basis,” as contemporary metaphor theorists might say? The Han-era dic-
tionary Shuowen jiezi 䁾᭛㾷ᄫ defines yin 䱄 ‘dark, female, negative’ as an
ye shui zhi nan shan zhi bei ye ᱫг∈ПफቅП࣫г ‘dark; south of water, north
of mountains’.100 Later texts echo this formula, and in modern Chinese
the saying shannan shuibei wei yang, shanbei shuinan wei yin ቅफ∈࣫⚎䱑,
ቅ࣫∈फ⚎䱄 ‘south of mountains and north of water is yang, north of
mountains and south of water is yin’ continues it.101 It makes sense that
the south sides of mountains should be yang (bright, sunny, and warm)
and the north sides yin (dark, shaded, and cool) because of the angle from
which the sun shines on the earth in the Northern Hemisphere. But why
would it be the opposite for water? Why, for water, is north especially
yang and south especially yin? This is not a minuscule academic question.
It has been important enough to affect the names of some very large
cities—Luoyang ⋯䱑 is on the north side of the Luo River, for example,
and Jiangyin ∳䱄 is on the south side of the Yangtze (Changjiang). A clue
to this puzzle appears in Cheng Dachang’s Song-period (960–1279 ce) Yu
Gong shanchuan dili tu ⾍䉶ቅᎱഄ⧚೪ (Geography of the mountains and
rivers in Yu Gong): gu yi shannan wei yang shuibei wei yang yi shouyang zhi
fang ming zhi ye স㕽ቅफ⚎䱑∈࣫⚎䱑ҹফ䱑ПᮍੑПг ‘the ancient no-
tion that both south-of-the-mountain and north-of-the-water are yang
comes from their both being places that receive sunlight’.102 Here I have
translated the second yang in the quotation as “sunlight” in order to em-
phasize what clearly appears to be the key natural fact that underlies the
100. Shuowen dazidian䁾᭛໻ᄫ‫(݌‬Shuowen jiezi 䁾᭛㾷ᄫ) (Taipei: Xuehai chuban-

she ᅌ⍋ߎ⠜⼒, 1982), 7:91.
101. I am indebted to Ye Minlei for reporting how she was asked to memorize this
line in elementary school in the 1990s. Email messages to the author, March 5 and 7,
2008.
102. Cheng Dachang ⿟໻ᯠ (1123–1195), Yu gong shanchuan dilitu ⾍䉶ቅᎱഄ⧚೪
[Geography of the mountains and rivers in Yu Gong] (Shanghai: Shangwu yinshu-
guan, 1936), vol. 1 ोϞ, p. 50. Also in Yingyin wenyuange sikuquanshu: jingbu ᕅॄ᭛⏉
䭷ಯᑿܼ᳌•㍧䚼 [ Jing section of the photocopied Wenyuange Complete Library in
Four Divisions] (Taipei: Taiwan Commercial Press, 1983), 56: 129.
usage: river banks on north sides of rivers get more direct sunlight than
do banks on south sides. (The opposite would be true in the Southern
Hemisphere.) The puzzle (if there still is one) is why early European map-
makers, who were located in the Northern Hemisphere, as the Chinese
were, chose to privilege the colder and darker north direction.
Several hundred years after the invention of the compass in China,
“pointing south” came to mean not just showing direction but showing
correct direction. In the poetry of Zhang Heng ᔉ㸵 (78–139 ce), Li Shang-
yin ᴢଚ䲅 (813–58 ce), and elsewhere, zhinan shifts from physical mat-
ters to the ethical domain and comes to mean “pointing the way to ap-
propriate behavior.”103 Eventually, the guidance that zhinan offered came
to refer to a wide spectrum of life. In 1936 when the Zhonghua Book
Company published a comprehensive guide to Shanghai, a city that people
from elsewhere in China could find mind-boggling, editors called it Da
Shanghai Zhinan ໻Ϟ⍋ᣛफ ‘Guide to great Shanghai’— or, more literally,
“pointing south in big Shanghai.”
Given this cultural favoring of south, it is interesting, and somewhat
puzzling, that maps in China’s Northern Song period (960–1127 ce) be-
gan to use a convention of “north is up.” Several stone-carved maps of the
whole of China, one from 1121 ce and others from 1136 ce and later,
clearly show the Yellow River, the Yangzi River, the Shandong peninsula,
and other well-known features of China’s topography. Two of these, the
jiuyu shouling tu бඳᅜҸ೪ ‘map of nine governing districts’ and the hua
yi tu 㧃་೪ ‘map of the Chinese and barbarians’, inscribe bei ࣫ at their
“top” (away from viewer) edges, nan फ at the bottom edges, xi 㽓 at the
left, and dong ᵅ at the right. Both maps, moreover, contain hundreds of
characters all of whose “top” sides consistently point north.104 In sum the
“north is up” orientation is strong and unambiguous. Because the elapsed
time between these maps and their Han predecessors is nearly thirteen
103. Zhang Heng, “Dongjing fu” ᵅҀ䊺, contains the line xing jianzhinan yu wu zi
ᑌ㽟ᣛफѢ਒ᄤ ‘fortunate in receiving guidance from you’, and Li Shangyin, “Xianji
xianxianggong qi” ⥏䲚䊶Ⳍ݀ଧ, includes the line wei baidai zhi zhinan, jiuzhou zhi
muduo ⚎ⱒҷПᣛफ, бᎲП᳼䨌 ‘as the [moral] compass for a hundred generations,
and [moral] teacher for China’. Cihai bianji weiyuan hui, Cihai, p. 1586.
104. See Cao Wanru ᳍ဝབ et al., eds., Zhongguo gudai dituji Ё೑সҷഄ೒䲚 (Bei-
jing: Wenwu chubanshe, 1990), vol. 1 (zhanguo-yuan), plates 62, 63, and 65.
Metaphor 169
centuries, we cannot speak of a “sudden” change, but it is a 180-degree

change and raises the interesting question of what brought it about. Was
there some influence from the Indo-European world, either over the
Silk Route or over the oceans? The question is made no easier by the
paucity of maps from the intervening years. One does exist, though not
from the heart of China. A circular “world map” drawn in Kashgar (in
contemporary Xinjiang) probably between 1072 and 1076 ce by a Turkic
scholar named Mahmud al-Kjshgarq shows Arabic script that implies an
orientation in which east is the “up” (away from viewer) direction, west
is down, north is left, and south is right.105 What to make of this is hard
to say.
Consciousness
Scholars who have studied metaphor in European languages have ob-

served that “up” and “down” relate to consciousness. Lakoff and Johnson
note that in English “conscious is up, unconscious down.”106 We “wake
up,” but “fall asleep,” “drop off” (into sleep), or “sink” into a coma. We
have a “subconscious” and can “go under” hypnosis. Lakoff and Johnson
claim that the “physical basis” for this conceptual metaphor is the fact
that human beings sleep lying down and stand up (or sit up) when con-
scious. We wake “up” because we get up. One problem with this physical-
basis theory is that Chinese people also sleep lying down but do not use
the “conscious is up, unconscious is down” metaphor in their language— at
least not before it was influenced by Western language. Modern Chinese
does use xia yishi ϟᛣ䆚, literally “below consciousness” for the subcon-
scious, but this term is clearly a borrowing, established in order to trans-
late Sigmund Freud, Carl Jung, and others. When we fall asleep in Chi-
nese we normally shui zhao ⴵⴔ, where shui is “sleep” and zhao is a verbal
complement indicating that the verb “takes effect” or, as we might say in
105. Zhang Guangda ᓴᑓ䖒, “Shiyi shiji de yuanxing ditu” कϔϪ㑾ⱘ೚ᔶഄ೒

[A round map from the eleventh century], in Cao Wanru et al., Zhongguo gudai dituji,
pp. 19–22; see also plate 42.
colloquial English, “kicks in.” In zhao there is no sense of up or down.

Spatial metaphors for going into or out of consciousness do exist in Chi-
nese, but they do not have us going up or down; they have us “crossing
over” an imaginary line within a horizontal plane. When xing 䝦 ‘wake
up’ uses a spatial metaphor, it is usually xing lai ‘wake come’ or xing guo lai
䝦䖛ᴹ ‘wake across come’. We the wakeful are on this side of a line, and
those in slumber, stupor, or coma are on the other. To faint is to yunguoqu
ᰩ䖛এ ‘swoon across go’.
It would be too simple to say that English always uses concepts of up
and down for entering and exiting consciousness while Chinese uses only
crossing back and forth, because there are exceptions, or quasi exceptions,
on both sides. For “wake up” in English we can say “come to,” which is not
far from the Chinese xing lai. And in the previous section we saw how one
of the Chinese metaphors for memorize, bei xialai, literally means “mem-
orize down [toward the speaker],” while not being able to recall from
memory can be shuo bu shanglai, literally “saying [it] cannot come up.”
With memory, of course, we are dealing with something less radical than
completely entering or exiting consciousness, but something similar is
involved; we are not conscious of things in our memories until we “recall”
certain items from memory’s storehouse, and this process, in Chinese, is
sometimes conceived as going up or down.
Despite these exceptions, it seems clear that Eu ropean languages and
Chinese have two different tendencies in conceiving the “direction” of en-
tering and exiting consciousness and that these tendencies, in the terms of
modern cognitive science, are “conceptual metaphors” that affect how
people view the world. If this is so, it becomes interesting to ask whether
the two conceptual tendencies might lead people in the two cultural
traditions—including their philosophers—to ask somewhat different ques-
tions. Consider, for example, the famous butterfly dream from chapter 2 of
the Zhuangzi (third–fourth century bce). In Burton Watson’s translation:
Once Chuang Chou [i.e., Zhuangzi] dreamt he was a butterfly, a but-

terfly fl itting and fluttering around, happy with himself and doing as
he pleased. He didn’t know he was Chuang Chou. Suddenly he woke
up and there he was, solid and unmistakable Chuang Chou. But he
Metaphor 171
didn’t know if he was Chuang Chou who had dreamt he was a but-
terfly, or a butterfly dreaming he was Chuang Chou.107
Watson translates jue 㾎 as “woke up,” which is certainly a defensible

translation, yet “up” is not part of what Zhuangzi thought had happened
to him.108 In Zhuangzi’s Chinese, the slide back and forth from dreaming
to waking seems somehow easier than up and down suggest in modern En-
glish. This difference may arise because (as we saw in the previous section)
up and down connote so much else in modern English—more or less of
some thing, higher or lower status or moral standing, and so on. Zhuang-
zi’s puzzle feels more elegantly simple— and therefore more puzzling—
in Zhuangzi’s ancient Chinese. A speaker (and dreamer) of modern En-
glish might also have come up with this puzzle, but perhaps not as readily.
One wonders how much the wide differences between the Chinese and
European philosophical traditions lie less in their different answers to
questions than in the different questions that they are induced to ask and,
moreover, how much the matter of “what questions are asked” is rooted in
differences of language systems. I will revisit this question later when the
question of nouns and “containers” arises.
The Self in Ancient Thought
The contemporary field of cognitive science has done considerable work

with the notion that the “self” is often metaphorically conceived as two
different things at the same time.109 In a sentence like “I could not control
myself,” one self, which is the seat of consciousness, acts (or fails to act) on
another; the second self is “out there” somewhere, being acted on, even
though, in a larger sense, it is still the same person. “I dragged myself out
of bed” is another example, as is “you’re pushing yourself too hard,” or “she
107. Burton Watson, trans., The Complete Works of Chuang Tzu (New York: Colum-
bia University Press, 1968), p. 49.
108. In his translation of Zhuangzi as Wandering on the Way (New York: Bantam
Books, 1994), Victor Mair avoids this problem by translating jue as “awoke” (p. 24).
109. See Lakoff and Johnson, Philosophy in the Flesh, ch. 13.
plopped herself down on the couch.” In such cases, theorists have called
the seat-of-subjectivity self a Subject and the self that is acted upon a Self.
(The capital letters are conventional here.) Another version of the “two
selves” metaphor occurs in examples like “I was not myself today,” where
one self is more exterior and ephemeral and another is deeper and more
essential. In these cases, some theorists set aside the term “Subject” and
refer to Self 1 and Self 2.
All such phrases are considered metaphorical because what Subject
does to Self (or the relation between Self 1 and Self 2) is not literal. I do
not literally drag myself out of bed and I am not literally something other
than myself. What the metaphors say can often be rephrased in nonmeta-
phorical language. “I was not myself today” might be said nonmetaphor-
ically as “today I did not behave as I usually do.” “You’re pushing yourself
too hard” might be rephrased nonmetaphorically as “you are trying too
hard and might suffer as a result.” A sentence like “I lifted my left arm” is
nonmetaphorical if you reach over with your right arm to do the lifting
but metaphorical if you use the muscles of your left arm itself.110
Edward Slingerland, in his book Effortless Action: Wu-wei as Conceptual
Metaphor and Spiritual Ideal in Early China,111 uses distinctions between
“Subject” and “Self,” and between “Self 1” and “Self 2,” to try to shed light
on what he calls the “paradox of wu-wei” among pre-Qin philosophers. Wu-
wei ⛵⠆, literally “[taking] no [intentional] action” is an ideal human state
for Laozi, Zhuangzi, Confucius, Mencius, and Xunzi. The metaphors these
thinkers employed suggest a variety of methods for achieving wuwei, and
each method, one way or another, involves a version of the paradox of “try-
ing not to try.” Some of these thinkers—Zhuangzi, Laozi, and Mencius—
emphasize not trying, according to Slingerland. Their view, in his para-
phrase, is that “we already are good, and we need merely to allow this
virtuous potential to realize itself.”112 For this group of thinkers the difficult
question becomes: if we are already there, why does anything at all have to
be attempted? Why do we need philosophers to point “a way” for us? Other
110. See ibid., p. 271.

111. Edward Slingerland, Effortless Action: Wu- wei as Conceptual Metaphor and Spiri-
tual Ideal in Early China (Oxford: Oxford University Press, 2003).
112. Ibid., p. 12.
Metaphor 173
thinkers, among whom Confucius and Xunzi are Slingerland’s examples,

hold that “wu-wei is a state acquired only after a long and intensive regime
of training”—that is, we do have to try. For this group the puzzle becomes:
how does one apply effort toward reaching a state of applying no effort?
Because “wu-wei is understood as an achieved state,” each thinker must
“specify some sort of effort-ful program for attaining this state.”113
Slingerland names these two approaches to wuwei “internalist” and
“externalist.” He finds that neither approach can solve the paradox but
“merely chooses a horn of the dilemma on which to impale itself.”114 He
then applies contemporary metaphor theory to the issue and asks, about
the paradox, “Does metaphor theory solve anything?” He wisely demurs.
It does not solve the paradox, he writes, but it does clarify certain issues by
“demonstrating more concretely” points that before had been “merely
intuitive connections.”115 For example, when we try to understand what
pre-Qin thinkers meant by metaphors involving cong ᕲ, yin ಴, yi ձ, and
shun ䷚— all of which have to do with “following” nature—the distinction
between Subject and Self can help, because we can see the Subject as fol-
lowing the Self. When I follow my nature, “the Subject,” Slingerland
writes, “is able to be free of exertion [i.e., wuwei] because the Self is al-
lowed to do all the work”; therefore, wuwei can be “a state in which action
is occurring even though the Subject is no longer exerting force.”116 In
another kind of metaphor, which appears especially in the Zhuangzi, the
Subject seems to “lose” or “forget” the Self. This can be a bad thing, as
when a second-rate shamen quails before a true Daoist, “loses himself” (zi-
shi 㞾༅), and runs away.117 But “losing oneself” can also be related to a state
of wuwei, as when one “forgets [wang ᖬ] liver and spleen, loses [yi 䙎] ears
and eyes, and un-self-consciously [mangran 㣿✊] roams outside the dusty
realm, wandering easily [xiaoyao 䗡䘭] in the ser vice of wuwei.”118 Similarly,
113. Ibid., p. 38.

114. Ibid., pp. 12 and 265.
115. Ibid., p. 270.
116. Ibid., p. 28.
117. Ibid., p. 31; Watson, Complete Works of Chuang Tzu, p. 96.
118. This translation is from Slingerland, Effortless Action, p. 32, except that I have
excised “his.” Slingerland notes that his rendition is based on others by Burton Wat-
son and A. C. Graham.
the distinction of a relatively superficial Self 1 and a more essential Self 2

can be useful in interpreting ziran 㞾✊, or the “so-of-itself,” that one
fi nds in the Laozi, where ziran is “the way a thing is when it follows its
own internal essence.”119
These attempts to apply modern metaphor theory to the concept of
wuwei have to do with what wuwei is, not how to achieve it, and therefore
can be useful in interpreting the thought of both the “internalists” and the
“externalists,” as Slingerland refers to them. For the externalists, however,
metaphor theory has the added advantage of making sense of method, that
is, how to “try not to try.” The separation of Subject and Self makes it
easier to conceive the Subject, here and now, as “trying”—through learn-
ing, discipline, or whatever—to achieve wuwei for a Self, projected into the
future, who then will not need, want, or use any effort to achieve it.
Can we say that any intellectual advance is involved? One might argue
that clarity itself is a contribution, but is there anything else? Probably not,
as far as I can tell— at least not for the serious study of Chinese thought.
But on another score there is, in my view, a significant intellectual ad-
vance. This is simply the discovery that metaphor theory based primarily
on modern English and other modern Western languages can make good
sense when applied to pre-Qin Chinese texts. If the distinction between
Subject and Self, or between ephemeral Self 1 and essential Self 2, applies in
language and life as distantly separated as pre-Qin China and the contem-
porary United States— and this much, at least, Slingerland has shown—
then there are good grounds for hypothesizing that such distinctions are
fairly deep in human experience. There does appear to be a sense in which
we human beings, when we look inwardly on ourselves, can view ourselves
as bifurcated.
Privilege in Dyads
In many languages, one of the items in paired categories is standardly

privileged over the other. The pattern is sufficiently pervasive that some
theorists have hypothesized that it, too, may be “universal.”
119. Slingerland, Effortless Action, p. 35.

Metaphor 175
These paired categories, sometimes called “dyads,” are things like up

and down, front and back, good and evil, more and less, before and after,
and love and hate. There are many, many of them. In many cases, they are
not much different from what grade school teachers call “opposites.”120
Linguists have noticed a pattern in the ways languages list dyads: a “plus”
item often comes first and a “minus” item second.121 Here “plus” and “mi-
nus” are conceived broadly, to encompass such distinctions as good-bad,
primary-secondary, prior-subsequent, cause-effect, and others.
Studies of dyads in the 1970s were based on European-language exam-
ples, and the results were sometimes presented, with due caution, as only
“our culture’s view.”122 Are dyads different in different cultures? Between
Chinese and English, at any rate, they are strikingly similar. In Chinese
we say hao huai དണ ‘good and bad’, not huai hao; and shang xia Ϟϟ ‘up
and down’, not xia shang. These examples match English, and so do the
overwhelming majority of others: shi fei ᰃ䴲 ‘right and wrong’, shi fou ᰃ৺
‘is and isn’t’, dui cuo ᇍ䫭 ‘correct and incorrect’, qian hou ࠡৢ ‘front and
back’, da xiao ໻ᇣ ‘big and small’, lao shao 㗕ᇥ ‘old and young’, gao di 催Ԣ
‘high and low’, chang duan 䭓ⷁ ‘long and short’, kuan zhai ᆑじ ‘wide and
narrow’, ai hen ⠅ᘼ ‘love and hate’, and so on.123 In English, “black and
white” may seem an exception to the pattern—and so it seems in Chinese
as well. The five main colors are conventionally listed in Chinese as hong
huang lan bai hei 㑶咘㪱ⱑ咥 ‘red, yellow, blue, white, black’, with bai com-
ing before hei, but in most other expressions, such as heibai dianshi 咥ⱑ⬉㾚
‘black-and-white television’ or diandao hei bai 乴‫צ‬咥ⱑ ‘invert black and
white—turn the truth upside-down’, hei comes first. But exceptions like
“black and white” only underscore the strength of the general pattern.
120. Not in all cases. Now and then, and here and there, are dyads, but not exactly
“opposites.”
121. On this point and in much of the following discussion, I owe much to David
Moser, “Covert Sexism in Mandarin Chinese,” Sino-Platonic Papers 74 (January 1997),
pp. 1–23. No one writes more clearly or incisively than Moser on the topic of struc-
tural metaphor in Chinese.
122. William E. Cooper and John Robert Ross, “World Order,” in Robin E. Gross-
man, L. James San, and Timothy J. Vance, eds., Functionalism (Chicago: Chicago
Linguistic Society, 1975). The phrase “our culture’s view” is from Lakoff and John-
son, Metaphors We Live By, p. 132.
123. This list of examples is from Moser, “Covert Sexism,” pp. 2–3.
In both Chinese and English, the “privilege” of the first members in

dyads is reflected not only in the fact that they come fi rst but in their be-
ing the standard term for measuring attributes of their kind. For example,
if we want to know Sam’s height, we say “How tall is Sam?” not “How
short is Sam?” We can, of course, ask “How short is Sam?” but the ques-
tion would suggest a special purpose. If Sam, although thirty- one years
old, is not allowed on a roller coaster because of a height requirement, we
might ask “How short is Sam?” But if we just want to ask how tall he is in
the normal way, we don’t use “short.” Similarly, we ask “How big is Rhode
Island?” (even though it is small, as states go) and can ask “How long is a
microbe?” It is entirely natural to ask “How wide is the Mississippi at its
narrowest?” but would seem odd to ask “How narrow is the Mississippi at
its widest?” These examples show that the fi rst member of a dyad is a “de-
fault” category. Chinese works the same as English in this regard. We ask
in Chinese Zhang San you duo gao? ᓴϝ᳝໮催? ‘How tall is Zhang San?’
(not— except in special circumstances—Zhang San you duo ai? ᳝໮ⷂ?
‘How short is Zhang San?’). The principle also holds in the “choice type”
questions that are so common in Chinese. “Was what she said right?” is ta
shuo de dui bu dui? ཌྷ䇈ᕫᇍϡᇍ? not ta shuo de cuo bu cuo? ཌྷ䇈ᕫ䫭ϡ䫭.
Similarly, Wenti da bu da? 䯂乬໻ϡ໻? ‘Is the problem big?’ (and very sel-
dom wenti xiao bu xiao 䯂乬ᇣϡᇣ ‘Is the problem small?’); Haizhe hao bu
hao chi? ⍋㳛དϡདৗ? ‘Does the jellyfish taste good?’ (never haizhe huai
bu huai chi? ⍋㳛ണϡണৗ? ‘Does the jellyfish taste huai—bad?’).
In a very small number of cases, Chinese and English differ in their
privileging. North and south is an example. In English we talk about the
North and South Poles and in Chinese the nanbeiji फ࣫ᵕ ‘South and
North Poles’; talks between the developed and developing countries are
the North-South dialogue in English, but the nanbei duihua फ࣫ᇍ䆱
‘south-north dialogue’ in Chinese.
It is also worth noting that in both Chinese and English, the conven-
tional order of a dyad can be reversed if one has a conscious reason for
reversing it. In contemporary English, when people want to compensate
for the bias inherent in saying “he and she,” they sometimes say “she and
he.” Among the talking points on “the Korea question” that Hu Jintao,
the president of China, brought with him on his tour of North America in
2005 was beinan Chaoxian zhi zheng shi xiongdi zhi zheng ࣫फᳱ剰Пѝᰃ‫ܘ‬
Metaphor 177
ᓳПѝ ‘the struggle between North and South Korea is a struggle be-
tween brothers’.124 Here the reversal of the conventional nanbei to an un-
conventional beinan apparently was done to grant primacy to North, not
South, Korea. The switch was made all the more important because of the
parallel phrase xiongdi ‘older brother and younger brother’. Unless nanbei
were reversed, South Korea would be parallel with “older brother,” which
Hu apparently did not want.
The striking congruities between Chinese and English in the way dy-
ads are conceived raises the question why this might be so. Borrowing is
out of the question; the independent roots of usages in the two languages
are far too obvious for that. But might there be some kind of commonality
in human experience that provides an “experiential basis” for the expres-
sions in both Chinese and English— and, perhaps, in human languages
generally? Lakoff and Johnson, citing work by William Cooper and John
Ross, have suggested a “me first” basis for the ordering of dyads.125 The
idea is that the first item in a dyad is somehow “closer to me”—the seat of
subjectivity—than the second item. Lakoff and Johnson list “up, front,
active, good, here, and now” as examples of the “me first” position that ex-
plains why we say—in these particular orders—“up and down, front and
back, active and passive, good and bad, here and there, now and then.” This
claim raises some questions. We must assume, first of all, that “me” is con-
ceived as good more than bad; this assumption is probably all right, in view
of general human nature. We also must see “me” as closer to now than to
then, and as closer to here than to there. Since “me” at other times would
be me then (and might also be at another place, hence me there), we are
obliged to understand “me” as meaning “my present subjectivity”— or, to
use Slingerland’s terms, Subject, not Self. This, too, seems fair enough. The
biggest problem among the “me first” examples Lakoff and Johnson pres-
ent is why up should be closer to “me” than down is. They recognize this
problem but do not address it effectively.126
124. Zhang Liang, email message to author, August 29, 2005.

126. They argue that “nearest is fi rst,” which would explain why up would be listed
fi rst once one had decided that it was nearest. But on the question whether up is in fact
“nearest,” they only state the point without arguing it. They write: “Of the two con-
cepts up and down, up is oriented nearest to the prototypical speaker” (ibid., p. 133).
Insofar as the “me first” theory works, it works almost as well in Chi-
nese as in English. If front, good, and active are “me first” in English and
hence listed before back, bad, and passive, similarly in Chinese qian, hao, and
zhudong Џࡼ are conventionally listed before hou, huai, and beidong 㹿ࡼ.
Now and then are a bit problematic because xianzai ⦄೼ and nashi 䙷ᯊ are
not an idiomatic pair in the way now and then are in English. A more idi-
omatic pair in Chinese (at least in northern Mandarin) would be zhehuir
䖭Ӯ‫‘ ܓ‬this moment—now’ and nahuir 䙷Ӯ‫‘ ܓ‬that moment—then’, and
this pair does observe the “me fi rst” pattern. A more significant diver-
gence from the “me first” pattern in Chinese is that zher 䖭‫ ܓ‬or zheli 䖭
䞠 ‘here’ do not always refer to the position of the speaker, as here does in
English. For example, in planning a meeting, if someone says in English
“let’s meet here at your place,” we infer that the speaker is presently at your
place. ( Were he or she somewhere else, the sentence would be “let’s meet
there at your place.”) This convention is mandatory in English but in Chi-
nese is not. One occasionally hears in Chinese sentences like zamen dou dao
ni zheli lai jianmian ba અӀ䛑ࠄԴ䖭䞠ᴹ㾕䴶৻ (literally, “let’s all meet
here at your place”) even if the speaker is not presently at your place.127 In
such cases, the speaker projects the center of things (“here”) to the loca-
tion that is associated with the listener. This pattern is not common in
Chinese, where dao ni nali qu ࠄԴ䙷䞠এ is still much more common in such
circumstances; but the fact that the usage exists at all shows Chinese to
have a conceptual flexibility that English does not allow.
The “me first” explanation of the internal order in dyads is not ideal,
but I do not know a better one. To say that the “plus-minus” congruities
between Chinese and English in this regard are mere chance is a much
weaker theory. Something is going on, and it goes on at a level that speak-
ers of the two languages do not normally notice. David Moser has astutely
127. I have noted this kind of usage in my own experience on several occasions. It
also appears occasionally in literature, for example when we read in Hu Fayun’s 㚵থѥ
novel [email protected] བ⛝ @sars.come [So it [email protected]] (Beijing: Zhongguo
guoji guangbo chubanshe, 2006), from a narrator who is not located at anyone’s house,
that Damo zhiyao huicheng, jiu changchang dao Wei laoshi zher lai 䖒ᨽা㽕ಲජ, ህᐌᐌࠄ
ि㗕Ꮬ䖭‫ܓ‬ᴹ ‘whenever Damo returned to the city he made it a habit to come [here]
to Teacher Wei’s place’ (p. 31).
Metaphor 179
observed that the evaluative connotations of “plus-minus” pairs often go

unnoticed even when they relate to major and controversial social issues.
Moser shows how the privileging of one or another pole can be “covert,”
and illustrates with the male-female dyad.128
In both Chinese and English, male comes first across a wide range of
uses. In English we have man and woman, he and she, brothers and sis-
ters, husband and wife, Mr. and Mrs., sons and daughters, guys and dolls,
and many other examples. We also have paired names like Jack and Jill,
Romeo and Juliet, Hansel and Gretel, Anthony and Cleopatra, Samson
and Delilah, Roy Rogers and Dale Evans, and many others in which the
male’s name comes first.129 Similarly, in Chinese we have nannü ⬋ཇ
‘male and female’, fuqi ໿ྏ or fufu ໿ཛ ‘husband and wife’, fumu ⠊↡
‘father and mother’, baba mama ⠌⠌ཛྷཛྷ ‘Dad and Mom’, xiongdi jiemei
‫ܘ‬ᓳྤྍ ‘brothers and sisters’ and ernü ‫ܓ‬ཇ or zinü ᄤཇ ‘sons and daugh-
ters’. Slightly more abstractly, but still “male first,” there are longfeng 啡勇
‘dragon [male] and phoenix [female], qiankun ђസ ‘male and female—the
cosmos’, and caizi jiaren ᠡᄤՇҎ ‘[implicitly male] talent and [implicitly
female] beauty’. Male-first pairs of names are also common in Chinese:
Liang Shanbo ṕቅԃ and Zhu Yingtai ⼱㣅ৄ; Jia Baoyu 䊜ᇊ⥝ and Lin
Daiyu ᵫ咯⥝.130 Moser notes exceptions to the pattern, such as yin 䱄
‘female element’ and yang 䱑 ‘male element’, which are always yinyang in
Chinese. In English, “ladies and gentlemen” appears to be an exception,
although Moser notes that it is based in a kind of chivalry that makes the
case problematic. “Bonnie and Clyde” is a more interesting exception be-
cause it violates not only the male-first pattern but also the rhythmic
preference of TAH-ta TAH-ta over ta-ta-ta-TAH. (As we saw in chapter
1, “salt and pepper” is generally preferred to “pepper and salt.”) In search-
ing for an explanation for Bonnie and Clyde, Moser wonders if it may lie
in “the sheer novelty of a female bank robber.”131 The English-language
exception of Mom and Dad (or Mom and Pop) is not mirrored in Chinese,
128. David Moser, “Covert Sexism,” pp. 1–2.

129. Ibid., p. 3.
130. Ibid., pp. 4–5.
131. Ibid., p. 3, n. 3.
where bama ⠌ཛྷ and fumu are standard. Moser notes that mufu “sounds
utterly wrong to Chinese ears.”132
Beyond the question of privileging within dyads is a subtler but more
telling fact: categories that in theory are gender-neutral are often—in
both Chinese and English—implicitly understood as male. “Male” is of-
ten a default interpretation of phrases in which gender is formally am-
biguous. Moser cites stories from Chinese joke books to illustrate. One
joke begins you yige ren, laopo si le ᳝ϔ‫ן‬Ҏ㗕ယ⅏њ ‘there was a person
whose wife died’. Here ren Ҏ ‘person’ is theoretically gender-free, but the
hearer understands the term as male, and this is confi rmed when we hear
that “the wife” died. What happens, Moser asks us to observe, if we say
the same sentence but have a husband die? The sentence you yige ren,
zhangfu sile ᳝ϔ‫ן‬ҎϜ໿⅏њ ‘there was a person whose husband died’
generates, according to Moser, “a slight sense of strangeness” for Chinese
speakers.133 Ren conjures (by the default principle) a certain expectation
that the joke will be about a man; then, when we hear that “the husband”
died, something doesn’t fit. The listener needs to go back and recalibrate.
Chinese jokes about females standardly do not begin by calling them ren
but use a gendered term from the beginning. For example: you yiwei xiaojie
zai haitanshang shai taiyang, chuanzhe sandianshi youyongyi . . . ᳝ϔԡᇣྤ೼
⍋☬Ϟ᳀໾䱑, こⴔϝ咲ᓣ␌⋇㸷 ‘there was a young lady sunbathing at
the beach, wearing a bikini . . .’. If we substitute ren in this example,
Moser’s insight is confirmed. The phrase generates a mild sense of con-
tradiction. In hearing you yige ren chuan sandianshi youyongyi ᳝ϔ‫ן‬Ҏこϝ
咲ᓣ␌⋇㸷 ‘there was a person wearing a bikini’, the listener needs to
understand a small but awkward gender shift halfway through.
The male gendering of “neutral” terms is further confi rmed by the fact
that whenever the protagonist of a joke is explicitly tagged as female ( you
yiwei xiaojie . . . ᳝ϔԡᇣྤ ‘there was a young woman’, you yiwei lao taitai . . .
᳝ϔԡ㗕໾໾ ‘there was an old lady’), it raises an expectation that her
gender is going to be relevant to the story. This is not true if the protago-
132. Ibid., p. 4, n. 4. I have a Chinese American friend who wrote to her parents,
who were natives of Taishan ৄቅ in Guangdong, addressing them as mufu ↡⠊. But
this usage may have been influenced by English.
133. Ibid., p. 7.
Metaphor 181
nist is male. Whether tagged explicitly ( you yige laotour . . . ᳝ϔ‫ן‬㗕丁‫ܦ‬

‘there was an old man’) or implicitly ( you yige ren . . . ᳝ϔ‫ן‬Ҏ ‘there was
a person’), there is no parallel implication that his sex is going to matter. It
might or might not matter. Essentially the same is true for English-
language jokes. A joke that starts “this guy goes into a bar . . .” can be
about anything. But if a joke begins “this lady went into a bar . . .” the lis-
tener will wait to hear why it matters that she was female. You can test this
effect by choosing any “this guy goes into a bar” joke and substituting
“woman”, “lady,” or “gal” for “guy” but changing nothing else in the joke.
Then watch your listeners’ reactions. Do they wonder why you supplied
“unnecessary information”? Moser cites a parallel case in which one might
begin a joke by saying, “This 650-pound man walks into a bar . . .” and
then telling a story that has nothing to do with the size of the man. Listen-
ers would wonder why the joke-teller supplied extraneous information.134
Moser’s choice of jokes as source materials for studying implicit atti-
tudes is a probably a good one, because jokes are told in relaxed and rela-
tively unguarded contexts where “default” attitudes can emerge. When
telling jokes, people focus on their punch lines, not their conceptual cat-
egories. But it is important to note that this default phenomenon is not
limited to jokes. It extends to many other language contexts. Liu Binyan’s
work of reportage Yige ren he ta de yingzi ϔϾҎ੠Ҫⱘᕅᄤ (A ren [man]
and his shadow) tells the story of a courageous young man and would not,
I think, have borne that title if the protagonist had been female.135 When
a protagonist is female, people use a gendered term to make her female-
ness explicit. The Chinese language sometimes folds nü ཇ ‘female’ or mu
↡ ‘mother’ into a term that is normally used for men in order to produce a
version of the term that is earmarked, as it were, “female special case.” Ex-
amples are cainü ᠡཇ ‘female talent’, which is the “female exception” ver-
sion of caizi ᠡᄤ ‘talented scholar’, or shengmu ೷↡ ‘sage mother—female
sage’, the female version of the (normally male) shengren ೷Ҏ ‘sage’. More
common than substitution, though, is the simple adding of nü as a prefi x, as
in nü xuesheng ཇᄺ⫳ ‘girl student’, nü daifu ཇ໻໿ ‘woman doctor’, nü
134. Ibid., p. 6, n. 9.

135. Liu Binyan, Liu Binyan Baogaowenxue ࡝䊧䲕ฅਞ᭛ᅌ (Hong Kong: Mingch-
uang chubanshe ᯢにߎ⠜⼒, 1987), pp. 67–136.
qiangren ཇᔎҎ ‘female strongman’, nü zhurengong ཇЏҎ݀ ‘female pro-

tagonist’, nü siji ཇৌᴎ ‘female [car or truck] driver’, nü minbing ཇ⇥݉
‘female militia member’, nü yingxiong ཇ㣅䲘 ‘heroine’, nü yongdongyuan
ཇ䖤ࡼਬ ‘female athlete’, and many other examples. For none of these
terms is nan ⬋ ‘male’ an appropriate prefi x. The term that contains nü is
sometimes shortened for the sake of rhythmic balance (e.g., gongren ᎹҎ
‘worker’ becomes nügong ཇᎹ ‘female worker’, jianyu ⲥ⣅ ‘prison’ becomes
nüjian ཇⲥ ‘women’s prison’, etc.), but the use of nü as a tag that suggests
“special case” is the same. Niang ࿬ ‘woman’ can be used either as a prefi x
(niangzijun ࿬ᄤ‫‘ ݯ‬women’s army’) or a suffi x (laobanniang 㗕ᵓ࿬ ‘boss
lady’); in either event the “special case” coloration of the gendered term
remains the same. A male army is a jundui ‫ݯ‬䯳 ‘army’, not a nanzijun ⬋ᄤ‫ݯ‬
‘men’s army’; and the male boss is a laoban 㗕ᵓ ‘boss’, not (not normally,
anyway) a laobandie 㗕ᵓ⠍ ‘boss daddy’.
The principle that the female case needs “special marking” applies even
to the internal structure of characters. Moser points out that standard
modern Chinese includes about 275 characters that use the nü ཇ ‘female’
radical, whereas a nan ⬋ ‘male’ radical does not even exist. The generic
“human being” (ren Ҏ) radical apparently does all the marking that males
need.136 In the early 1930s the poet Liu Dabai ࡝໻ⱑ invented a character
for “him” by combining ⬋ and г to correspond to the recently invented
ta ཌྷ ‘she, her’, but his effort died for want of followers.137
Yet we must not overstate the extent to which terms like ren ‘human
being’ carry an implicit male connotation. Theoretically, the term is un-
gendered, and in many contexts it actually is. Moser points out that when
someone is using a public telephone and we say you yige ren zheng zai yong
᳝ϔϾҎℷ೼⫼ ‘someone is using it now’ there is no presumption about
the sex of the user.138 When we talk about ren zhi changqing ҎПᐌᚙ
‘common sense in human relations’, it is natural to think of both sexes. If
an elevator holds a maximum of twelve ren, no one checks sexes before
deciding whether to allow a thirteenth to board. There are, moreover, a
small number of ungendered terms for which “female,” not “male,” is the
136. David Moser, “Covert Sexism,” pp. 12–13.

137. Ibid., p. 11.
138. Ibid., p. 5.
Metaphor 183
default interpretation. These are the names of roles in society that usually
are filled by females, such as hushi ᡸ຿ ‘nurse’ or laoshi 㗕Ꮬ ‘teacher’—
provided it is a teacher at nursery or primary school. (At higher levels of
education, laoshi takes on a male “default” interpretation.) English, too,
observes these kinds of exceptions to the “male default” rule.
The close resemblance of Chinese and English in the conception and
use of gendered terms seems as strong as for the other dyads we have con-
sidered above. We will look next at other examples of how far this resem-
blance extends.
Metaphors That Chinese and English (Pretty Much) Share
When we study conceptual metaphor across language and culture, one of

the attractions is our curiosity to discover whether there are different
ways to “conceive the world.” A book on Chinese and English metaphor
by Dilin Liu captures this aspiration in its title, Metaphor, Culture, and
Worldview.139 Do Chinese and English have different “worldviews”? How
much? If it is true, as a variety of philosophers and cognitive scientists
have claimed, that many conceptual metaphors are either determined by
the common features of human experience or, perhaps, built into the
structure of the human mind, then we should expect there to be consider-
able congruity in worldview across languages. Much as we might enjoy
taking note of cultural differences, there is no getting around the fact that
much of human life everywhere is obviously—boringly, one might say—
similar. People have two arms, ten fingers, and a nose, walk upright, live
in families, eat, sleep, feel pain when pinched, and have DNA that differs
by less than 1 percent from anyone else’s. Much of language, too, falls into
this area where commonality is taken for granted. The basic equipment of
the human mouth determines a common repertoire of oral sounds. And
studies have revealed what seem to be commonalities in the brain’s equip-
ment as well. To give just one example, it seems that in all human lan-
guages, the categories that have primacy within taxonomies are those that
139. Dilin Liu, Metaphor, Culture, and Worldview: The Case of American English and
the Chinese Language (Lanham, Md.: University Press of America, 2002).
are most often usable in daily life. Chair, for example, is more basic than
furniture or wooden things on the one hand and rocker or Granny’s rocker on
the other.140
We need to recognize that when we study “culture” or “worldview” we
usually have a bias toward focusing on differences. We do this because the
differences are more interesting. I would not get many students in my
Chinese culture courses if I spent my lectures explaining that Chinese
people have two arms, ten fi ngers, a nose, and so on, even though these
are extremely fundamental facts about Chinese life. It is more interesting
to compare the culture of chopsticks with that of knife and fork, even
though chewing, swallowing, sitting on a chair, digesting, defecating, and
so on comprise an area of overlap that is immensely larger and more com-
plex than the difference between chopsticks and a fork.
Yet it remains important (not merely interesting) to study cultural dif-
ferences, in large part just because we human beings make such a big deal
of them. (It is ironic that our zest for noticing differences is one of the
many ways we are the same.) Our differences set the stage for much of
what we mean by “understandings” and “misunderstandings” across cul-
tures and also provide the basis for a wide range of activity—from univer-
sity departments of national languages all the way to international rival-
ries and even wars. A focus on ten fi ngers, the nose, and DNA overlap
would not lead in these directions.
In this section and the next, I want to look at a range of examples that
metaphor theorists have found in English and ask how much Chinese may
differ from them. I will begin with examples that seem significant for
their similarity and move toward ones that seem interesting for their dif-
ferences. The reader should not expect any movement toward systematic
contrasting “worldviews,” however. I do not find such things, and doubt
that they are there to be found.
In looking for conceptual metaphors that are similar between the two
languages, we need first to note the special case of borrowing. Chinese,
140. Objective criteria for what is meant by “primary” or “basic” here can be spelled
out. See Lakoff, Women, Fire, and Dangerous Things, pp. 32–34; Lakoff credits Roger
Brown, “How Shall a Thing Be Called?,” Psychological Review 65, pp. 14–21, for a “clas-
sic” statement on the matter.
Metaphor 185
beginning especially in late-Qing times, has borrowed a number of con-

ceptual metaphors from English and other European languages. Some of
these metaphors have taken root in Chinese in ways that diverge from how
they were used in Eu ropean languages. (We saw, for example, how “yel-
low” evolved from “scandal-mongering” toward “pornographic” after the
metaphor entered Chinese.) Other borrowings, especially more recent
ones, such as langfei shijian ⌾䌍ᯊ䯈 for “waste time,” carry the obvious
flavor of a loan. Sometimes the sense of “translatese” is overbearing. Ning
Yu cites the example of Zhongguo jingji zhengzai ruan zhuolu, guore de
zhuangtai yijing jiangwen Ё೑㒣⌢ℷ೼䕃ⴔ䰚, 䖛⛁ⱘ⢊ᗕᏆ㒣䰡⏽ ‘Chi-
na’s economy is making a soft landing, and its overheated state is cooling
down’.141
But having noted the problem of borrowings, let us now generally set
these examples aside. They have limited use in revealing the conceptual
habits of originally different languages. Let us turn to conceptual meta-
phors that are similar across languages even though they seem to spring
from independent roots.
A widely noted example, for many languages, is the metaphor “affec-
tion is warmth.” In English we say things like “a warm welcome” or “a
lukewarm friendship,” and in Chinese relie huanying ⛁⚜⃶䖢 ‘warm [in-
deed, hot] welcome’ and reqing jiedai ⛁ᚙ᥹ᕙ ‘warm reception’. The
counterpart metaphor “unaffection is coldness” has been noted as well,
and in English we have “cool reception” and “sexual frigidity” just as, in
Chinese, we have lengmo ‫‘ ⓴ދ‬cold and detached’, lengxiao ‫ދ‬ュ ‘cold
laugh— sneer’, or lengbingbing ‫‘ ބބދ‬frosty (in attitude)’. To brush some-
one off in English we can give a “cold shoulder”; in Chinese we do the
same thing with the eye: lengyan ‫‘ ⴐދ‬cold eye’. Lakoff and Johnson sug-
gest that the origin of the “affection is warmth” metaphor is our experi-
ence as infants of a parent’s embrace, in which warmth and affection are
delivered simultaneously.142 “Unaffection is coldness” presumably arises
as a natural corollary to “affection is warmth.” John Searle, on the other
hand, has argued that a metaphor like “unaffection is coldness” works quite
well without an experiential base and without any need for one. Searle

shows in detail how the metaphorical statement “Sally is a block of ice”

cannot be interpreted in terms of literal statements about Sally and ice
blocks that show the two to have any common attributes.143 Yet the meta-
phor is not only understandable but similar to others that have arisen in a
variety of human cultures. Searle concludes that “the notion of being cold
just is associated with being unemotional.”144
If warmth means “affection,” it is also true—in Chinese, English, and
apparently many languages—that when the degree of warmth gets high,
the metaphorical meaning turns from affection to anger. We get “hot
under the collar” in English and fa huo ⱐ☿ ‘emit fire’ in Chinese. A re-
lated metaphor sees anger as “fluid under pressure” or “fluid in a container,”
as when you get “steamed up,” “blow your stack,” or “explode” in English
or, in Chinese, sheng qi ⫳⇨ ‘produce vapor—get angry’, fa piqi থ㜒⇨ ‘emit
spleen gas—throw a tantrum’, or become qihuhu de ⇨੐੐ ‘gas puff-puffy—
panting with rage’. Is there a universal physical basis for these metaphors?
Lakoff is impressed with experimental results from Paul Ekman and col-
leagues that show increased skin temperatures and heart rates when hu-
man beings become angry.145 Lakoff hypothesizes that among “languages
of the world, we will not find any that contradict the physiological results”
of this study.146
This claim is open to doubt. Several studies, which go well beyond my
brief analysis here, have shown that there is indeed considerable congru-
ity between Chinese and English uses of “anger is heat” metaphors.147 But
143. Searle, “Metaphor,” p. 96.

144. Ibid., p. 98.
145. Paul Ekman, Robert W. Levenson, and Wallace V. Friesen, “Autonomic Ner-
vous System Activity Distinguishes among Emotions,” Science 221, no. 4616,
pp. 1208–1210.
146. Lakoff, Women, Fire, and Dangerous Things, p. 407.
147. Ning Yu, Contemporary Theory of Metaphor, pp. 52– 60; Zoltán Kövecses, “The
Concept of Anger: Universal or Culture Specific?,” Psychopathology 33 (2000), pp. 159–
170; Brian King, “The Conceptual Structure of Emotional Experience in Chinese,”
Ph.D. diss., Ohio State University, 1989. Ning Yu feels that he has discovered a minor
difference between Chinese and English metaphors for anger in that English uses
“hot fluid in a container” (p. 51) while Chinese uses “hot gas in a container” (p. 54).
But this seems simply a misunderstanding of English, where “fluid” means “liquid or
gas,” not just “liquid.”
Metaphor 187
whether or not actual body heat is the cause of the congruity seems far
less certain. The increases in blood temperature that Ekman and colleagues
found during anger were only 0.08– 0.10 degrees Centigrade. These are
very small differences— so small that one might reasonably ask why they
are not associated with “warm” affection rather than “hot” anger. “Nor-
mal” body temperatures vary in a person during a day by much larger
margins. If such small differences really are the basis for “anger is heat”
metaphors, it would have to be true that people notice them during anger
and associate them with the anger. Do they? Are such small differences
noticeable, especially when one’s psyche is presumably dominated by an-
ger? Moreover, since metaphors presumably are invented not at the very
moment of anger but some time later, when one is not experiencing the
anger but only recalling it, or perhaps observing it in others (in whom one
would need to infer the temperature rise empathetically), we would need to
postulate that the tiny temperature differences are either remembered in
oneself or inferable from observation of someone else. Neither postulate
seems likely.
Another problem is that Ekman and colleagues found that anger is as-
sociated not only with rises in body temperature but also with increases in
the rate of heartbeat that were, proportionally speaking, much larger.
During anger, heartbeat rates increased to eight more beats per minute
than normal. So why, on Lakoff’s hypothesis, would it not have been more
likely that we generate an “anger is faster heartbeat” metaphor? Some
have suggested that heartbeat increase underlies the “fluid under pres-
sure” metaphor, and there is some prima facie plausibility to this. We do
say things like “I was so angry that I almost blew an artery.” But there are
problems with this conjecture as well. Worry about bursting a blood ves-
sel depends on modern understanding of the circulatory system, some-
thing that came much later than the formation of conceptual metaphor in
both English and Chinese. In order to test whether an increased heart-
beat could have given rise to the metaphor “anger is fluid under pressure,”
we would need subjectively to feel a sense of pressure when angry. Do we?
I think so. But is it because of the increased heartbeat?
Yet another problem in identifying the origins of the “anger is fluid
under pressure” metaphor is that anger is not the only feeling with which
we associate an “internal pressure” metaphor. In English we say things
like “bursting with pride” and in Chinese manxin huanxi ⒵ᖗ⃶୰ ‘the
heart is filled with happiness’ or annabuzhu xinzhong de xiyue ᣝ᥎ϡԣᖗ
Ёⱘ୰ᙺ ‘cannot contain the joy in the heart’. In addition to studying
subjects who were experiencing anger, Ekman and colleagues studied
subjects experiencing happiness, and in the happiness cases found very
slight drops (0.03– 0.07 degrees Centigrade) in skin temperature; heartbeat
rates went up, as they did during anger, but to a lesser extent. These fi nd-
ings raise several questions: if both happiness and anger lead to increased
heartbeat, and if heartbeat indeed is a basis for the “fluid under pressure”
metaphor, then would we not be better off postulating a generic “emotion
is fluid under pressure” metaphor?148 This would accord with the physio-
logical data as well as with the metaphors that tell us that both anger and
happiness can overflow, be hard to contain, and so on. It would lend sup-
port to Lakoff’s “physiological universality” hypothesis, but at a higher
level of generality than Lakoff proposed. It would, however, solve prob-
lems only for the “fluid under pressure” metaphor, not the “anger is heat”
metaphor, for which the other problems I have pointed out would remain.
Let’s move to another example. Chinese and English are very close in
their use of “stinky is bad” as a conceptual metaphor. Here the underlying
experiential basis is fairly obvious. The undesirability of foul odor is, like
two arms and ten fi ngers, probably precultural. Of course, there are a few
things that are stinky and yet have been attractive to human beings (cod
lutefisk, jackfruit, camembert cheese), but these items are rare enough
that they only underscore the validity of the basic pattern. In English, any
number of things can “stink” metaphorically— a bad idea, a dull movie, a
naïve plan, your school’s volleyball team. Something might even “stink to
high heaven.” “I smell a rat” means that I have picked up an olfactory sign
that something is wrong. In Chinese, chou 㟁 is negative across a similar
range of application. When someone’s reputation is bad, he or she is chou.
A stupid move in chess is a chouqi 㟁ẟ ‘stinking chess piece’, a wild shot in
basketball a chouqiu 㟁⧗ ‘stinking ball’. When chou is used as an adverb it
still carries a strong negative flavor, as in choumei 㟁㕢 ‘show off disgust-
ingly’ or chouchi chouhe 㟁ৗ㟁ୱ ‘eat and drink with ugly abandon’. All of
148. Ning Yu does speak of an “emotions are fluids in a container” metaphor (Con-
temporary Theory of Metaphor, p. 67).
Metaphor 189
these uses are in basic harmony with “stinking” behavior in English. Chi-
nese does go a bit beyond English when chou metaphorically means “se-
verely” or “intensely,” as in chouma 㟁偖 ‘stinky scold— deliver a tongue-
lashing’, but the overall connotation remains negative. Chou doufu 㟁䈚㜤
‘stinky [fermented] bean curd’ might seem an exception to the “stinky is
bad” metaphor, because people genuinely do like it. But chou doufu is not
an exception because it is not a metaphor. It stinks literally, like lutefisk or
camembert. (What separates Chinese people from Swedes or French in
this regard is only truth in labeling.) The point of the term chou doufu is
not to say metaphorically that the bean curd has a bad reputation or is
showing off.
Other kinds of metaphor that Chinese and English share are metaphors
for difficulties.149 “Difficulties are burdens” is one example, as when we
say in English “she’s weighed down by responsibilities.” Lakoff and Johnson
suggest that the experiential basis for this metaphor is “the discomforting
or disabling effect of lifting or carry ing heavy objects.”150 Chinese simi-
larly uses fudan hen zhong 䋳ᢙᕜ䞡 ‘burdens are heavy’ both literally and
metaphorically. When burdens are relieved, one can say in Chinese ru shi
zhongfu བ䞞䞡䋳 ‘as if setting down a heavy load’. Even more common, in
both languages, are metaphors in which difficulties are blockages or im-
pediments. In English we sometimes try to “get around” regulations, “get
through” a trial, or “hack our way through” a bureaucratic jungle. Some-
times we “run into a brick wall.”151 In Chinese, too, difficulties can be
zhang’ai 䱰⹡ ‘obstacles’ that zu’ai 䰏⹡ ‘block’ one’s progress. The chal-
lenge is to xiaochu ⍜䰸 ‘dispel, remove’ or paichu ᥦ䰸 ‘shove aside, get rid
of’ such things. During the Cultural Revolution, Mao Zedong enjoined
the Chinese people to paichu wannan, qu zhengqu shengli ᥦ䰸ϛ䲒এѝপ㚰
߽ ‘push aside every difficulty and go pursue victory’. Specific images in
Chinese often differ from those of English even while the basic conceptual
metaphor remains the same. “Stumbling block” in English uses the same
image as banjiaoshi ㌚㝇⷇ ‘foot-tripping stone’ in Chinese; lanluhu ᫨䏃㰢
149. For a variety of examples see Lakoff and Johnson, Philosophy in the Flesh,
pp. 188–190, and Ning Yu, Contemporary Theory of Metaphor, pp. 202–211.
151. The examples are adapted from ibid., p. 189.
‘road-blocking tiger’ uses a different image but still illustrates the concep-
tual metaphor “difficulties are impediments.”152
The English word “comprehend,” which comes from Latin com ‘jointly’
plus prehendere ‘grasp’, illustrates the metaphor “understanding is grasp-
ing.” This metaphor appears elsewhere in modern English in phrases like
“I can’t grasp transfinite numbers.”153 Chinese uses a similar metaphor
when bawo ᡞᦵ ‘grasp’ means “confidence” in a mental ability: kai jiaoche
wo you bawo, kai kache mei bawo ᓔ䕓䔺៥᳝ᡞᦵˈᓔव䔺≵ᡞᦵ ‘I know
what I’m doing driving sedans, not driving trucks’. There is a subtle dif-
ference, though. Bawo in Chinese is used for having confidence in how to
do something, not for the more purely mental experience of “getting it”
about something like transfi nite numbers. For that meaning, Chinese of-
ten uses mingbai ᯢⱑ, literally “bright white,” as in wo mingbai ni de yisi ៥
ᯢⱑԴⱘᛣᗱ, literally “I bright-white your meaning.” This usage illus-
trates another systematic metaphor in which brightness, in one way or an-
other, is “good.” As an adjective, mingbai can mean “clear, pellucid” in a
positive sense. Ming by itself can mean “open, honest” as opposed to an
ᱫ, literally “dark” and metaphorically “hidden, covert,” as in mingren bu-
zuo anshi ᯢҎϡ‫خ‬ᱫџ ‘honest people don’t do dark [untransparent] things’.
Yanming ‘eyes are bright’ means sharp-eyed in yanming shoukuai ⴐᯢ᠟ᖿ
‘sharp of eye and deft of hand’ or yanming xinliang ⴐᯢᖗ҂ ‘sharp eye
and bright mind— see and think clearly’. “Eyes bright” also contributes in
the phrase ercong muming 㘇㘾Ⳃᯢ ‘ears sharp eyes bright’, whose short
version, congming, is the common word for “smart.” Liang ‘bright’ partici-
pates in the “bright is good” metaphor independently of ming; when we
dakai chuanghu shuo lianghua ᠧᓔに᠋䇈҂䆱, we “open the window and say
bright words”—that is, say what we really mean. After we do this, some-
one might say wo xinli liang le ៥ᖗ䞠҂њ, literally “my mind is bright
now,” that is, I understand. Piaoliang ⓖ҂ ‘spiffy-bright’ is the common
word for “pretty.” English uses essentially the same metaphor in phrases
like “a bright child” and “dark days ahead.” Exceptions to the pattern tend
to apply only in restricted ranges; “dark” is good in “tall, dark, and hand-
152. Some of the examples in this paragraph are borrowed from Ning Yu, Contem-
porary Theory of Metaphor, p. 203.
153. Lakoff and Johnson, Philosophy in the Flesh, pp. 54, 125.
Metaphor 191
some,” but this phrase applies only in certain contexts, often to dashing
young men of a certain kind.154
Lakoff and Johnson fi nd “important is big” to be a conceptual meta-
phor of English— as when we say “tomorrow is a big day.”155 In Chinese,
too, we say that an important official is a daguan ໻ᅬ ‘big official’ and a
major surgical operation is a dashoushu ໻᠟ᴃ ‘big operation’. Lakoff and
Johnson conjecture that the connection between “big” and “important”
may derive from childhood, when, in a child’s view, parents are consis-
tently both big and important. They also suggest that the connection
between “big” and “important” has a parallel in the form of language,
specifically that “more of form is more of content.” For example, to say that
something “is bi-i-i-i-ig,” drawing out one’s voice, gives the impression
that it is bigger than if one just says it is big.156 Although they do not say
so, here Lakoff and Johnson are relying on another conceptual metaphor:
“longer in time is more (of something),” which works in piggyback combi-
nation with “more of form is more of content.” In any case, their insight,
for this kind of example, does apply in Chinese as well. Something hen
da-a-a-a-a! in Chinese is bigger than something merely hen da ᕜ໻
‘big’.157 Big and da might be confusing examples because the words them-
selves mean “big,” so the idea of big is coming from both the form of the
word and its referent. But the principle holds with other vocabulary as
well. To emphasize that one can do something, in English one might say
“I ca-a-a-a-n!” and in Chinese wo h-h-u-u-u-i-i! ៥Ӯ- - -.
Lakoff and Johnson cite duplication as another case in which “more
form is more content.” When a noun, verb, or adjective stands for some-
thing, the same word repeated can stand for more of the same or a higher
degree of it.158 They write that the duplication principle applies, as far as
they know, to “all languages of the world,” but for Chinese it works only
154. I am grateful to Nicholas Admussen for this and other examples in this
section.
157. When hen is not stressed, hen da means ‘big’, not ‘very big’. See Ta-tuan Ch’en
et al, Chinese Primer: Notes and Exercises, rev. ed., vol. 2 (Princeton, N.J.: Princeton
University Press, 2007), p. 27, n. 7.
unevenly. It often holds for nouns and auxiliary nouns, such as renren ҎҎ
‘person-person—everybody’, jiajiahuhu ᆊᆊ᠋᠋ ‘every house and home’,
tiantian ໽໽ ‘every day’, and so on. When Gao Xingjian describes a pot-
holed highway as daochu kengkeng wawa ࠄ໘ഥഥߍߍ ‘pits and cavities
everywhere,159 the idea is that there are a great many of both keng ‘pits’
and wa ‘cavities’. But the principle does not work for other nouns, such as
baba ⠌⠌ ‘Daddy’, meimei ྍྍ ‘younger sister’, taitai ໾໾ ‘Mrs.’, baobao
ᅱᅱ ‘baby’, xingxing ‘star’ ᯳᯳, xingxing ⣽⣽ ‘ape’, and many other ex-
amples.160 Sometimes, especially in southwestern Mandarin (around Yun-
nan), noun repetition implies smallness, as in qiuqiu ⧗⧗ ‘small ball or
bead’.161 This tendency for repetition to suggest small or less contradicts
the hypothesis that “repetition is more.”
Lakoff and Johnson hold that “more form is more content” applies to
verbs and adjectives as well, that is, that “more of the verb [or adjective]
stands for more of the action [or property].”162 But here, too, Chinese fits
only unevenly. Repeated adjectives in Chinese do often intensify the de-
gree of something, as in chi de baobao de ৗᕫ佅佅ⱘ ‘eaten good and full’,
zhongzhong de da 䞡䞡ഄᠧ ‘beat fiercely’, jiejieshishi de 㒧㒧ᅲᅲⱘ ‘strong
and solid’, and so on. Sometimes, though, the repetition of adjectives di-
minishes the strength of the adjective, or at least makes it a bit fuzzy, as in
wuzili heihei de ሟᄤ䞠咥咥ⱘ ‘darkish in the room’ or yanlei wangwangr de
ⴐ⊾∾∾‫‘ ⱘܓ‬teary-eyed’.
Repeated verbs, for their part, almost never signal “more” of the verb.
They are commonly “verb plus cognate object,” as in kankan (short for
kanyikan ⳟϔⳟ) ‘take a look’, chichi ৗৗ ‘have a bite (of something)’, and
so on, and, somewhat less often, are verb-plus-complement, as in men kai-
kai le 䮼ᓔᓔњ ‘the door has been opened’. In both these cases, the second
verb has a clear function, but it is not that of “adding more verb.” There is
the problem, too, that in Chinese, repetition occurs in a large variety of
159. Gao Xingjian, Lingshan 䴜ቅ (Soul mountain) (Taipei: Lianjing chubanshe 㙃㍧

ߎ⠜⼒, 1990), p. 1.
160. See Yuen Ren Chao, A Grammar of Spoken Chinese (Berkeley: University of
California Press, 1968), pp. 198–210. Several of my examples here are drawn from
these pages of Chao.
161. Ibid., p. 202.
Metaphor 193
onomatopoeic and other lively expressions that are not onomatopoeic but
are close to it. Dogs say wangwang ∾∾ ‘bowwow’; something sourish is
suanbuliuliude 䝌ϡ⑰⑰ⱘ (literally ‘sour-not-slippery-slippery-ish’); a
chubby child is pangdudude 㚪௳௳ⱘ (literally ‘fat-bunch-bunch-ish’). It is
hard to pinpoint in these examples what the repetition is telling us; but it
is something considerably more subtle than “more.”
In a similar argument, Lakoff and Johnson hold that the metaphor
“closeness is strength” can apply to form as well as to content.163 When we
say “the people closest to the prime minister” we might mean his or her
loved ones, and, in political discourse, can mean the people who have the
most influence on the prime minister. Chinese offers a range of similar
examples: people can be hen jin ᕜ䖥 ‘very close’, or guanxi hen miqie ݇㋏
ᕜᆚߛ ‘tightly related’, or, more colloquially, tie gemenr 䪕હӀ‫‘ ܓ‬iron
buddies’. That much of the “closeness is strength” metaphor seems clearly
to be shared between Chinese and English.
The additional claim that “closeness is strength” applies to the form as
well as to the content of language holds that a tighter grammatical struc-
ture can by itself suggest a tighter link between the items that are men-
tioned in a phrase. Some of Lakoff and Johnson’s examples involve the way
negatives are put. A sentence like “I think he won’t come” is both shorter
and stronger, they point out, than “I don’t think he will come.” Other
examples involve direct and indirect objects: “I taught Harry Greek” is
tighter and stronger than “I taught Greek to Harry”; the former implies
more clearly than the latter that some Greek actually got into Harry. “I
found the chair comfortable” is tighter and stronger than “I found that the
chair was comfortable,” because the former implies that the knowledge
comes from direct experience: I actually sat on the chair and found it
comfortable. “I found that the chair was comfortable” could also mean,
depending on context, that I observed my grandmother and discovered it
to be comfortable to her, or did a survey and found what consumers felt in
general, or something like that.
Does this “grammatical closeness is strength” metaphor work for
Chinese? On balance it does, although there are problems. In Chinese
the most common words for “think,” kan ⳟ and xiang ᛇ, cannot be
163. Ibid., pp. 128–132.

negated,164 so there is no handy counterpart for sentences like “I don’t

think he will come.” We can say in Chinese wo kan [or xiang] ta buhui lai
៥ⳟ[ᛇ]ҪϡӮᴹ ‘I think he won’t come’ but we do not say wo bukan [or
buxiang] ta hui lai ‘I don’t think he will come’. (René Descartes would have
loved Chinese. The language can make it awkward to “not think” some-
thing.) Using juede 㾝ᕫ ‘feel’, however, it is possible: wo juede ta buhui lai
៥㾝ᕫҪϡӮᴹ ‘I think [feel] that he won’t come’ and wo bujuede ta hui lai
៥ϡ㾝ᕫҪӮᴹ ‘I don’t think [feel] that he will come’ are both all right.
If we look for a parallel in Chinese to the case of comparing “I taught
Harry Greek” and “I taught Greek to Harry,” we might try saying wo jiao
le Zhang San Riwen ៥ᬭњᓴϝ᮹᭛ ‘I taught Zhang San Japanese’ and wo
ba Riwen jiao gei le Zhang San ៥ᡞ᮹᭛ᬭ㒭њᓴϝ . For these two sen-
tences it does seem, as Lakoff and Johnson predict, that the shorter ver-
sion gets more Japanese into Zhang San than the longer one. But a prob-
lem with this experiment is that in Chinese, neither of the two sentences
seems very natural. They are the kind of thing one thinks of when trying
to look for Western-language counterparts (which is just what we are do-
ing here). Chinese people in normal contexts would be much more likely
to say something like wo shi Zhang San de Riwen laoshi ៥ᰃᓴϝⱘ᮹᭛㗕Ꮬ
“I was [and am] Zhang San’s Japanese teacher’. In short, we will probably
do better if we test the “grammatical closeness is strength” hypothesis on
more native turf.
In Chinese, the example of the verbal complement might be useful. A
verbal complement is a second verb that immediately follows the first and
tells how the action of the first ends up. For example ting dong ਀ព, literally
“listen understand,” suggests that listening (to oral instructions, a lecture,
etc.) ends up in understanding (them, it). The formal relation of the two
verbs is immediate juxtaposition, and the semantic connection between
their senses is similarly tight. If the verbs are separated by the particle de ᕫ,
the connection is a bit weaker: now the idea is only that the fi rst verb can
result in the second. Ting de dong ਀ᕫព means something like “can under-
stand from having listened.” If bu ϡ ‘no’ intervenes, the connection be-
164. Xiang can be negated in its meaning of ‘would like to’, followed by a verb, but
that is a different matter.
Metaphor 195
tween ting and dong is explicitly negated but still carries the “can” idea; ting
bu dong is “cannot understand from having listened.” It is probably signifi-
cant that when even more syllables separate ting and dong in this kind of
construction, the meaning is always negative, that is, always expresses a
weak connection between ting and dong. We can say, for example, ting bu
tai dong or ting bu da dong ਀ϡ໾ (໻) ព, both meaning “can’t understand
very well from listening”; we do not say either ting de da dong or ting de tai
dong, “can understand well from listening.”165 This is hardly overwhelm-
ing evidence for Lakoff and Johnson’s claim that “closeness is strength” in
form as well as content. But it is supportive evidence, and it is authentic.
To move to another kind of example, it might seem that the “fronts” and
“backs” of things— animals, theaters, railroad trains, and many things—
might be metaphors that Chinese, English, and other languages share.
But it turns out that this is only partly true. In order to understand why,
we need to distinguish three different criteria for determining “front and
back”:
1. Things whose fronts and backs are defined by the direction in

which they move (e.g., animals and vehicles).
2. Things whose fronts and backs are defi ned by the layout and
functions of their features (e.g., stores or theaters, where signs hang at
the “front,” one usually pays near to the “front,” etc.).
3. Things (such as mountains and trees) that do not move and do not
have intrinsic fronts or backs but have “fronts” and “backs” attributed
to them according to the position of a speaker or writer.
In category 1, the “fronts” of animal bodies are commonly understood

as the sides that face in the direction in which the animals normally move.
The “fronts” of an ant, a hippopotamus, and a snake are all obvious by
this criterion. Some animals, including humans, can walk backward, but
we have no problem identifying such motion as exceptional. Crabs that
165. It is possible to say things like ting de shifen dong ‘completely understand’ 㙐ᕫ
कߚព or ting de wanquan dong ‘entirely understand’ 㙐ᕫᅠܼព, but these are not
verb-complement constructions. They are predicate constructions.
“go sideways” might be a special problem, but here category 2 can help—
and it helps for many animals other than crabs—because the front is the
“operational” side, the side where the eyes, mouth, and nose all do their
important work. A crab eats at its “front,” and appears even to be “look-
ing” from that side; in this case we do not give serious thought to the pos-
sibility that its “front” is the edge that leads the way when it moves.
Lakoff, Johnson, and others point out that our conceptions of the
“fronts” of inanimate things are sometimes done by analogy to animate
ones.166 The “front” of a bus, a train, or an airplane is the end that leads
the way when the thing gets going. This rule seems to hold consistently in
both Chinese and English. For things that do not move, like stores or
theaters, the “front,” as with animals, is the operational side, where facades,
signs, box offices, and so on are located. In the case of theaters, it is inter-
esting to note that once we enter the building and get into the big room
where the show is held, front and back instantly switch. The “front row”
is near the back of the building. This is because the conception “opera-
tional side” has switched. Once one is inside the theater, the stage is where
the important activity is. All these points are the same in Chinese and
English.
But there is a problem in holding that English and Chinese “share
metaphors” of front and back. This is because what we have just consid-
ered are not metaphors but definitions of front and back. To observe that
they are “the same” is, at bottom, tautologous. To see why this is so, try
the following experiment. What if Chinese conceived the “fronts” and
“backs” of things like trains and buses oppositely from English? Then
qian would be the back and hou the front. But if a train came along, and a
Chinese speaker pointed to the end that was in the direction of the train’s
motion and said houtou ৢ༈ ‘back’, an English speaker would infer—
correctly—that houtou in Chinese is the same as “front” in English. Simi-
larly, if the side of a theater building bearing the facade and harboring the
box office were called the “back” in English, a Chinese speaker would
have no trouble inferring that “back” is qiantou ࠡ༈. If we were to assign
nonsense words to the box office side of the theater— say, “blint” in En-
glish and fing in Chinese—then “blint” would mean fing, and fing would

Metaphor 197
mean “blint,” and life would go on. To say that the languages “coincide”
on this point would be meaningless.
However, using category 3, it is genuinely possible (and does happen)
that different languages conceive front and back differently. Mountains—
unless they are carved, like Rushmore—have no intrinsic fronts or backs.
But a tongue twister in Chinese begins shanqian you ge Cui Cutui, shan hou
you ge Cutui Cui . . . ቅ᳝ࠡϾየ㉫㝓, ቅৢ᳝Ͼ㉫㝓የ . . . ‘in front of the
mountain was a Cui Fatleg, and behind the mountain was a Fatleg Cui . . .’.
Here the front and back of the mountain are defined relatively. The front
is the side that the speaker and anyone standing near the speaker can see
on a clear day. The back is the opposite side. If the same people climbed
the mountain and went to the other side, front and back would switch. We
do the same in English. Boulder, Colorado, is in front of the mountains if
we are in Denver, and behind the mountains if we are in Grand Lake. But
this convention is arbitrary. It could be the other way around. We could
conceive the other side of the mountain, which we cannot see, as “out in
front,” as it were, while we are behind it. We could stand in Boulder and
see Grand Lake as “in front” of the mountains. Lakoff and Johnson re-
port that Hausa does conceive front and back in this alternate way.167
For a mountain itself, Lakoff and Johnson observe that English uses the
metaphor “a mountain is a person.” They cite the word “foothills” as evi-
dence; in English, mountains have feet. But they classify this as an “idio-
syncratic, unsystematic, and isolated” metaphor, not a “metaphor that we
live by,” because the “unused portions” (presumably the head, shoulders,
knees, etc. of a mountain) are not standard phrases in English and would
be taken as novel metaphors if a poet or other creative person were to
choose to use them. Chinese also uses the metaphor “a mountain is a per-
son” and does “live by it” a bit more than English, because in Chinese
shantou ቅ༈ ‘mountain head’ and shanyao ቅ㝄 ‘mountain waist’ (i.e., half-
way up) are just as acceptable as shanjiao ቅ㛮 ‘mountain foot’.
There are myriad ways the instantiations of conceptual metaphors
might differ among languages, and to try to catalogue them all would be
tedious. In using the human body to describe vegetables, for example, we
say yitou dasuan ϔ丁໻㩰 ‘a head of garlic’ in Chinese, but a “bulb” in
167. Ibid., p. 34.

English; on the other hand we use “head” in English for lettuce and cab-
bage. The comparisons between languages that are more significant in
revealing “worldviews” are those that differ in more systematic ways, and
we turn to them now.
Metaphors in Chinese That Diverge from English in Significant Ways
Lakoff and Johnson make a broad claim that “the most fundamental val-
ues in a culture will be coherent with the metaphorical structure of the
most fundamental concepts in the culture.”168 They are careful to point
out that “be coherent with” does not mean “will always actually exist,” but
only that whatever values do exist will be “consistent with” the metaphorical
system.169 They suggest—but only occasionally illustrate—that different
cultures observe metaphorical patterns that reflect different values.
In his book Metaphor, Culture, and Worldview, Dilin Liu takes up this
question for the cases of modern Chinese and American English. Liu iden-
tifies a few of what he calls “dominant metaphors” in the two languages. By
“dominant” he means, in part, common, but also apparently something
like what Lakoff and Johnson mean by “conceptual” or “structural” meta-
phor in the sense of metaphor that not only reflects the ways we conceive
things but actually shapes conceptions as well. Liu argues that dominant
metaphors are different in modern Chinese and American English, and
that we therefore can probably infer that some of the values and priorities
of people who use the two languages diverge along similar lines.
For example, Liu points out that when Americans are unconvinced of
an argument they might say “I don’t buy that,” whereas in Chinese, in a
similar situation, one might say wo bu chi ni nei yi tao ៥ϡৗԴ䙷ϔ༫ ‘I
don’t eat that stuff of yours’.170 He lists many other examples where mar-
keting metaphors are used in American English and eating metaphors in

169. Ibid., p. 23.
170. Dilin Liu, Metaphor, Culture, and Worldview, p. 9. The example is Liu’s, the
translation mine.
Metaphor 199
Chinese. Unfortunately, he does not do actual counts of metaphor fre-

quency in texts or oral recordings, and gives only anecdotal evidence for
his claim that the thinking and values of the two cultures correspond to
their metaphor preferences. Still, his broad claims do seem plausible. In
addition to the marketing and eating examples, he lists sports (e.g., “the
ball is in your court”) and the driving of vehicles (e.g., “my wheels are
spinning”) as especially fertile sources of metaphor in American English
and, for modern Chinese, family relations (e.g., disidai ‘fourth generation
[of leaders]’) and stage performance (e.g., chang gaodiao ‘sing high-sounding
words’) as especially common.
It is intuitively obvious that metaphorical uses of chi ‘eat’ in Chinese far
exceed, both in frequency and in range of applicability, uses of “eat” (or
“ate,” “eaten,” etc.) in English. Very broadly speaking, the person who
metaphorically “eats” something in Chinese can either gain by the eating
or suffer from it. On the gain side, there is chi yige zir ৗϔ‫ן‬ᄤ‫‘ ܓ‬eat a
piece’, for capturing a piece on a chessboard, or chidiao dijun ৗᥝᬠ‫‘ ݯ‬eat
up the enemy army’ for annihilating enemy troops. To enjoy popularity is
chi xiang ৗ佭 ‘eat fragrance’, and to receive a kickback is chi huikou ৗಲᠷ,
literally ‘eat return discount’. Chi taiping fan ৗ໾ᑇ佁 ‘eat great-peace
food’ means passing one’s days in comfort. Chi doufu ৗ䈚㜤 ‘eat bean
curd’ is a subtler example but also illustrates the principle that the “eater
gains,” because here bean curd stands for the soft white flesh that a male
metaphorically nibbles when he flirts with a female. In chi laoben ৗ㗕ᴀ
‘eat original capital—live on past gains’, the long-term consequences are a
loss, but the immediate result of chi is still one in which the eater gains.
On the other hand— and more commonly—chi can suggest that the
eater suffers in some way. Chi kui ৗѣ ‘eat loss’ means to get the worst of
something or, in a corresponding English idiom, to “come out on the short
end of the stick.” Chi zui ৗ㔾 ‘eat crime’ is to take the blame for some-
thing. Chi ku ৗ㢺 ‘eat bitterness’ means to suffer hardship, and chi jin
ৗ㋻ ‘eat tension’ to be tense or hard pressed. Chi jing ৗ᚞ ‘eat surprise’
is to be startled or shocked; chi bai zhang ৗ䋹ҫ ‘eat defeat in battle’ is
to lose in battle; and chi guansi ৗᅬৌ ‘eat a lawsuit’ is to be charged in a
suit. Here the “eaten” things are all abstract concepts, but concrete
nouns, understood metaphorically, can also be “eaten,” as in chi heizaor
ৗ咥ᵷ‫‘ ܓ‬eat a black date’ for being hit by a bullet, or chi cu ৗ䝟 ‘eat vin-
egar’ for feeling jealous. In chibuxiao ৗϡ⍜ ‘eat and cannot digest’, it is
not clear what is eaten or suffered except that, whatever it is, the thing is
unpleasant. When a ship is loaded and, as English puts it, “draws” water
to a certain depth, Chinese says chi shui ৗ∈ ‘eats water’ to the same
depth. Here the question of gain or loss seems moot, but the metaphor
remains apt.
Wanting to test objectively whether Dilin Liu is right that chi is used
metaphorically in Chinese more than “eat” is in English, I did a word
search of two novels, chosen essentially at random: Lao She’s novel The
Philosophy of Lao Zhang (Lao Zhang de zhexue 㗕ᔉⱘ૆ᅌ) in Chinese and
Mark Twain’s novel The Mysterious Stranger in American English. In
Lao She’s text, chi accounts for one in every five hundred characters, and
in Twain’s text “eat” (including “ate” and “eaten”) appears once in every
twenty-five hundred words.171 This difference is accountable in part to the
fact that eating in a literal sense gets more attention in Lao She’s novel
than in Mark Twain’s. (That fact, though, can be viewed as confi rmation
of Liu’s claim that both references to actual eating and metaphors based
on eating are especially salient in Chinese discourse.) The contrast in
metaphorical uses of chi and “eat” was sharper: there are five instances in
Lao She’s novel, none in Twain’s.
Another of Liu’s examples of dominant metaphor in Chinese is “gov-
ernment is family.” For this example, he is able to show how the metaphor
is strong enough, or “conceptual” enough, to travel through time and
over varied political terrain.172 That the “Five Relations” of Confucianism
list jun ৯ ‘sovereign official’ and chen 㞷 ‘subordinate official’ alongside
family relations like fuzi ⠊ᄤ ‘father and son’ and fuqi ໿ྏ ‘husband and
wife’ is already suggestive of a conceptual parallel between political hier-
archy and family relations. Traditionally, guojia ೟ᆊ, literally ‘country-
family’, had several senses, one of which was “emperor,” and in modern
171. I am grateful to Mao Sheng for research assistance in this effort. Mao found
177 appearances of chi in The Philosophy of Lao Zhang, which amounts to 0.19 percent of
the total character count. In The Mysterious Stranger there were 9 instances of “eat,”
7 of “ate,” and 1 of “eaten,” which was 0.041 percent of the total word count.
172. Dilin Liu, Metaphor, Culture, and Worldview, p. 8, 55– 64, 131–133.
Metaphor 201
times guojia came to be the standard word for both “nation” and “state.”
In Qing times, a popular term for “county magistrate” was fumuguan ⠊↡ᅬ
‘parental official’. There are many ways to refer to “the people” in modern
Chinese, one of which is zhonghua ernü Ё㧃‫ܦ‬ཇ ‘sons and daughters of
China’. Chinese is not unique in using family metaphors for the nation, of
course; there are “founding fathers” in English, la patrie ‘the fatherland’
in French, Mat’ Rossiia in Russian, and many other examples. But Dilin
Liu may be right that the metaphor is especially salient in Chinese. De-
spite its traditional— or what the Communists in other contexts have
called “feudal”—roots, it has extended even into the informal lingo of the
Communist movement itself. When Marx and Lenin arrived in China,
the Party called them lao zuzong 㗕⼪ᅫ ‘old ancestors’. During the Mao
era, factory workers were sometimes gongren dage ᎹҎ໻હ ‘worker elder
brothers’, farmers nongmin bobo ‫⇥ݰ‬ԃԃ ‘peasant elder uncles’, and sol-
diers jiefangjun shushu 㾷ᬒ‫ݯ‬নন ‘liberation army uncles’.173 The Soviet
Union, until relations with Communist China turned sour, was a laodage
㗕໻હ ‘old big brother’. And the Hu Jintao cohort of top leaders was
known as the “fourth generation.”
Of the “dominant” metaphors that Dilin Liu finds in modern Chinese,
the one that has, in my view, the deepest implications is the one he calls
the “opera/acting metaphor.”174 Liu cites a number of examples to show
how public presentation of behavior is important in Chinese culture and
how metaphors of stage performance, many derived from Chinese opera,
are used to talk and write about it. (In reference to China, by the way,
“opera” should not be thought of as an elite art. Some of its modern forms
are indeed elite, but most of what have for centuries been called xi ៣ are
popular performing arts that have a wide variety of local traditions. Be-
fore the twentieth century, xi were the staples of popular Chinese enter-
tainment.) Liu Dilin notes that for an official to take office is to shang tai
Ϟৄ ‘go up on stage’, while to leave office is to xia tai ϟৄ ‘come down
from the stage’. In a broader sense, any person (not necessarily an official)
who takes a publicly visible stand can be said to deng tai ⱏৄ ‘ascend the
stage’. A person can get stuck on stage (i.e., publicly committed to an
173. Ibid., p. 58.

174. Ibid., pp. 103–109, 133–135.
awkward position), in which case his or her adversary, who might not
want to let him or her “off the hook,” as we might say in English, can rang
ta xiabuliao tai 䅽Ҫϟϡњৄ ‘make it so he/she cannot get off stage’.
While a person is “on stage”—that is, visible to people who are
watching—how that person performs is sometimes described using chang
ଅ ‘sing’. There are a number of ways to chang, metaphorically speaking.
To chang honglian ଅ㑶㜌 ‘sing the red face’ is to play the role of the good
guy. The phrase derives from the convention that a red face in opera
makeup usually indicates a good person. In popular usage, it does not
have to mean that one is a good character but only that one is presenting
the good-character pose. The pose can be insincere, so that the phrase
can mean, in effect, “pretend to be the good guy.” Because villains in op-
era often have white faces, chang bailian ଅⱑ㜌 ‘sing the white face’ means,
correspondingly, to play the role of a bad guy. Chang gaodiao ଅ催䇗 ‘sing
a high-pitched tune’ is to say fi ne-sounding things, often with the nega-
tive connotation of overweening self-righteousness; didiao chuli Ԣ䇗໘⧚
‘handle things at a low pitch’ is usually regarded as better. Chang dujiao xi
ଅ⣀㾦៣ ‘sing a solo’ is to do something without partners or allies. For set-
ting up a rival enterprise to someone else’s, Chinese can say chang duitai xi
ଅᇍৄ៣, literally ‘sing opera on an opposite stage’. If you want explicitly
to oppose someone, you can chang fan diao ଅড䇗 ‘sing a contrary tune’; if
you need to go along with somebody else’s pretences in order to get what
you need, you might jiaxizhenchang ‫؛‬៣ⳳଅ ‘sing a phony opera as if it
were real’. If you are only going through the motions of doing something
you just zou guo chang 䍄䖛എ ‘walk across the scene’, and if you play but a
minor role, running errands and the like, you pao longtao 䎥啭༫ ‘run the
dragon outfit’. (Imperial palace guards, when they played bit roles in op-
eras, wore “dragon outfits.”) If an enterprise collapses, one can say in Chi-
nese mei xi chang le ≵៣ଅњ ‘there is no opera to sing any more’. Virtually
any project at all, if it comes to naught, or was boring to begin with, can be
described as mei xi ≵៣ ‘no play’.
As Dilin Liu suggests, the extensive use of stage metaphors in Chinese
does indeed seem to correspond to some aspects of a Chinese “world-
view.” Recalling the Sapir-Whorf controversies, we might want to ask
whether the metaphors generate the worldview or the other way around.
For a case as complex as this one, I think “both” is no doubt the right
Metaphor 203
answer, and I believe we should be content to note the fact of correspon-

dence without undertaking to unwind the directions in which causality
runs. That said, there are two major features of “acting” in Chinese lan-
guage and culture that seem to me worth reflecting on: the importance
of outward per for mance and the moral value of doing the per for mance
correctly.
The association of formal language and correct per formance in China
is ancient. The earliest examples of written Chinese characters, on what
are called “oracle bones,” often had to do with advice about action: When
should the king go to war, or perform rituals? When should farmers
plant? The function of language was not to say how things were but to
advise on what to do. Speaking in general terms about pre-Qin philoso-
phy, Roger Ames and Henry Rosemont write that “language is both per-
formative and prescriptive; it both does something to the world and rec-
ommends how it should be.”175 They note that dao 䘧 ‘speak’ is the same
term as dao 䘧 ‘guiding [along the way]’.176
What might be called “Confucian psychology”—found in countless
texts but perhaps most classically in the Daxue ໻ᅌ (‘Great learning’)—
holds that classic texts contain a morality that can be internalized in a
person via memorization and lead to personal cultivation, which in turn
has the power to radiate outward and bring good effects, successively, to
the family, to the country, even to everything under tian ໽, the natural
order of the cosmos. The same assumption of the power of morality to
grow out of the study of classical language and to lead to personal cultiva-
tion and then to qualification to govern others was the main ideological
undergirding for China’s imperial examination system. Scholars who had
mastered classical learning were supposed to be able to display not only
technical ability in quoting texts or writing characters but also proper
moral responses to life situations— as, for example, in producing poems to
reflect the mood of poignant moments such as the seeing-off of a friend.
Peasant rebels, even if barely literate themselves, still often had one or
175. Roger T. Ames and Henry Rosemont, Jr., The Analects of Confucius: A Philo-
sophical Translation (New York: Ballantine Books, 1998), p. 31.
176. Ibid., attributing the insight to Chad Hansen, in Hansen, A Daoist Theory of
Chinese Thought (New York: Oxford University Press, 1992), pp. 33–54.
another kind of tianshu ໽᳌ ‘natural [righteous] document’, which served

as a moral warrant for their quest to change a dynasty. Even Mao Zedong,
the would-be “smasher” of all “old habits, old culture, old customs, and
old ideas,” encouraged a system in which, during the Cultural Revolution
in the late 1960s, the principles of what I am calling here “Confucian psy-
chology” were almost exactly reproduced: young Red Guards memorized
texts (containing now the words of Mao, not Confucius or other ancients),
internalized the unquestionable righteousness that the texts contained,
presented their resulting cultivation as qualification to set things right in
the larger society, and through it all forged a connection with the highest
authority under Heaven (now Mao, not an emperor).
The meanings of the correct words were paramount, of course. But the
powerful sense of their rightness could be enough to imbue even their
forms and vessels—written characters and oral sounds—with a special
moral glow. The art of Chinese calligraphy relies only in part on what its
written characters mean; the ways they are written—their “life” on
paper— are important signs of the character and cultivation of the callig-
rapher. (This can be true even when the characters themselves are hard to
identify.) At least until the mid-twentieth century, Chinese children were
warned to jingxi zizhi ᭀᚰᄫ㋭ ‘respect paper that bears characters’. Dur-
ing high Maoism this “old habit,” like others, persisted in altered form: it
became a political crime to use a piece of newspaper bearing a Mao quote
for casual purposes, such as stuffi ng shoes or wrapping fish.
As for sounds, the very sound of zao zi ᮽᄤ ‘sons soon’ explains why
dates (zaozi ẫᄤ) are eaten at weddings or, as noted in Chapter 1, why a
mother might pack chicken hearts ( jixin 叵ᖗ) in her son’s school lunch-
box in order to help him with his memory ( jixing 䆄ᗻ). The force packed
within the sounds of these two morally relevant syllables was assumed
to be strong enough to span the gap from label to object, from jixin to
chicken hearts, then to enter a boy’s alimentary canal, then his blood-
stream and his brain, then back across the mind-body gap from his
brain to his mind, and into his behavior—where the effects would be
“proper.”
Some scholars have argued that in ancient Chinese texts, the criterion
for deciding whether a statement is “right” is more nearly an ethical than
a cognitive criterion, that is, “do these words guide action properly?” and
Metaphor 205
not “are these words true?” Ames and Rosement write that “classical Chi-
nese has no close lexical equivalent for the English “true” and “truth”;177
A. C. Graham has held that followers of Mozi (ca. 470– 391 bce) “[do] not
use a single term corresponding to English ‘true’ [but hold that] a name or
complex of names applied to an object either fits (dang ⭊) or errs (guo 䘢),”178
and Chad Hansen has written that “[ancient] Chinese philosophy has no
concept of truth” and is built, instead, on “a pragmatic rather than a se-
mantic interest in language”; utterances are evaluated by whether they are
ke ৃ ‘admissible’ or ‘appropriate’, not whether they are true or false in the
sense that attracted much attention in ancient Greece.179
No one argues— or could, reasonably—that in daily life over the centu-
ries Chinese language ignores true-false distinctions any more than one
might say, to go too far in the other direction, that European languages
have trouble with appropriateness. Daily life in China, as anywhere, has
always been full of questions and answers about truth and falsity, and it is
worth noting that the kinds of Chinese literary expression that have been
closest to informal daily life— songs, storytelling, popular fiction, and so
on—have made much of deception, illusion, and other matters in which
truth and falsity are crucial. The great novel Dream of the Red Chamber (or
Story of the Stone) plays grandly with the true-false distinction, opening
with a dream by a gentleman archly named Zhen Shiyin ⫘຿䲅 (a hom-
onym for ⳳџ䲅 ‘true matters concealed’) and moving to the story of a Jia
Baoyu 䊜ᇊ⥝ (also “fake jade” ‫؛‬ᇊ⥝) who lives in a very large Jia (“false”)
family; yet there is much to say of interest, because, after all, jia zuo zhen
shi zhen yi jia ‫؛‬԰ⳳᰖⳳѺ‫‘ ؛‬when false poses as true, true is also false’,180
and so on. Here and elsewhere in Chinese language and culture, there is
no shortage of awareness of “true versus false.”
177. Ames and Rosemont, Analects of Confucius, p. 33.

178. A. C. Graham, Later Mohist Logic, Ethics, and Science (Hong Kong: Chinese
University Press, 1978), p. 39. Here I have converted Graham’s use of Wade- Giles
romanization to hanyu pinyin.
179. Chad Hansen, “Chinese Language, Chinese Philosophy and ‘Truth,’ ” Journal
of Asian Studies 44: 491–519 (1985), pp. 492, 504. I have added ‘appropriate’ to Hansen’s
‘admissible’ as a gloss for ke.
180. The phrase appears in chapters 1 and 5. Cao Xueqin ᳍䲾㢍, Hongloumeng ㋙ῧ໶
(Hong Kong: Youlian chubanshe, 1960), pp. 5, 45.
Still, those who note a Chinese cultural tendency to assume that proper
use of language is essentially ethical behavior—a performance, not a state-
ment, of correctness— do have an important point. This assumption helps
to explain why Dilin Liu can find “opera/acting metaphors” to be “domi-
nant” in Chinese even in modern times. Notions of correct performance are
embedded even in the grammar of daily-life Chinese, where they have sur-
vived through the turmoil of the modern era and its attendant language
change. To “say this” in Chinese one does not shuo zheige ‘say this’ (which
sounds awful), but zeme shuo 䖭М䇈, literally “say it this way”. If a child mis-
takenly says two plus three is four, her mother can correct her by saying
bushi neme shuo de—ϡᰃ䙷М䇈ⱘ “it’s not said that way” or “that’s not the
way you say it.” The child’s speaking-performance did not resemble the
right pattern. If the child strikes a sibling, throws food at guests, or misbe-
haves in any number of other ways, the mother might say ni zheyang bu-
xiangyang Դ䖭ḋϡ‫ڣ‬ḋ, literally “your acting this way does not resemble
the [right] pattern.” She might also say (and here we see how “correct lan-
guage” can stand be a stand-in for “correct behavior”) ni zheyang buxianghua
ϡ‫ڣ‬䆱 ‘your acting this way does not resemble [proper] words.’ She might
ask the child to xuehao ᄺད ‘learn to be good’ or xue zuoren ᄺ԰Ҏ ‘learn to
be [a proper] person’. Here xue, which is often translated as “study” or “learn,”
does not mean study or learn facts. It means “imitate.” A mother’s admoni-
tion to xue baba ᄺ⠌⠌ is a call to emulate Daddy, not to research him.
For matters of research, in which true-or-false is the governing ques-
tion, Chinese uses the word yanjiu ⷨお, but even here, more often than in
modern English, there can be a subtle tendency to assume that the object
of research is something worthy of emulation— a good person, good idea,
or whatever. In 1980, when I was studying contemporary literature in
China, I discovered some “hand-copied volumes” (shouchaoben ᠟ᡘᴀ)
containing detective stories, triangular love stories, martial arts stories,
and so on that had been secretly copied and passed around during the
Cultural Revolution for entertainment purposes. They struck me as fasci-
nating sources on popular thought during extraordinary times, and I
wrote a research paper on them.181 Some of my Chinese friends, though,
181. “Hand- Copied Entertainment Fiction from the Cultural Revolution,” in Perry
Link, Richard Madsen, and Paul Pickowicz, eds., Unoffi cial China: Popular Culture and
Thought in the People’s Republic (Boulder, Colo.: Westview Press, 1989), pp. 17–36.
Metaphor 207
were perplexed. Why do you write about that kind of literature? What is
there in it to “study” (by which they meant, implicitly, “imitate” or “learn
from”)? I had been aware that there were political and cultural sensitivi-
ties associated with shouchaoben, but the question in the minds of these
friends was something a bit different: their concern was that I could get
better moral sustenance by looking elsewhere.
During the same year, I did a survey of reading habits and preferences
among seventy-four students at Zhongshan University in Guangzhou,
and found that most of them listed Dream of the Red Chamber as their fa-
vorite work of fiction of all time, even though none acknowledged having
read it, at least not recently, or completely.182 Yet their answers to both
questions—“What do you read?” and “What do you prefer?”—were, I
feel sure, sincere. A foreign scholar was asking a formal question about
the way things should be, and “Dream of the Red Chamber” was the right
answer. Similarly, when Samuel Huntington’s Clash of Civilizations and the
Remaking of World Order appeared in 1997,183 some Chinese intellectuals
complained that Huntington should not be advocating civilizational
clash. “He should write a book about harmony, not clash,” a friend wrote
to me in an email. She assumed that Huntington was seeking to guide
behavior, not describe it.
In one interesting respect, even the grammar of modern Chinese
tends to reinforce the assumption that correctness is per for mance. This
happens with the extremely common structure that Y. R. Chao calls
the “predicative complement.”184 Chinese uses this structure in many
situations where speakers of Eu ropean languages would use adverbs. To
say, for example, “she sings (more) beautifully (than someone else),” Chi-
nese can say ta chang de haoting ཌྷଅᕫད਀, a literal English rendition
of which would be something like “her singing-manner is pleasant-
sounding.” In Chao’s analysis, the notion of “way” or “manner” is im-
plied between ta chang de ‘the [noun] of her singing’ and haoting ‘is
182. See Perry Link, “Fiction and the Reading Public in Guangzhou and Other
Chinese Cities, 1979– 80,” in Jeffrey C. Kinkley, ed., After Mao: Chinese Literature and
Society, 1978–1981 (Harvard University Council on East Asian Studies, 1985), pp. 256,
262–263.
183. Samuel Huntington, The Clash of Civilizations and the Remaking of World Order
(New York: Touchstone, 1997).
184. Chao, Grammar of Spoken Chinese, p. 355.
pleasant-sounding’.185 But the word-pattern itself leaves it theoretically

ambiguous whether the implied noun that follows de is “manner” or some
other noun. In ta chang de haoting, for example, the implied noun could
just as easily be “the song” that she sings, followed by the comment that it
is pleasant-sounding. Similarly ta shuo de hen hao ཌྷ䇈ᕫᕜད, literally
something like “her speech-performance was very good,” can mean either
“she said it very well” or “what she said was very good.” (It can also mean
both.)
Chinese speakers are accustomed to this ambiguity and very rarely feel
puzzlement, in context, about what is meant. The relevant distinctions
can always be made clear if necessary. Someone can ask, “Do you mean,
precisely, that her meaning was good— or that she expressed it well?” and
such a question is easy to understand and usually easy to answer.186 Still,
the fact that it is so very common in Chinese to speak in terms of “per-
forming” actions that are named by verbs does make it more natural, in
my view, that metaphors of stage performance are so pervasive.
In Chapter 3 we will explore ways in which “performance” of language
has had special importance in political contexts. In writing about Red
Guard activists during the Cultural Revolution, Anita Chan has observed
that the “playing of a role in China was more than a sociolog ical abstrac-
tion. Role-playing involved literal play-acting: a conscious assumption of
the mannerisms and ways of speaking appropriate to the activist status
and role.”187 More broadly— and not just for activists but for everyone
in Mao’s China—zhengzhi biaoxian ᬓ⊏㸼⦄ ‘political performance’ was
something that could determine job assignments, living conditions, ad-
missions to schools, and a variety of other crucial matters. Biaoxian comes
literally from biao ‘surface’ plus xian ‘appear’, and this combination of
notions puts the matter exactly right. A person had to present the right
185. Although de in the predicative complement is now commonly written ᕫ, not

ⱘ, Chao explains that ⱘ fits better with the genesis and logic of the pattern. Ibid.,
pp. 356–357.
186. This can be asked, for example, as ni shi shuo ta de neirong hao haishi ta de biaoda-
fangfa hao ne? Դᰃ䇈ཌྷⱘ‫ݙ‬ᆍད䖬ᰃཌྷⱘ㸼䖒ᮍ⊩དਸ਼? ’do you mean that her content
was good or her manner of expressing it?’
187. Anita Chan, Children of Mao: Personality Development and Political Activism in the
Red Guard Generation (Seattle: University of Washington Press, 1985), p. 214.
Metaphor 209
appearances in public even if they were different from what he or she was
feeling inside. The public persona formed of one’s biaoxian was on display
in the workplace or at school. It was especially relevant during political
study sessions. One had to pay attention to it, craft it, and guard it. By the
1970s, the maxim “Don’t make friends within the work-unit” had arisen
in Chinese cities. This was because life in the work-unit (factory, school,
government office, etc.) by necessity involved the formal presentations of
zhengzhi biaoxian, whereas friendships involved interchange that was more
frank and informal. The two levels did not mix well. Informal comments
made in the context of friendship, if raised to the level of political
performance—whether by accident or because of betrayal by a friend—
could bring serious trouble.
Conceptual Differences That Are Rooted in Metaphor
Dilin Liu’s claims that metaphors of sports, marketing, and the driving of
vehicles are salient in American English while metaphors of eating, family
relations, and acting are more common in Chinese are claims about de-
gree and emphasis only. Either language can use the metaphors of the other
to the extent that its speakers want to. Indeed this happens in direct-
translation borrowing, as when “bottleneck” in English becomes pingjing
⫊乜 in Chinese. But in some instances of metaphorical contrasts between
languages, the differences have more to do with the structure of thought
and are not so easily exchangeable. There is, in short, a useful distinction
between examples that are different by custom and those that are different
by concept.
A difference by custom, in this definition, is one in which arguably the
“same” conceptual metaphor is available in each of two languages, but one
language uses it more than the other. An example I have already discussed
is the vertical conception of time, as in “past is up, future is down.” We
use this metaphor in English only rarely, whereas Chinese, as Jenn-yeu
Chen has shown, uses it for about one-third of the instances in which a
“time is space” metaphor is used. Other examples are metonyms for fa-
mous events, where both times and places can serve the purpose in both
Chinese and English, although there is a marked tendency for Chinese to
prefer times and English to prefer places. For example, in English we “re-
member the Alamo,” but in Chinese we never forget jiu yi ba бϔܿ ‘nine
one eight’ (i.e., September 18, 1931, when Japan invaded China’s North-
east). In American English we have Watergate (and by extension other
“-gates” for other scandals) and can hope that “Afghanistan is not another
Vietnam.” In Chinese, by contrast, we have wusi Ѩಯ ‘five four’—the
May Fourth movement, 1919, and siwu ಯѨ ‘four five’—the April 5, 1976,
demonstration at Tiananmen to mourn Zhou Enlai’s passing. The Beijing
massacre of June 4, 1989, is liusi ݁ಯ ‘six four’ in Chinese and Tiananmen
in English. There are exceptions, though, to this pattern of the prefer-
ences the two languages show. We have “nine eleven” and the War of 1912
in English, and in Chinese lugouqiao श≳ḹ ‘Marco Polo bridge’ can
stand for the Japanese invasion of north China on July 7, 1937. Either kind
of metonym works in either language; the disproportion between the two
is a matter of custom.
A difference by concept, on the other hand, is one in which one language
uses a metaphor that another just does not use. An example is the meta-
phor “an instrument is a companion,” which in English leads to sentences
like “I sliced the salami with a knife” and “she plays Ping-Pong with her
left land.”188 Lakoff and Johnson claim that “with few exceptions . . . in all
languages of the world the word or grammatical device that indicates ac-
companiment also indicates instrumentality.”189 Chinese is defi nitely one of
the “exceptions” to this generalization, and indeed would be an extreme
one, because Chinese words that might be translated “accompany” (such
as gen 䎳 ‘follow, with’; pei 䰾 ‘accompany’; sui 䱣 ‘follow’) are never used
metaphorically to introduce an instrument. In the Chinese mind, to say
something like daozi pei wo qie rou ߔᄤ䰾៥ߛ㙝 ‘the knife accompanies
me in slicing meat’ creates a mood reminiscent of Alice in Wonderland, and
even in that mood the image would be hard to picture. To say “she plays
Ping-Pong with her left hand” as ta gen ta de zuoshou da pingpangqiu ཌྷ䎳
ཌྷⱘᎺ᠟ᠧЦЧ⧗ would ask a Chinese speaker to imagine that her left
hand is at the other end of the Ping-Pong table, playing her as her oppo-
nent. To convey the notion of instrument, Chinese uses a nonmetaphorical
188. Lakoff and Johnson, Metaphors We Live By, pp. 134–135.

189. Ibid., p. 135.
Metaphor 211
yong ⫼ ‘use’: wo yong zuoshou da pingpangqiu ៥⫼Ꮊ᠟ᠧЦЧ⧗ ‘I play

Ping-Pong with [i.e., using] my left hand’. English can, of course, employ
“use” in this way as well; but Chinese never uses the metaphor “an instru-
ment is a companion.”
Another family of metaphors that illustrates different conceptual ap-
proaches in Chinese and English are those that involve “looking at” or
“seeing” as metaphors for “thinking” or “understanding.” In some cases,
these metaphors are similar. In either English or Chinese, for example, if
we say someone is “covering up the facts” ( yan’gai shishi ᥽Ⲫџᅲ) we use
a visual metaphor that suggests that our access to the facts—to knowing
or thinking about them—is blocked. When we say that something is
“blurry” in English or menglong ᳺ㚻 ‘foggy’ in Chinese, we are again using
a metaphor of sight to describe our thinking or understanding. In some
other uses, though, the see conceptual metaphors of Chinese and English
are conceptually different.
Metaphor theorists who work from European languages have noted the
metaphor “understanding is seeing.” We say in English “I see what you
mean,” “I see they’ve changed their plans,” or “the judge just couldn’t see
it my way.” In these cases, “see” essentially does mean “understand.” But in
Chinese, neither kan ⳟ ‘look at’ nor kanjian ⳟ㾕 ‘see’ works in this way:
*wo kan nide yisi ៥ⳟԴⱘᛣᗱ ‘I look at your meaning’ and *wo kanjian
nide yisi ៥ⳟ㽟Դⱘᛣᗱ ‘I see your meaning’ are both clearly wrong.
The difference is systematic: kan and kanjian never work in this kind of
case. But kan does work as a conceptual metaphor for other concepts in
Chinese.
The most basic sense of kan ⳟ is deftly suggested by its Chinese char-
acter, which represents a hand over an eye, an arrangement that human
beings everywhere presumably have found useful in shielding the eyes from
sunlight in order to “look at” something. From the fundamental meaning
of “look at,” kan has extended meanings of watch (kan qiusai ⳟ⧗䌯 ‘watch
a ball game’), read (kan bao ⳟ᡹ ‘read a newspaper’), visit (kan pengyou
ⳟ᳟ট ‘visit friends’), care for (kan [first tone] haizi ⳟᄽᄤ ‘babysit’), “it
depends” (kan ta zenme shuo ⳟҪᗢМ䇈 ‘it depends on what he says’), and
many others. The sense of kan that comes closest to “understand” is “think,”
as in wo kan ta buhui lai ៥ⳟҪϡӮᴹ ‘I think he won’t come’. This sense is
close to the English “as I see it” or “in my view.”
Kan is crucially different from “see” in the literal sense because kan
implies only “look at,” not the successful perception that is implied when
we add jian 㽟 ‘perceive’ or dao ࠄ ‘arrive’. (Kanjian and kandao are “see” in
the literal sense.) Even the “at” in “look at” is a bit of an overinterpretation
of the original sense of kan. One can wangwaikan ᕔ໪ⳟ ‘peer outward’ or
xiangqiankan ৥ࠡⳟ ‘look forward’ without looking at anything. Kan
means only, as it were, “peering from [someone’s] vantage point”; it does
not imply seeing, let alone understanding what is seen, as the see metaphor
does in English.
In Molière’s play Tartuffe, Orgon tries to convince his mother that he
knows Tartuffe has been making advances toward his wife. His mother,
who venerates Tartuffe, cannot believe it, so Orgon says (act 5, scene 3): Je
l’ai vu, dis-je, vu, de mes propres yeux vu ‘I saw it, I tell you, saw, saw with my
own eyes!’. Here vu, similarly to the translation “saw” in English, implies
“know as an absolute certainty.” The line is funny in part because Orgon
did his witnessing while hidden beneath a table, from where he heard a
great deal but saw literally nothing of what he claims by je l’ai vu. Molière
exploits the distance between the literal and metaphorical senses of vu
rather in the manner of a pun. In Chinese, where kan falls well short of
meaning “know for certain” and kanjian is restricted to “see” in a literal
sense, Orgon would seem less funny— and more just a flat-out liar—if he
were to have said wo qinyan kanjian de ៥҆ⴐⳟ㾕ⱘ ‘I saw it with my own
eyes’.
But are these metaphors that “differ by concept” among languages re-
ally very significant for how speakers of the different languages think about
the world? When you understand what someone else has said, in English
you can say “I see what you mean,” using a knowing-is-seeing metaphor,
and in Chinese you might say wo mingbai ni de yisi ៥ᯢⱑԴⱘᛣᗱ ‘I
bright-white your meaning’, using a metaphor of “knowing is bright and
white.” In both cases, the metaphors “differ by concept,” that is, cannot be
imported into the other language: wo kanjian nide yisi in Chinese and “I
bright and white your meaning” in English are both far from acceptable.
But so what, in the end? Are these not just different devices of expression,
more or less like different words? The underlying ideas that get expressed
are practically indistinguishable. The “conceptual difference” seems
inconsequential.
Metaphor 213
In some cases, though, there do appear to be important consequences

for worldviews, including the conscious worldviews that philosophers con-
struct. To illustrate, I would like to examine a case in which the habits of
metaphor in Chinese can show how concepts that are embedded in En-
glish and other European languages may not be as “given” as Western
civilizations have taken them to be.
Here is the example in capsule. English and other European languages
use metaphors of “containers” or “enclosures” more than Chinese does;
European languages also tend, more than Chinese, to label metaphorical
containers, enclosures, and other things with nouns; and Western think-
ers tend, much more than do Chinese thinkers, to worry about what these
“things” named by nouns “are.” It is worth asking how much the philo-
sophical puzzles— and even more, the intensity of worry about them—
might be artifacts of linguistic habit.
In English we say that something is “in” a category or even “falls into”
it. Guppies are “in” the tropical fish category, and swordtails “fall into”
that category, too. In Chinese the simplest way to say the same thing is
kongqueyu shi redaiyu ᄨ䲔剐ᰃ⛁ᏺ剐, which is literally “guppy be tropical
fish.” In sentences like this, the single syllable shi does the work of the
entire English phrase “falls into the category of.” One reason why this
works easily in Chinese is that all Chinese nouns, unless they are pre-
ceded by auxiliary nouns (also called “measure words”), are by nature ab-
stract in the same sense that English words like “water” or “sugar” are
abstract (and can be concretized only by saying things like “a cup of water”
or “a lump of sugar”). To translate kongqueyu shi redaiyu even more literally,
one might gloss it as “guppyhood fall-into-category-of tropical-fish-ness.”
Similarly, to say in Chinese “Zhang San is in the People’s Liberation
Army,” one can easily dispense with in and just say Zhang San shi jie-
fangjun ᓴϝᰃ㾷ᬒ‫ݯ‬, literally “Zhang San be [in the category of] People’s
Liberation Army.” One can put in the notion of in if one wishes, by saying
something like Zhang San zai jiefangjun li ᓴϝ೼㾷ᬒ‫ݯ‬䞠 ‘Zhang San is
in the People’s Liberation Army’, but this is a modern usage that has been
influenced by Western-language grammar. Even in this Western-influenced
mode, however, Chinese does not go so far as to use “fall” as a metaphor for
“belonging to” a category. To say that something belongs to a category in
Chinese, one most naturally says something like jianweiyu shuyu redaiyu
zhi lei ࠥሒ剐ሲѢ⛁ᏺ剐П㉏ ‘swordtails belong to the category of tropi-

cal fish’.
Other things in English that can be expressed using container or enclo-
sure metaphors include relationships. We can say in English that we are
“in” a good relationship. If in a bad one, we might find it “confining.” “Fall-
ing” is sometimes involved as well; we can “fall” in love. These metaphors
occur in Chinese, too, but much less often. Chinese love stories can have
people sinking into enamored states, as in xianyu chiqing 䱋Ѣ⯈ᚙ ‘sink
into silly sentiment’; on the other hand, to fall in love with a person is of-
ten aishang ⠅Ϟ ‘love onto’. In English we can also “enter into” relation-
ships with people as varied as coaches, lawyers, and tax advisers, whereas
in Chinese we tend to zhao ᡒ ‘seek out’ this kind of person, or ting ਀
‘listen to’ them.
Using English examples, Lakoff and Johnson note that objects can come
out of substances, as when we make statues “out of” clay. Similarly sub-
stances can be put into forms, as when we turn clay “into” a statue. Identi-
fying these processes as modes of “causation,” Lakoff and Johnson show
how the metaphor “causation is emergence,” as a metaphor we live by, can
travel to other contexts: he shot the mayor “out of” desperation; he dropped
“from” exhaustion, and so on.190 But here, too, Chinese is clearly different.
We don’t go “into” or “out of” things for such purposes in Chinese. For
making clay into a statue, in Chinese one might most naturally say yong ni
zuo suxiang ⫼⊹‫خ‬ล‫‘ ڣ‬use clay to make statue’ or ba ni zuocheng suxiang
ᡞ⊹‫خ‬៤ล‫ڣ‬, literally “take clay make to form statue”; killing mayors or
falling from exhaustion would likely use a nonmetaphorical yin ಴ or yin-
wei ಴Ў ‘because’.
There is one very common “out of” metaphor in Chinese, but it does
not mean “caused by” as its English cousin does. As a complement to
verbs, chulai ߎᴹ ‘come out’ can signal that the verb results in something
emerging into the cognitive open, as it were, where it is more plain and
perceptible than it was before. If I am wondering whether people on a bus
are speaking Fuzhou or Amoy dialect, I might ask you if you can tingchulai
਀ߎᴹ ‘listen come out—listen and have the answer emerge’. Kanchulai
‘look come out’ can be used when something that was originally at a dis-
190. Ibid., pp. 72–75.

Metaphor 215
tance, or shrouded in fog, comes visually clear; it can also be used when
the visual data is right in front of you but for some reason harbors a ques-
tion whose answer needs to “come out”— as, for example, when you are
looking at a handsome young man but cannot tell by looking whether he
is Korean or Chinese. Xiangchulai ‘think come out’ relies on no percep-
tual data at all; it is entirely cerebral. It is used when I am trying to figure
out where I last saw someone (I cannot “think out come” where it was) or
how I can reasonably explain to my mother that I skipped the geometry
final (I can’t “think” a good excuse “out”). In all these cases there is, im-
plicitly, a metaphorical “place” from which the answer to a (stated or un-
stated) question emerges, or could emerge.
A somewhat similar conceptual metaphor in Chinese is kai ᓔ, literally
“open,” which, when used as a verbal complement, means something like
“free, away, unfettered.” Zoukai 䍄ᓔ is “walk away” (and as a command
can mean “get out of the way!”), where kai suggests release from some
tightness in the present situation. Jiekai 㾷ᓔ, literally “untie open,” can
be used literally for untying a knot or figuratively for solving a puzzle or
getting free of a hang-up ( jiekai geda 㾷ᓔ⭭⯽). Xiangkai ᛇᓔ or kankai
ⳟᓔ ‘think open’ is used for release from a mental burden, as when get-
ting over the death of a loved one.
Broadly speaking, these examples of chulai and kai, both of which sug-
gest a notion of “release into the open,” implicitly resemble some of the
“container” metaphors of English. But English (like other Indo-European
languages) still uses container metaphors much more than Chinese does.
These container metaphors have their advantages, as we will see, but also,
I will argue, might be conceptually confining in ways that Chinese avoids.
Can Conceptual Metaphors Generate Philosophical Problems?
In English, container metaphors are similar in some ways to what Lakoff

and Johnson call “ontological metaphors”: shorthand labels we give to
phenomena whose description in literal detail would involve inordinate
difficulty or tedium.191 To spell out fully what infl ation is would take many
191. Ibid., pp. 25–32.

more than three syllables; but once we “get the idea” of what it is, we settle
for the one-word shorthand; then, because the abbreviation is a noun, we
begin to use it in the ways we use other nouns, treating it as if it were a
“thing” and sometimes even an animate thing: hence “inflation is killing
us at the checkout counter,” “we need to combat inflation,” and so on. As
noted in the Introduction, Lakoff and Johnson go so far as to say that on-
tological metaphors “are necessary for even attempting to deal rationally
with our experiences.”192
Here the contrast with Chinese is instructive and in some ways far-
reaching. Many examples of ontological metaphors in English, including
ones cited by Lakoff and Johnson, are ideas that are most naturally said in
Chinese using verbs, and with no diminution at all in the “rationality of
dealing with our experience.” Lakoff and Johnson cite “my fear of insects is
driving my wife crazy,” where “fear,” which in Chinese is almost always
talked about using verbs, in English seems to be a “thing” that can do
something to something else. If we try to match the English sentence
closely in Chinese and say something like wo zhi pa kunchong ba qizi bi feng
le ៥Пᗩᯚ㰿ᡞྏᄤ䘐⮃њ ‘my fear of insects drives my wife crazy’, we
get an awkward sentence that clearly smacks of borrowing from Western
language. It would be more natural in Chinese to say wo zeme pa kunchong
qizi shoubuliao ៥䖭咑ᗩᯚ㰿ྏᄤফϡњ, literally something like “I so fear
insects that wife can’t take it.” In the end there is no way to say that either
the verb-heavy Chinese sentence or the noun-studded English one is a
“more rational handling of life.” One uses ontological metaphor, and the
other does not; at least in this case, the use of ontological metaphor is a
cultural choice, not a necessity.
Another of Lakoff and Johnson’s examples is “it will take a lot of patience
to finish this book.” Here, again, experience that could well be expressed
with verbs—is made into a noun, as if patience were “stuff.” A Chinese
speaker, encountering the same boring book, would be unlikely to think
of patience as stuff. A natural, idiomatic way to respond in Chinese would
be to say zhei ben shu burenzudu 䖭ᴀкϡᖡद䇏 ‘this book (one) cannot
bear to finish reading’, in which ren ‘tolerate’, zu ‘finish’, and du ‘read’ are
all verbs. A Chinese sentence somewhat closer to the English—but still
192. Ibid., p. 26.

Metaphor 217
not borrowing Western grammar—would be zhei ben shu duqilai hen feijin
䖭ᴀк䇏䍋ᴹᕜ䌍ࢆ. It is a telling fact that this Chinese sentence is very
hard to put into English without reverting to an ontological metaphor,
even though there is no ontological metaphor in the Chinese. Of the six
syllables in the predicate, duqilai hen feijin, only one, jin ‘energy’, is a noun.
Four are verbs and one is an adverb. A superliteral translation would be
“[this book] read rise come very spend energy.” In more readable English,
but still trying to preserve the spirit of the Chinese, one might say “it
takes a lot of energy [or takes a lot out of you] to read this book.” But note
what happens when we do this. The metaphor of energy as “stuff”—that
gets used up, or gets taken out of you—has subtly crept back in. English
likes to do this.
The divergent preferences of the two languages in this regard are suffi-
ciently strong that it can seem, at times, that some things that are sayable in
English just cannot be said in Chinese. “Her feelings of frustration over-
came her,” for example, is hard to put into Chinese without resorting to
phraseology so Westernized that it hurts the ears. But Westerners have no
monopoly on the experience of frustration, of course, and it is easy to find
ways to talk about frustration in Chinese—just not as a “thing” or “stuff”.
English phrases that use container metaphors are especially hard to
translate into Chinese. So long as container metaphors are not involved,
we can say that something “disappears” in English and render it as shi-
zongle ༅䏾њ ‘has lost traces’ in Chinese; we can also say in English that
something “is no longer there” and say buzaile ϡ೼њ ‘no longer is [some-
where]’ in Chinese. But major problems arise when the container meta-
phors of English come along. For essentially the same idea as “disappear,”
we can say in English “go out of existence”— as if one thing “exits” an-
other. We can also say in English that a thing “comes into existence” or
“is in existence.” Chinese balks at all of this. Just by itself, the English
noun “existence” translates into Chinese only awkwardly. A direct trans-
lation of the phrase “is in existence” as zai cunzaili ೼ᄬ೼䞠 sounds al-
most cretinous in Chinese, in part because “is” and “exist” (zai and cunzai)
somehow seem redundant, but even more because, obviously, no “con-
tainer” is needed here.
It is important to note that the nominalization English prefers is a
matter of choice, not (as Lakoff and Johnson seem to feel) necessity. The
poet Wang Wei ⥟㎁ (699–761), when he describes moonlight with the

line yuese youwuzhong ᳜㡆᳝⛵Ё ‘beauty of moon between there and not
there’, achieves a superb elegance. An English speaker viewing the same
moon phenomenon that Wang Wei beheld might well have said that the
moon was moving into and out of something—plain sight, the clouds, or
even, poetically, “existence.” There should be no doubt that a description
can be effective either way, with or without container metaphors.
In saying that “either way” is viable, we must guard against supposing
that one way is always the “Chinese” way and the other always the “West-
ern” one. English can say “is no longer” as easily as it can say “out of exis-
tence,” and Chinese sometimes uses ontological and container metaphors.
They can be found even in the lively, natural language of comedians’ dia-
logues. Wang Guoxiang’s 1955 piece Hu, ji, dan ໎䲲㲟 (Pot, chicken,
egg) uses both a container metaphor and an ontological metaphor when
it says zhei sange “ji” litou . . . 䗭ϝ‫ן‬ÿᗹÿ㺣丁 ‘inside [i.e., among] these
three “urgencies” . . .’.193 Ning Yu cites many ontological and container
metaphors in Chinese, including examples like wo you xinxin ៥ֵ᳝ᖗ ‘I
have confidence’ (where “confidence” is an ontological metaphor) and
guoyou qiye chuyu lianghao zhuangtai ೑᳝ӕϮ໘Ѣ㡃ད⢊ᗕ ‘state enter-
prises are [located] in good condition’ (where a container metaphor is
at work).194 One could object that these examples are obviously based on
borrowings from Western grammar, and that is true; but that does not
mean that such examples, by now, are not authentic Chinese. By the early
twenty-first century, usages such as these had become thoroughly natu-
ral, especially among younger generations of native speakers. Consider
the English “inflation is killing us.” Set the ontological metaphor aside
and think just of the meaning. Most Chinese speakers, wanting to express
such a thought, would probably say something like dongxi gui de yaoming le
ϰ㽓䌉ᕫ㽕ੑњ ‘things are getting unbearably [literally, “life-demandingly”]
expensive’. A literal mirroring of the English—tonghuo pengzhang zheng-
zai sha women 䗮䋻㝼㚔ℷ೼ᴔ៥Ӏ ‘inflation right now is killing us’—
would seem very awkward. But something in between— a hybrid like
tonghuopengzhang yuelai yue lihai 䗮䋻㝼㚔䍞ᴹ䍞ढ़ᆇ ‘inflation is getting
193. In Shuoshuochangchang 䁾䂀ଅଅ [Telling and singing], no. 3 (1955), p. 46.

194. Ning Yu, Contemporary Theory of Metaphor, pp. 152, 216.
Metaphor 219
more and more severe’—by the early twenty-first century, sounds quite
normal.
Does English really tend to prefer nouns and Chinese tend to prefer
verbs? Can we put this to an objective test? As an experiment, I chose at
random a page from each of two classic works of fiction, Charles Dick-
ens’s Oliver Twist and Cao Xueqin’s Dream of the Red Chamber, and counted
the nouns and verbs on the pages.195 I chose Dream of the Red Chamber in
part because I wanted a Chinese sample that would be free from Western-
language influence. Some aspects of the experiment required arbitrary
judgment: for example, should a Chinese cliché like xiaodao ュ䘧 ‘said
with a smile’ count as one verb or two? Should “it” in an English phrase
like “so it is” count as a noun? But such problems were relatively few, and
in the end this kind of borderline question seemed insignificant because
the overall result was very clear: there were 96 nouns and 38 verbs on one
page of Dickens (a 2.5:1 ratio) and 130 nouns and 166 verbs on one page of
Cao Xueqin (a 0.8:1 ratio). The experimental results did confirm what I
thought they might.
If English grammar tends to give borders (metaphorically, the edges of
containers) to “things” like existence and frustration, while Chinese is
more fluid about these matters, it is worth noting that the same tendency
seems to hold for the very grammatical categories of the two languages.
In any language, terms can move from one part of speech to another. In
English we can read a book and we can book a room; the freeway admits
us at the rate of “one car per green”; and so on. But such fluidity across
grammatical categories is demonstrably more common in Chinese than
in English. I cited earlier the example of ji ᗹ used as a noun and trans-
lated it as “urgency.” But ji much more commonly is a verb, as in xian bie ji
‫߹ܜ‬ᗹ ‘fi rst don’t excite’ (“don’t get excited before you have to”) or is an
adjective (sometimes called a “stative verb”), as in xinli hen ji ᖗ䞠ᕜᗹ ‘in
the heart excited— anxious’. I also cited dao 䘧 ‘speak’ as a verb, but dao is at
least as famous as a noun meaning road or way (or exalted Way). An obvi-
ous reason why Chinese is more fluid in this regard is that morphological
195. I used Oliver Twist (New York: Barnes and Noble Classics, 2003), p. 64, and
Hongloumeng (Hong Kong: Youlian chubanshe, 1960), p. 487. The page selection was
utterly at random.
change occurs much less in Chinese than in most other languages. Fazhan
থሩ is “develop,” as in fazhan nide caineng থሩԴⱘᠡ㛑 ‘develop your
talents’, but also “development,” as in zuijin de fazhan ᳔䖥ⱘথሩ ‘recent
developments’, and “developed,” as in fazhan guojia থሩ೑ᆊ ‘developed
countries’ (although the latter is more common as fada guojia থ䖒೑ᆊ). If
we add zhong Ё ‘within’ to the latter phrase, to get fazhanzhong de guojia
থሩЁⱘ೑ᆊ ‘developing countries’, the zhong might be viewed as a mor-
phological change, but even if so, it is not as clear an example of such
change as is – ing in the English word “developing.”
Like Chinese verbs, Chinese adjectives (or stative verbs) also cross the
borders of categories more easily than their English counterparts do.
Chinese adjectives, when used alone, are always implicitly comparative
with something else. I do not mean that they are “implicitly comparative”
with a general reference group of the kind we need in order to interpret
“Mary is tall” depending on whether Mary is a fourth-grader or an adult
volleyball player, which, as Searle has pointed out, is a kind of implicit
comparison necessary in any human language. I mean that, for example,
gao 催 in Chinese, when used alone, does not mean “tall” but “taller,” and
not taller than things of its kind in general but taller than some specific
other thing. Zhang San gao ᓴϝ催 does not mean “Zhang San is tall”; it
means “Zhang San is taller (than someone or something else).” It is the
answer to the question Zhang San gao haishi Li Si gao? ᓴϝ催䖬ᰃᴢಯ催
‘who is taller, Zhang San or Li Si?’. In order to say “Zhang San is tall” in
general—parallel to the way we say “Mary is tall” in English— one needs
to insert a relatively empty adverb before gao. Hen ᕜ ‘very’ is the most
common of these, although (in some northern dialects) ting ᤎ ‘rather’
and (in some southern ones) man 㸏 ‘quite’ perform the function just as
well: Zhang San hen (ting, man) gao is how one says “Zhang San is tall
(compared to some appropriate reference group).” My point here is that
the fundamental conception of Chinese adjectives leaves them intrinsi-
cally less static and entified than English adjectives.
A comparable “border-crossing” is visible in the conception of Chinese
verbs. It is instructive to ask why the English word “try”— and compara-
ble words in other languages— are hard to translate into Chinese. Today
Chinese uses shi 䆩, but this is a modern borrowing. Was it perhaps the
case that premodern Chinese people “tried” things less than other peoples?
Metaphor 221
Of course not. The difference is that the Chinese language handles the
description of “trying” in a different way. The idea of “try” is built into
most Chinese verbs. For example shui ⴵ ‘sleep’ is not, precisely, “to sleep.”
It is something more like “head for sleep,” “set about sleeping,” or “try to
sleep.” When a person actually falls asleep (i.e., when the attempt ends
and the real experience begins), Chinese adds a complement to the verb,
shui zhao ⴵⴔ, to signal that the shui effort “takes effect.” Chinese uses
this kind of verbal complement with great frequency, and English hardly
at all. (I noted in Chapter 1 that “tickled pink” and “scared stiff” are ex-
amples in English of the kind of thing that verbal complements in Chi-
nese do.) What English achieves by putting “try” before a verb is often
the same as what Chinese achieves by using a verb and omitting any
complement. When you try something in Chinese, you set about the ac-
tion of the verb; if it works, you tack on a complement to indicate that it
worked.196 James Tai and Jane Chou have raised the telling example of
how to say kill in Chinese. Chinese-English dictionaries list sha ↎ as
‘kill’, but this is not as precise as it perhaps should be. “Try to kill” might
be better. In Chinese one can say Zhang San shale Li Si liang ci dou mei ba
ta sha si ᓴϝᴔњᴢಯϸ⃵䛑≵ᡞҪᴔ⅏ ‘Zhang San tried twice to kill Li
Si but didn’t kill him’. In awkward English that better reflects the struc-
ture of the Chinese, one might say “Zhang San twice set about the killing
of Li Si and neither time killed him into deadness.” Similarly zisha 㞾ᴔ,
which dictionaries list as “commit suicide,” might be more precisely listed
as “attempt suicide.” One can zisha in Chinese many times but commit
suicide in English only once.
But setting aside, for now, the question of how much it might be true
that Chinese verbs and adjectives “cross borders” more easily than their
English counterparts do, I want to focus especially on the case of nouns,
where the Western-language preference for putting things into contain-
ers has, it seems to me, some fairly profound implications.
196. Many complements add the additional flavor of how something worked. Verb-
wan ᅠ says that you fi nished, -hao ད that you fi nished and the result was fi ne, - cheng
៤ that you formed something into something else, -buliao ϡњ that you couldn’t get
to the end of the task, and there are many other examples. They all say what happened
after you tried.
We sometimes insert nouns into English phrases for no useful purpose

at all. Why, for example, do many English speakers say “on a daily basis”
instead of “every day,” or just “daily”? Put directly into Chinese, the “ba-
sis” idea sounds stupid. Wo tiantian shang ke ៥໽໽Ϟ䂆 ‘I go to class every
day’ sounds fi ne, but wo zai tiantian de jichushang shang ke ៥೼໽໽ⱘ෎⻢
ϞϞ䂆, literally, “I go to class on a daily basis,” the zai . . . jichushang is
superfluous and a bit ridiculous. The same kind of superfluousness can
invade one-word ontological metaphors in English. For example, in the
late 1990s it became fashionable in literary study in the West to speak not
just of “position” but of “positionality.” Both words are nouns, of course,
but the -ality was assumed to be adding something— something new, styl-
ish, even a bit mystical. Students from China could master the word in
English, but how could they put it into Chinese? If “position” is weizhi
ԡ㕂, then what is “positionality”? Weizhixing ԡ㕂ᗻ seems inane, but to
leave off xing ᗻ is to lose the -ality, and then position is only position.
How often, in English, do we conceive things in terms of nouns when it
is not necessary to do so? For example, when electric impulses are speed-
ing along neurons in the brain, might not a verb be best? Why does a
distinguished cognitive scientist, George Lakoff, create a noun, “neural
connectivity,” and then conceive it as a thing that can act on other things,
for example, “makes it natural for complex metaphorical mappings to be
built”?197 Elsewhere, in writing about conceptual categories, Lakoff ex-
plains that certain examples of a category (e.g., robins and sparrows, for the
category of birds) are its “prototypes” and, through application of rules in-
volving similarity and other criteria, can be “generators” of extension of the
category to include emus, ostriches, penguins, and other birds. So far, the
nouns in this explanation are probably all necessary. But then Lakoff also
uses—as a writer in Chinese would not, indeed could not, without severe
awkwardness—the abstract noun “prototypicality,” which in this case al-
lows something else, yet another abstract noun, “generativity,” to occur.198
Similarly, Amanda Scott, whose work on vertical space-for-time metaphors
I have already referred to, quotes R. H. Lauer as observing not that time is

198. Lakoff, Women, Fire, and Dangerous Things, pp. 12, 41, 44.
Metaphor 223
a fact but that “temporality is a facticity.”199 A medical researcher at the

University of California at San Francisco in 2003 discussed mad cow dis-
ease in terms of its “infectivity.”200 It is easy to find similar examples, espe-
cially in the world of Western academe.
What is not so easy to say is whether this bent for nominalization does
harm or good in the world of thought. Derk Bodde and others have ar-
gued that Indo-European languages, because they deal easily in abstract
nouns, make abstract thinking itself more natural than it is in languages
like Chinese, which see the world, as Bodde puts it, more “organically.”
This difference in languages helps, in Bodde’s view, to explain why the
scientific method originated among speakers of Eu ropean languages.201
Bodde may be right. Certainly it is a mistake to discount his view (as some
have) because of its alleged disrespect for Chinese culture. What I want
to do, though, is to suggest a hypothesis that leans in the opposite direc-
tion. It does not contradict Bodde’s view, and both could well be true. My
hypothesis, put as a question, is this: what kinds of problems might be
caused by excessive nominalization in Indo-European languages? Might
such problems be less troublesome when one interprets the world through
verb-heavy Chinese?
Where excessive nominalization can begin to cause trouble, in my view,
is the point at which a person who thinks in English begins to assume that
a noun somehow says something more “real” than a verb or adjective
does. “The neurons connect well” and “the neural connectivity is good”
say essentially the same thing. (Good writers might prefer the former as a
matter of style, but that is a separate question.) Problems enter when
people begin to suppose that “neural connectivity” somehow is something
special, something that adds to “neurons connect.” Such a supposition can
lead to several problems. One very simple problem is that we can be mis-
led into thinking that a mere tautology is intellectually significant. If I
199. Amanda Scott, “Vertical Dimension of Time in Mandarin,” p. 312; R. H.
Lauer, Temporal Man (New York: Praeger, 1981), p. 26.
200. Joanne Silberner, “Infected Cow Born before Feed Ban Took Effect,” National
Public Radio, December 30, 2003.
201. Derk Bodde, Chinese Thought, Society, and Science: The Intellectual and Social
Background of Science and Technology in Pre- modern China (Honolulu: University of
Hawaii Press, 1991), ch. 7 and p. 357.
were to say, for example, “her neurons connect well because she has good
neural connectivity,” the emptiness of the explanation would be plain. Yet
even an experienced writer like Lakoff comes close to doing this when he
writes, for example, “What gives human beings the power of abstract
reason? Our answer is that human beings have what we will call a concep-
tualizing capacity.” 202 Lakoff goes on to explain what he means by “capac-
ity,” and his explanation, like the thing explained, is heavy with nouns.
He writes that the conceptualizing capacity is an “ability to form symbolic
structures . . . in our everyday experience.”203 My point here is not to criti-
cize Lakoff’s idea; it is to note that his thought, as he has expressed it,
would be very hard to put into Chinese without completely reconceiving
it (and then, if one did completely reconceive it, the question would arise
as to whether or not it is the “same” thought). In Lakoff’s English sen-
tence, people reason abstractly because they “have” something (an ability,
a capacity, etc.); in Chinese it is much easier to say that people reason ab-
stractly because they “do” something. Chinese clearly prefers verbs to say
this kind of thing: fenxi ߚᵤ is the modern word for “analyze”; xiang tong
ᛇ䗮, literally “think through,” and xiang mingbai ᛇᯢⱑ, literally “think
clear,” are more deeply rooted vernacular expressions; gao qingchu ᧲⏙Ἦ
‘get clear’ is a twentieth-century colloquialism that grew out of the Com-
munist movement. But whatever their provenance, all of them are verbs.
Among them only fenxi can be used as a noun, and as such is limited to
modern, Western-influenced language where it corresponds to the noun
“analysis” in English. There is no way in Chinese, without seeming bi-
zarre, to push fenxi further, toward a word like “analyticity.”
The bent of Western languages toward nominalization might create
some problems that run deeper than the relatively minor danger of re-
dundant explanation, but these deeper problems are harder both to defi ne
and to solve. They arise mainly from an unexamined assumption that
“what nouns refer to” are “things” that might or might not “exist.” Most
people, most of the time (here I am excluding only solipsists, certain mys-
tics, and a few others), feel comfortable with the notion that nouns like
“book,” “chair,” and “snowflake” refer to things that exist. A noun like
202. Lakoff, Women, Fire, and Dangerous Things, p. 280. Emphasis in original.
203. Ibid., p. 281. My emphasis.
Metaphor 225
“unicorn” does not refer to a thing that exists, but it does refer to the kind
of thing that could exist in the way that other referents of nouns do. Hence
it is natural for speakers of Western languages, who have this kind of
mental habit about nouns in their minds, to assume that nouns like “infla-
tion” and “connectivity” also refer to things that do or might “exist.” For
example, when Lakoff refers to “analyticity” as a “phenomenon” that we
can “make sense of,”204 how much are we led to think in terms similar to
“chair” as a “thing” that we can “put in the corner”? Because ontological
metaphor works so smoothly in Western language, we normally do not ask
a question like this.
Lakoff comes close to addressing this problem when he writes: “cul-
tural categories are real and they are made real by human action. Govern-
ments are real. They exist.”205 There are two questions here. One is why we
use the noun “action” to define government. Could we use the verb “acts”
instead? Government collects taxes, fi nes you if you run a red light, runs
public schools, protects you if the Canadians invade, and so on. These are
things that it does. (Such a list would be indefi nitely long.) We can list the
actions as nouns or— as I just have— as verbs. Either works. The second
important question is why, if we do decide to use nouns instead of verbs,
we choose to say that abstract nouns—like “government”— are “real” and
that they “exist.” It is abstract to say that a government “does” things,
because humans are the actual doers. Lakoff suggests as much by refer-
ring to “human” action. He probably believes that the human actions that
comprise government are themselves “real,” but he does not say this; his
point is that human actions are what turn governments into real things.
When this happens, governments “exist.” The question that needs to be
raised— and that I believe, broadly speaking, Chinese grammar prods us
to raise—is to what extent English speakers are induced by the habits of
their language to jump too easily from the level of “chairs exist” to the level
of “governments exist,” and then to conceive the two things as more simi-
lar than they are.
Does it happen, for example, that speakers of English are drawn to be-
lieve that certain things exist because nouns that claim to be their labels
204. Ibid., p. 118.

205. Ibid., p. 208.
exist? Might it be that only the labels exist? For example, Hoyt Alverson,
the anthropologist I cited earlier for his argument that crosscultural simi-
larities in the experience of time suggest a “universal template of human
experience,” describes the relativist position on this question, which Al-
verson opposes, as claiming that the “ontogeny” of time is indeterminate.206
He explains “ontogeny” as meaning the “character” of something’s “be-
ing.” We have, then, the proposition that the character of the being of time
is indeterminate. Do the nouns in this proposition refer to things that
exist? In addition to time, is there a “being” of time? And if there is, is
that being the kind of thing that can possess something else, as here it
is supposed to possess a “character”? These problems are by no means
Alverson’s alone; he writes in a mode that is fairly common in English. In
Chinese, though, it is almost impossibly awkward to refer to “the charac-
ter of the being of time.” A noun-studded literal translation like shijian de
cunzai de xingzhi ᯊ䯈ⱘᄬ೼ⱘᗻ䋼 ‘nature of existence of time’ is opaque
in Chinese; to Chinese ears it signals that “this came out of a Western
language and you might well go there to figure out what it is supposed to
mean.”
As noted earlier, the very word “being” (or “existence”) is hard to put
into Chinese. Cunzai is a modern term, and by the late twentieth century
it had entered Chinese sufficiently to sound natural as a verb (although it
remained awkward as a noun). Ancient Chinese thinkers who were con-
cerned with being and nonbeing often used you ᳝ ‘there is’ and wu ⛵
‘there is not’ to express these concerns. But you and wu are both verbs.
Westerners have often translated them using nouns like “being” and
“nonbeing,” but within Chinese grammar they do not feel natural as
nouns. This grammatical difference makes it hard, in Chinese, to pose
philosophical questions about what being “is” or what “character” it might
“have.” From here, one can speculate on the inadequacy of the Chinese
language, because certain questions cannot be raised in it. Or, to turn
the matter around, one can speculate on the flaws of Western language,
in which grammar permits one to raise (and waste time on) questions
that are not real questions. Are English speakers better off because they
can think about the “properties” of “being,” or are Chinese speakers
206. Hoyt Alverson, Semantics and Experience, p. 3.

Metaphor 227
better off because they are spared this word-trap? In short, the issue of
a preference for nouns in Eu ropean languages and a preference for verbs
in Chinese might have implications for formal philosophy in the two
traditions.
Plato puzzled over “the good,” “beauty,” “justice,” “substance,” and
other noun-named things some of which can be discussed just as easily as
adjectives or verbs. Later Western philosophers have spent time on “mind,”
“autonomy,” and “free will,” while religious thinkers have pondered
“God” and “man,” “sin,” and the “immortality” of the “soul.” It is possible
to overdraw the differences between Chinese and Western approaches in
such matters, but there does seem to be a systematic difference between
the two traditions, and others have noticed it. Standing back from his
detailed study of pre-Qin thought, Edward Slingerland generalizes that
the knowledge that ancient Chinese thinkers pursued “was not abstract
knowledge that the good was to be defined in a certain way, but concrete
knowledge concerning how to act in a way that was good.”207 Roger Ames
and Henry Rosemont, generalizing over a broader range of Chinese
thought, write that “we want to claim that English (and other Indo-
European languages) is basically substantive and essentialistic, whereas clas-
sical Chinese should be seen more as an eventful language.”208 Ames and
Rosemont feel that the distinction between Chinese and English holds
even in modern language. They compare the sentence “The young
woman who just entered the room is very bright” with a translation—
gangcai dao wuzili lai de xiaojie feichang congming ࠯ᠡࠄሟᄤ㺣՚ⱘᇣྤ䴲
ᐌ㙄ᯢ. Then they note— correctly, in my view—that somehow the young
woman in the English sentence feels more “substantial” and the xiaojie
‘young lady’ in the Chinese sentence feels more “dynamic” than the coun-
terpart in the other sentence.209
207. Slingerland, Effortless Action, p. 3. Emphasis in original.

208. Ames and Rosemont, Analects of Confucius, p. 20.
209. Ibid., pp. 22–23. I have inserted dao ࠄ into the Chinese example to make it a bit
more natural. Readers who do not know Chinese and wonder why the Chinese seems
more “dynamic” and the English more “substantive” might wish that I give a transla-
tion to make this clear. I apologize that I cannot improve on the translation that Ames
and Rosemont provide, but I do agree with their judgment about the subtle difference
in the effects the two languages create.
It would be unfair to Western philosophers to hold that their perhaps

excessive focus on nouns is something that they have not noticed and that
only a language like Chinese can help them to escape. Western philoso-
phy since Plato has not lacked for critics of Plato’s conception of Forms or
his claiming that Forms somehow exist separately from the rest of the
world.210 Moreover, in the twentieth century, and quite without any as-
sistance from Chinese, Ludwig Wittgenstein and others did much to
show how philosophical puzzles can be born from the unexamined ways
people use language. But even at the end of the twentieth century, and
well after Western philosophers had had a chance to absorb Wittgenstein,
the danger of creating nouns and then assuming that “things” correspond
to them seems to have persisted among Western phi losophers. In 1999
Colin McGinn published his little book The Mysterious Flame: Conscious
Minds in a Material World.211 McGinn owns a daunting intellect, and his
writing is clear and incisive. In choosing his book in order to see whether
the “noun trap” was still at work in Western thinking at the end of the
twentieth century, I am purposefully choosing a powerful example for the
test: Is even a writer as astute as McGinn vulnerable to the peril?
McGinn’s book addresses the classic “mind-body problem”: what is
consciousness, and how does it exist in the physical world? At one point he
focuses on the curious fact that our perceptions of the world are often per-
ceptions of things in space, yet the perceptions themselves occupy no space.
He writes:
We need to make a distinction between the object of awareness and

the awareness itself. When I sit in [a] ski lift and feel fear about the
distance between me and the ground, the object of my fear is a spatial
fact: my distance from the earth. It is not that the fear itself is a spa-
tial thing—it is not a hundred feet in length!212
210. Peter Abelard (1079–1142), William of Ockham (1288–1347), Thomas Hobbes

(1588–1679), and Jeremy Bentham (1748–1832), among others, all polemicized in one
way or another against the notion of assuming that every noun is a name for a real
thing.
211. Colin McGinn, The Mysterious Flame (New York: Basic Books, 1999).
212. Ibid., p. 109.
Metaphor 229
Certainly the distance between a suspended chair on a ski lift and the
ground below is different from a fear of falling. But why do we use nouns
for both, and then believe that we are dealing with two parallel “things”?
It is fairly clear that McGinn does this, that is, takes the distance and the
perception of the distance as two somethings. He writes, “My fear has
space as its object, but that which has this object—the mental state of fear
itself—is not to be confused with that object.”213
In Chinese pa ᗩ ‘fear’ and haipa ᆇᗩ ‘fear’ are both verbs. They can be
transitive verbs, as in pa she ᗩ㲛 ‘fear snakes’ or “stative verbs,” as in hen
haipa ᕜᆇᗩ ‘be very much afraid’. But neither can be a noun— except in
modern (and notably awkward) imitation of Western usage. Kongju ᘤ័
‘dread’ works better as a noun, but it, too, is a modern term. In short,
while both nouns and verbs are possible in Chinese, when I am danger-
ously suspended in a ski lift it is much more natural in Chinese to think “I
fear” than to think “I have a fear.” But right there, perhaps, is the begin-
ning of the problem. If I think “I have a fear,” then it makes sense, at least
grammatically, for me to raise the question why that fear is “not a spatial
thing” whereas the distance to the ground is a spatial thing, and this can
become a puzzle. On the other hand, if I think “I fear,” then no question
arises of why one “thing” has spatiality and the other one does not. Again
we need to ask: is it a loss or a gain that this question does not arise? Is
Chinese deficient for closing off an avenue of analysis, or is English defec-
tive for allowing its grammar to entangle us in a nonquestion? The ques-
tion cannot be decided on the criterion Lakoff and Johnson have suggested,
that is, what we need “in order to deal rationally with our experience.”
Haipa ‘fearing’ in Chinese and “feeling a fear” in English are both unobjec-
tionably rational ways to deal with the distance between a ski lift and the
ground.
McGinn gives another example of how physical things occupy space
and mental things do not:
Consider the visual experience of seeing a red sphere two feet away
with a six-inch diameter. The object of this experience is of course a
spatial object with spatial properties, but the experience itself does
213. Ibid.
not have these properties: it is not two feet away from you and six
inches in diameter. . . . When we reflect on the experience itself, we
can see that it lacks spatial properties altogether. 214
The key phrase here is “the experience itself.” Is there such a thing? The
noun “experience” exists, but that is not the question. Does the experience
exist? We might feel intuitively that it does. But does this intuition arise,
in part, from the grammatical habit of using nouns like “experience” and
assuming that they refer to things? Is there a way we can test whether our
intuitions indeed are being shaped by nouns?
The English word “experience” is perhaps not the best example for doing
such a test, simply because it has the same form as both noun and verb.
“Feeling” might work better, because the noun (“feeling”) and the verb
(“feel”) have different forms. In most cases, two statements of the forms “I
feel X” and “I have a feeling of X” will not differ much, if at all, in mean-
ing. But now consider this: If I say “I feel X,” you cannot grammatically
ask me in English “Does your feel have spatial properties?” You could ask,
“Do you feel with (or in) length and color?” but this question, although
grammatical, does not make sense. No matter how you put them, ques-
tions about the spatiality of X are hard to phrase if you use the word “feel”
instead of the word “feeling.” But if, on the other hand, I say “I have a
feeling of X,” then the same question—“Does your feeling have spatial
properties?”—now does make sense. It not only makes grammatical sense,
but makes enough philosophical sense to get into the writing of an excel-
lent philosopher like Colin McGinn. So we can see here that from a start-
ing point where there is no real difference in daily-life usage (i.e., between
“I feel X” and “I have a feeling of X”), the choice of which to use can lead
to (or perhaps generate?) a great philosophical puzzle if one goes in one
direction and lead to no puzzlement if one goes in the other.
McGinn goes on to point out that numbers, like the experience of red
spots, do not occupy space. “We cannot sensibly ask how much space the
number 2 takes up relative to the number 37,” he writes. “It is hardly true
that the bigger the number the more space it occupies.”215 Then he writes:
214. Ibid.
215. Ibid., p. 110.
Metaphor 231
To attribute spatial properties to numbers is an instance of what phi-

losophers call a category-mistake, trying to talk about something as
if it belonged to a category it does not belong to. Only concrete
things have spatial properties, not abstract things like numbers or
mental things like experiences of red.216
In my imagination, a pre-Qin Chinese philosopher might well accept this

point but then ask McGinn: Why do you experience life as “abstract
things”? Is that not also a category-mistake? If I see a red spot, do I not
simply see a red spot? The red spot, yes, is a thing, but “I see” is not a thing,
either concrete or abstract. I see is I see. If you change it into “my sight” or
“my experience of seeing,” you are performing a grammatical act, but that
grammatical act has no power to change the way the world is.
The Significance of Similarities and Differences among

Conceptual Metaphors in Different Languages
I began my study of how conceptual metaphor compares in Chinese and

English with a hope that the results might reveal alternative “worldviews”
in some sort of systematic way. That hope has been largely dashed. Not
only do the two languages not stand apart very clearly in their overall pat-
terns of metaphors but— and here is a much more fundamental problem—
even within just one of the languages, Chinese or English, there is a lot of
metaphorical incoherence. In space-for-time metaphors, as we have seen,
“before” in English can mean either past or future; and so, in Chinese,
can qian ࠡ. Examples such as these— and there are many— show how
conceptual metaphors in a language do not have to fit together coherently.
Less like auto parts that form a whole, they are more like tools in a tool
bag that a person takes out, uses for a purpose, and then puts back—with
no thought of how they fit together or form anything larger. When we say
houdai ৢҷ ‘future generations’, where hou points to the future, and also
say qiantu ࠡ䗨 ‘future path’, where qian means future, it does not matter
at all that qian and hou are opposites. The tools in the tool bag work, each
216. Ibid.
in its own way. So much for my imaginings of grand conclusions about

worldviews.
But then a different sort of grand conclusion, one I did not expect, be-
gan to emerge. I noticed that the separate tool bags of Chinese and En-
glish conceptual metaphors contained some remarkably similar items.
Some of these were modern borrowings, which were easy to explain and
not very interesting. To recognize that pingjing ⫊乜 is “bottleneck” does
not reveal anything very deep. More interesting were examples that are
deeply and separately embedded in the two languages yet turn out to be
almost uncanny in their similarity. The three different ways Chinese
conceives space-for-time, analyzed earlier, are not simple patterns, either
in themselves or in how they fit together. The fact that they match up so
well with the ways English conceives space-for-time metaphors has been
surprising to me, indeed astonishing. So have other similarities, such as
the patterns of privilege within dyads.
Why do human languages as different as Chinese and English turn out
to be so very similar in certain ways? As we have seen, there have been,
broadly speaking, two kinds of explanation for the similarities: one is that
the structure of the human brain and its perceptual apparatus determines
that these things be as they are, and the other is that the commonality of
human experience makes it natural for human beings everywhere to develop
similar concepts . Both sorts of explanation seem clearly right in part, even
though much remains to be understood about most of the details. What-
ever their explanations, the commonalities have turned out to be much
stronger than I anticipated, and there is something very refreshing in that
fact. It seems to reinforce the valuable notion of a single human family.
In cases where conceptual metaphors differ between Chinese and En-
glish, there may be no overarching conclusions to draw yet a gardenful of
interesting insights on particular questions. Cultural differences are al-
ways fun to notice, and sometimes they are enlightening. I have noted, for
example, how accompaniment is instrumentality in many languages; in En-
glish, you can play Ping-Pong “with” your left hand, but this does not
work in Chinese, where, if you say such a thing, it sounds as if your left
hand has gone to the other end of the table to play against you.
There is often an even larger benefit in noticing cultural differences,
however, and that is in realizing that aspects of one’s culture that one has
Metaphor 233
taken for granted are in fact arbitrary—because things can be conceived

differently. In the modern West, north is so well associated with up (not
only on maps but in words like “uptown” and even phrases like “north of
thirty” to mean “over thirty years old”) that it can be interesting, indeed
can “turn the world upside down,” to think of south as “up.” But this is
how it was in China until about a thousand years ago (and again, not only
on maps but in words like zhinan ᣛफ ‘pointing south’ to mean “compass”
or to talk about a person’s moral bearings). My suggestion that excessive
use of ontological metaphor in Indo-European language might help to
explain puzzlement over the mind-body problem in Western philosophy
draws on essentially this same point. It is possible here that habits of con-
ceptual metaphor in one’s native language might prevent one from seeing
that if one did not have such habits, things might look different. I doubt
that insight of this variety will ever “solve” the mind-body problem. But at a
minimum, it can help to explain the extraordinary amount of concern
that has gone into centuries of dealing with the problem in Western lan-
guages. In Chinese the problem does not present itself in a vexing way. It
is therefore not as intellectually stimulating, in Chinese, to get into the
ring and wrestle with it. And that very fact may undo some of the original
problem.
3
Politics
Formal political language, in many societies and in many times, has tended
to diverge from ordinary talk. Vocabulary can differ, rhythms and tones
of voice can diverge, and even grammar can be affected. In most cases,
what creeps into political language is officiousness. A stuffy tone can claim
a special authority for the speaker, who then can assume a position of el-
evation above an audience. This makes glibness easier to achieve and can
provide a slick suit of clothes for questionable or even groundless claims.
George Orwell has written that “political language . . . is designed to make
lies sound truthful and murder respectable, and to give an appearance of
solidity to pure wind.”1 Orwell was writing about English, but his principle
applies broadly.
1. George Orwell, “Politics and the English Language,” in A Collection of Essays

(Garden City, N.Y.: Doubleday, 1954), p. 177.
Politics 235
A Bifurcation
Officialese has a long pedigree in China. In late imperial times, the lingua
franca of the empire was called guanhua ᅬ䁅 ‘official talk’. It was the lan-
guage officials were obliged to use because they (and not many others, in
those times) needed to travel among China’s complex patchwork of lan-
guage areas. (The term guanhua is reflected in the English word “Manda-
rin,” which originally referred to these officials.) During China’s late-Qing
and Republican years (from about 1860 until 1949), grammatical influences
from Western languages as well as a large amount of new vocabulary based
on Japanese and Western terms flowed into China’s official language. In
most ways, though, the official language remained continuous with the
earlier guanhua.
The cataclysmic events of the late Mao era, from the late 1950s to the
early 1970s, then brought a significant change. The bifurcation between
official language and ordinary language in China grew much sharper and
more pervasive than it had ever been before. In the Anti-Rightist Move-
ment of 1957, people were forced to explain in public how they had come
to be, officially speaking, “anti-Party” and “antisocialist”— even though
these words did not correspond at all to their inner feelings. During the
Great Leap Forward, people had to speak of “great bountiful harvests” at
the same time that the largest famine in world history was unfolding. Not
only was the gap between official language and daily-life reality suddenly
much larger than before; it was now no longer just the officials but nearly
everyone in society who had to learn to negotiate the official language.
Moreover, the penalties for missteps were more severe than they had even
been.
China’s eminent journalist Liu Binyan has written of his personal dis-
covery of this language bifurcation. Liu was labeled a “Rightist” in 1957
and sent to work among dirt-poor farmers during 1958– 60. There, in a
gully in the mountains,
a struggle began to rage deep inside me: how could two diametri-
cally opposed “truths” coexist in the world? The longings of the
peasants were one truth, and the policies of the higher-ups and the
propaganda in the newspapers were quite another. Which should I

follow?2
After some soul-searching, Liu opted for the “bottom up” truths that he
saw in the daily lives of farmers. But at the same time (and here he was like
most people in Chinese society) he learned to handle both versions of the
Chinese language. A person had little choice. If you wanted to get certain
things done—buy a bicycle, get a marriage license, or take a family photo,
for example— simply saying so would not get you very far; you had to ma-
nipulate the official language rather as one manipulates the pieces in a
chess game, referring to such things as “revolutionary needs,” “class
stands,” or your desire that “policy be correctly implemented,” being
careful to put things in such a way that small matters like your bicycle,
your marriage license, or your photograph seemed almost incidental. If
you suffered political attack, the need for skill in language manipulation
became even more critical; your words offered in defense had to fit with
the words in the attack or they would not count; wrong political terminol-
ogy could confirm that your politics were wrong, and that could confirm
the original charges.
Even when the official language was not needed for such vital purposes,
it was a good idea to stay in touch with it. It was, after all, “another kind of
truth.” It could tell a person about policy, and policy was important to know
about regardless of what one thought of it. When policy during the Great
Leap Forward called for planting rice stalks much closer together than
ever before, it suddenly became crucial for farmers to know the policy even
if they also knew that the stalks would die if planted in such a way. One’s
livelihood, even one’s life, could depend on knowing what the policies
were. In addition to learning about policy, sophisticated readers could fi nd
ways to mine official language for ordinary-life truths that were buried
inside it. Wu Zuxiang, an author of splendid short stories in the 1930s and
later a professor of Chinese literature at Peking University, put it this way:
2. “Listen Carefully to the Voice of the People,” in Liu Binyan, Two Kinds of Truth:
Stories and Reportage from China, ed. Perry Link (Bloomington: Indiana University
Press, 2006), p. 31.
Politics 237
There is truth in Chinese newspapers, but you have to know how to

find it. This often means reading upside down. If they say great
strides have been made against corruption in Henan, you know that
corruption is especially bad in Henan. If they say dozens of police
were hurt in a clash with students, you know hundreds of students
were injured if not killed.3
Wu was speaking in 1980, four years after the death of Mao Zedong and
at a time when Chinese society was emerging from the extreme repres-
sion of the late Mao years. By 1980 ordinary talk about nonpolitical things
had returned to the public sphere in China in ways that had not been pos-
sible under Mao. But the juxtaposition of the two kinds of language in
public life only accentuated the sense of artificiality of the official lan-
guage. It became more and more apparent that in using it people were
only shuffling words that had been drained of the ideals and principles
they had originally represented.
The bifurcation was sufficiently well established by 1980 that people
living their daily lives tended not to notice it. It was just the way things
were in the Chinese language. But to an outsider like me, who fi rst lived
in China during 1979–80 (I had been to China before, but only for short
trips), the distinction of levels was both startling and fascinating. In
newspapers, on the radio, in classrooms, at formal meetings, or in official
welcomes to foreign guests, there was one kind of language. Buying fish
in the market, asking your sister-in-law to pass the soy sauce, or shouting
at a child to get out of the rain, there was a very different kind of lan-
guage. By “language” I do not mean just tone, or what linguists call “reg-
ister.” I mean also vocabulary, pronunciation, and grammar. To be sure, it
was all “Chinese,” but the two kinds of Chinese presented a sharp and
pervasive distinction. Even later, in the 1990s and the early twenty-first
century, as linguistic influences surged into China from the outside world
through travel, television, and the Internet, the official language has re-
mained largely intact. It has kept its distinctive diction, grammar, and
3. Interview, July 18, 1980, Beijing.

aura, and has continued to occupy a special plane within Chinese lan-
guage use.4
One way to test for the large but often unnoticed gap between the of-
ficial and unofficial languages in China is to observe how mixing of the
two levels produces incongruity that people fi nd laughable. Judith Shap-
iro and Liang Heng, observing Chinese student life in 1983, note that “a
form of black humor was common among young Chinese . . . the repeti-
tion of any of the [official] slogans in ordinary conversation was almost
certain to bring a laugh.”5 I observed this phenomenon myself around the
same time at a dinner party in Shanghai. Someone put a delectable morsel
on the plate of a friend, and the friend responded, as is normal, bie keqi, ziji
lai ߿ᅶ⇨, 㞾Ꮕᴹ ‘don’t be polite, help yourself’. The morsel deliverer then
responded—not normally—by saying buyaojin, wo wei renmin fuwu ϡ㽕㋻,
៥ЎҎ⇥᳡ࡵ ‘that’s all right, I’m serving the people’. The phrase precipi-
tated sharp laughter from everyone within earshot. The Maoist slogan
“Serve the People” belonged to such a different level of language that it
was highly incongruous, and therefore funny, to insert it into such a re-
laxed and informal context. (The laughter probably also derived, at least
in part, from the sense of relief that comes when the tension involved in
use of the official language is made to dissolve in mirth.) Near the end of
Jiang Zemin’s rule in China in 2002, there were popular jokes about “three
[people] wearing wristwatches” (sange dai biao ϝϾ᠈㸼) because the phrase
was a pun on Jiang’s pretentious phrase for the new roles of the Commu-
nist Party as “The Three Represents” (sange daibiao ϝϾҷ㸼).6
4. The last time I visited China, in the summer of 1996, I was detained at the Bei-
jing airport (no reason was given) and held overnight in a nearby hotel by four Chi-
nese policemen before being sent back to the U.S. The police used official language to
tell me the rules of my stay: I could not exit the room, I could not make a phone call, I
could not leave their company, etc. That done, the four of them, who were young and
curious, reverted to a very different level of language to ask: “How much did your
watch cost?” “How did you learn Chinese?” etc. For details see my essay “Beijing
yiyou” ࣫Ҁϔ䘞, in Banyang Suibi ञ⋟䱼ㄚ [Notes of a semiforeigner] (Taipei: San-
min chubanshe, 1999), pp. 93– 98.
5. Judith Shapiro and Liang Heng, Cold Winds, Warm Winds: Intellectual Life in
China Today (Middletown, Conn.: Wesleyan University Press, 1986), p. 41.
6. The three things that the Communist Party should represent, according to Jiang
Zemin at the Sixteenth Party Congress in November 2002, are (1) the requirements of
Politics 239
Even officials sometimes played with this kind of pun, and records of
such play sometimes found their ways into internal Party documents. At a
1979 Party meeting to discuss how to repress a cartoon strip that, in the
view of certain high officials, showed the violence of the Cultural Revolu-
tion in excessively stark terms, officials who opposed the repression sati-
rized it by punning on politically charged terms. What the people on the
repressing side stood for, they quipped, was not lishi weiwuzhuyi ग़৆ଃ⠽
ЏН ‘historical materialism’ but lishi weiwu zhuyi ग़৆Ў᮴ЏН ‘history-
is-nothing-ism’. And it was the same people, they said, who had turned a
wenhua dageming ᭛࣪໻䴽ੑ ‘Cultural Revolution’ into a wuhua dageming
℺࣪໻䴽ੑ ‘militarized revolution’.7
Puns by nature trade on the incongruity of two levels of language, but
in this kind of politically loaded pun the contrast is sharper, and the resul-
tant laughter often stronger, than what puns normally provide. When I
did research at Zhongshan University during the 1979–80 academic year,
students of English sometimes sought me out for language practice. On
one occasion a group wanted to learn the game of charades. After I ex-
plained the game’s principles, some of the students went into an adjacent
room to plot what they were going to act out, and they suddenly burst into
irrepressible laughter. Someone had suggested that they do “class struggle,”
which, as a euphemism for attack on people during the Cultural Revolu-
tion, was still a phrase that inspired fear. In this case the release of tension
was far more than what a mere pun could have achieved.
I was visiting the university to study “scar” literature, which consisted
primarily of short stories that lifted the curtain on the fear, violence, and
other trauma of the late-Mao years. Scar literature was parallel in many
ways to the “thaw” literature that helped Russians begin to look at the
pain and repression of the Stalin years after Stalin died in 1953. One reason
why Chinese readers loved scar stories was that they could help to release
tensions and animosities that had long been repressed; at the same time,
the development of China’s advanced productive forces, (2) the orientation of the de-
velopment of China’s advanced culture, and (3) the fundamental interests of the over-
whelming majority of the people in China.
7. Guanyu lianhuanhua “feng” de zuotanhui fayan jilu ݇Ѣ䖲⦃⬏“ᵿ”ⱘᑻ䇜Ӯথ㿔㑾ᔩ
(Transcript of the conference on the comic strip “Maple”), comp. Editorial Offices of
Lianhuanhua 䖲⦃⬏ [Comic strips], August 11, 1979, pp. 17, 18.
and in part because of precisely that incendiary potential, the stories were
sometimes “politically sensitive.” Officials at several levels worried about
whether they personally might be recognizable in the ostensibly fictional
portrayals of villains, while the rulers at the very top monitored the ques-
tion of whether the overall outcry might rock their whole system and
threaten their monopoly on power. These two sides of the response to scar
literature—popular enthusiasm and official apprehension—were expressed,
respectively, in both of the two levels of language, official and nonofficial,
that we have been considering here. In classrooms in which I was an audi-
tor, professors of Chinese literature used the official language to lecture
to students on the Party-approved history of modern Chinese literature;
but during the breaks between classes (which, because of the infectious
interest in scar literature, grew longer than they were scheduled to be), the
same professors and students spoke in lively ordinary language about the
latest scar stories.8 During the day, government bureaucrats pronounced
warnings, in austere official language, about how some scar stories were
going too far, but in the evenings, after hours, those same officials some-
times went home to discuss the same stories with their children, in infor-
mal language and in much more positive terms. Hu Yaobang, minister of
propaganda of the Communist Party of China at the time, delivered a
major speech on scar literature in February 1980, in which he acknowl-
edged this problem of public versus private commentary on politically
sensitive works. “What you might say casually at home,” Hu told an assem-
bly of officials, “doesn’t matter very much. But to speak in public is to do
official thoughtwork, and this produces social effects . . . that can be good
or bad.”9 Hu was accepting the political bifurcation of the Chinese lan-
guage and simply trying to ensure that the two kinds of language operated
in their proper spheres.
8. Personal observation of classes on modern Chinese literature taught by Profes-

sors Wang Jinmin ⥟ᰟ⇥and Huang Weizong 咘ӳᅫ, Zhongshan University, spring
1980.
9. Hu Yaobang 㚵㗔䙺, Zai juben chuangzuo zuotanhuishang de jianghua ೼࠻ᴀᎹ԰Ӯ
䆂Ϟⱘ䆆䆱 (Talk at the conference on playwriting), February 12–13, 1980. Published
in 1980 as a pamphlet with no date or place listed. Once available on the Internet at
http://news.xinhuanet.com /ziliao/2005-02/04 /content _2548212.htm, but not available
as of June 29, 2012.
Politics 241
In certain ways the official language in contemporary China has always

had more power than the ordinary language. It conveys the ideas of “lead-
ers,” and these ideas, one way or another, do have important effects on
ordinary life. It hardly follows, however, that the official language is itself
more infused with life; indeed the opposite is almost always the case. The
official language normally is colorless and boring. People tolerate it, and
sometimes study it, because of its possible impacts—but not, usually, be-
cause of any intrinsic attraction to it as language. Mark Salzman, who
taught English for two years in China after he graduated from Yale in
1982, wrote this after observing some official meetings at his school:
The Chinese have, by necessity, increased their endurance manyfold

by making listening optional. During meetings they talk with one
another, doze, get up and stretch or walk around, and in general do
not pretend to pay attention. This does not seem to offend the speaker,
who, in general, does not pretend to be interested in what he or she is
saying.10
If the speaker is as bored as the listeners, why, one might ask, do such
meetings happen at all? In broad terms, the answer to this question is that
their purpose is largely ritual; the point is to mark the fact that everyone
present subscribes to the content of the speech. At meetings where deci-
sions are actually made or argued about— and these would be meetings
that a young foreigner like Salzman certainly would not have been al-
lowed to observe—the atmosphere is more lively and, significantly, the
language itself is often much more informal. At a confidential Party meeting
on literature and art in Guangxi in the winter of 1980, for example, some
delegates were arguing against the continued influences of Maoism on
literary policy by saying that Mao could not even control his wife, Jiang
Qing (leader of the discredited “Gang of Four”). People present at the
meeting report that this thought spread among the conference delegates
in the earthy phrase lian laopo dou mei fa guan, zenme neng guan wenyi ne?
䖲㗕ယ䛑≵⊩ㅵ, ᗢМ㛑ㅵ᭛㡎ਸ਼? ‘[Mao] had no way to handle even his
old woman, how could he handle literature and art?’. In response, Hu
10. Mark Salzman, Iron and Silk (New York: Vintage Books, 1986), p. 162.
Yaobang, the Party’s top propaganda official, reportedly warned “we have
to get away from this tendency” of saying Mao “could not even corral his
own wife.”11 Around the same time Hu Qiaomu, who for years had been a
secretary to Mao, was deputed to deliver good news to a young playwright
named Zhao Zixiong 䍉ṧ䲘. The message was that Zhao’s play The Fu-
ture Beckons (Weilai zai zhaohuan ᳾ᴹ೼ী૸), which had drawn heavy
criticism from Party authorities, would not likely land the youngster in
prison. The published criticism of Zhao’s work had all been in formal lan-
guage. But Hu Qiaomu’s reprieve, which was delivered privately, used lively
informal language. These words were, as reported by witnesses, xi ge zao,
shui ge jiao, zhe shi bu hao xi ⋫Ͼ╵, ⴵϾ㾝, 䖭ᰃ䚼ད៣ ‘have a bath, take
a nap, this is a good play’.12 More recent examples of informal language at
very high levels in Chinese politics are not hard to find. In the 1990s Deng
Xiaoping was widely quoted as saying that the Communist Party’s Polit-
buro “should have only one mother-in-law” (i.e., one boss, himself ); other-
wise, it would be a “many-headed horsecart” and get nothing done.13
In considering the differences between official and unofficial language,
therefore, it is important not to think of them as two different kinds of
language that two different groups of people speak. They are, rather, two
ways of speaking, and people of many kinds draw on them as needed. To
be sure, officials use the official language more than ordinary people do,
but that is only a matter of degree. The two kinds of language can even be
interlarded within the same conversation. In a formal interview a person
might break from official language to say something informal, as if paren-
thetically. Sometimes the volume of a person’s voice will drop in such
11. The reports of what was said at the meeting came from Chen Huangmei 䰜㤦✸,
the deputy director of the Institute for Literary Research of the Chinese Academy of
Social Sciences, and Li Yingmin ᴢ㣅ᬣ, the acting chief of the literature and art
bureau of the Propaganda Department of the Communist Party of China. They
were relayed to me by Wang Jinmin ⥟ᰟ⇥, professor of Chinese at Zhongshan
University.
12. The sources here are the same as those in note 11.
13. See Bao Pu, Renee Chiang, and Adi Ignatius, eds., Prisoner of the State: The Secret
Journal of Zhao Ziyang (New York: Simon and Schuster, 2009), p. 209; and Yan Jiaqi
Ϲᆊ݊, “ ‘Santou mache’ de lishi kaocha” “ϝ༈偀䔺” ⱘग़৆㗗ᆳ (A historical exami-
nation of ‘the three-headed horsecart’), appendix to Gao Gao 催ⱟ, Santou mache shidai
ϝ༈偀䔺ᯊҷ (The era of the three-headed horsecart) (New York: Mirror Books, 2009).
Politics 243
cases. A softer tone can seem to say “here’s a little unvarnished item for
you.” Credibility can go up as volume goes down.
In the next section I will explain some of the formal and conceptual
features that set China’s official language apart from the language of daily
life.
Characteristics of the Official Language
Every living language changes, and so does what I am here calling China’s
“official language.” Today’s modern bureaucratese has roots in China’s
late-Qing years and took further shape during the republican era. My fo-
cus below is on Communist-era official language, but even that has been
through changes. In 1951 the distinguished Chinese linguists Lü Shu-
xiang and Zhu Dexi lent their scholarship to a government-sponsored ef-
fort to standardize the national language and “purify” it, that is, rid it of
several things, including Western syntax, classical influences, and bor-
rowings from dialects.14 None of this was easy to do, or could be achieved
to perfection, but after a government campaign to get the messages out
in print and over the radio, considerable standardization did result, at
least in the official language. Some Chinese writers have argued that Lü
and Zhu laid the foundation for what became known (and sometimes de-
plored) as the “Maoist literary style” (mao wenti ↯᭛ԧ) that dominated in
the 1960s and 1970s. A number of scholars, including T. A. Hsia, H. C.
Chuang, Lowell Dittmer, Chen Ruoxi, Xing Lu, and Ji Fengyuan, have
written incisively on what might be called “the language of high Maoism”
and on its influences on how people conceptualize things.15 My own focus
14. Lü Shuxiang and Zhu Dexi, Yufa xiuci jianghua 䇁⊩ׂ䕲䆆䆱 (Lectures on
grammar and rhetoric) (Shanghai: Kaiming shudian, 1951).
15. T. A. Hsia, Metaphor, Myth, Ritual, and the People’s Commune (Berkeley: Center
for Chinese Studies, University of California, 1961), A Terminological Study of the
Hsia-Fang Movement (Berkeley: Center for Chinese Studies, University of Califor-
nia, 1963), and The Commune in Retreat as Evidenced in Terminology and Semantics
(Berkeley: Center for Chinese Studies, University of California, 1964); H. C. Chuang,
The Great Proletarian Cultural Revolution: A Terminological Study (Berkeley: Center
for Chinese Studies, University of California, 1967), and The Little Red Book and Cur-
rent Chinese Language (Berkeley: Center for Chinese Studies, University of California,
in what follows will be primarily on the official language of China’s late-

Mao and post-Mao years, from the 1960s through the early twenty-first
century. No cut of the temporal spectrum can be clean, however. Much of
what I address below are habits inherited from earlier times.
Another line that cannot be drawn clearly is any between post-Mao of-
ficial Chinese language and the official languages of societies and govern-
ments elsewhere in the world. There are differences, yes, but also many
common features. I have already noted the relevance of Orwell’s famous
essay “Politics and the English Language,” and there is plenty to study in
the spin-meistering of American politics. Comparisons between Chinese
officialese and the official languages of other authoritarian governments,
especially Marxist-Leninist authoritarian governments, will be even more
obvious, and admirers of Czeslaw Milosz, Miklos Haraszti, and Vaclav
Havel are sure to find many resonances in what I address below.16 Compari-
sons can be useful, and I will make some, but I do ask readers to bear in
mind that my focus below is on post-Mao China. When I assert something
about the political language of post-Mao China, I do not mean to imply
that the same is either true or not true anywhere else. Where I want to
make those further claims, I will make them.
Let us now turn to five features of post-Mao official Chinese, which I
will call, for convenience, lexicon and metaphor, grammar and rhythm,
moral weight, goal orientation, and “fit” as a kind of truth.
1. Lexicon and metaphor. In Chapter 2 we saw how English, like other Indo-
European languages, tends to invent nouns to handle abstract concepts.
1968); Lowell Dittmer, “Thought Reform and Cultural Revolution: An Analysis of

the Symbolism of Chinese Polemics,” American Political Science Review 71 (1977), 67–
85, and with Chen Ruoxi, Ethics and Rhetoric of the Chinese Cultural Revolution (Berke-
ley: Center for Chinese Studies, University of California, 1981); Ji Fengyuan, Lin-
guistic Engineering: Language and Politics in Mao’s China (Honolulu: University of
Hawaii Press, 2004); Xing Lu, Rhetoric of the Cultural Revolution: The Impact on Chi-
nese Thought, Culture, and Communication (Columbia: University of South Carolina
Press, 2004).
16. Czeslaw Milosz, The Captive Mind, trans. Jane Zielonko (New York: Knopf,
1953); Miklos Haraszti, The Velvet Prison: Artists under State Socialism, trans. Katalin
and Stephen Landesman (New York: Basic Books, 1987); Vaclav Havel, Living in Truth
(London: Faber and Faber, 1989).
Politics 245
When neurons connect we have neural “connectivity”; and so on. Lakoff

and Johnson called such uses ontological metaphors,17 and we saw how
they are much less common in Chinese— at least in Chinese before the
end of the nineteenth century, when neologisms derived from Western
languages began to enter the Chinese language in large numbers. Nouns
like zhengzhi ᬓ⊏ ‘politics’, jingji ㍧△ ‘economics’, and shehui ⼒᳗ came
in, and such terms were called mingci ৡ䖁 ‘terminology’, which itself was
another example of the new noun-borrowing. Nearly all of the new terms
had been invented in Meiji Japan, where Japanese modernizers had used
Chinese characters to invent counterparts to Western abstract nouns. For
Chinese, to borrow the terms felt like only half-borrowing, since they
were already expressed in Chinese characters. Later, Chinese began mak-
ing adaptations of their own, and eventually there were nouns like zi-
youhua ‘liberalization’ 㞾⬅࣪, zhutixing ‘subjectivity’ Џԧᗻ, and many
others. Many items of ordinary vocabulary in Chinese (da ᠧ ‘beat’, gou ⢫
‘dog’, xi ⋫ ‘wash’, tou 丁 ‘head’—the examples are legion) are monosylla-
bles. Most of the new Western-derived terms, though, were polysyllabic.
In 1959 a survey found that 98 percent of “new verbs” and 95 percent of
“new nouns” were polysyllables.18
This influx of polysyllabic, Western-derived abstractions into the Chi-
nese language played a major role in the formation of modern official Chi-
nese. Premodern official Chinese could also be pretentious—turgid, some-
times excessively ornate, and sometimes opaque. But the distinctive flavor
of modern official Chinese— simultaneously austere and vacuous, intimi-
dating yet elusive, in short stuffy and puffy at the same time— owes its
foundations to European habits of abstract conception that were brought
into Chinese in the late nineteenth century. Orwell recognized essen-
tially the same problem when, in 1946, he deplored the “pretentious dic-
tion” of European (especially Marxist) political language:
Chicago Press, 1980), pp. 25–32.
18. See San Duanmu, “Stress and the Development of Disyllabic Words in Chi-
nese,” Diachronica 16 (1999), p. 31. For “old” (i.e., traditional) vocabulary, 27 percent of
verbs and 83 percent of nouns (which included nouns with vestigial suffi xes like zhuozi
Ḡᄤ ‘table’ and yizi ộᄤ ‘chair’) were polysyllables.
Words like phenomenon, element, individual (as a noun), objective, cate-

gorical, effective, virtual, basic, primary, promote, constitute, exhibit, ex-
ploit, utilize, eliminate, liquidate, are used to dress up a simple state-
ment and give an air of scientific impartiality to biased judgments.19
Orwell’s worry about “an air of scientific impartiality,” which he finds

“pretentious,” has to do with the texture of language—the use, for example,
of multisyllabic words of Latin or Greek origin when plainer Anglo-Saxon
words would be clearer and livelier. But it also has to do importantly with
abstraction— speaking of phenomena instead of sunsets, of insects instead of
ladybugs, and so on.
The Chinese political commentator Cao Changqing has found the
same tendency toward abstraction in contemporary official Chinese and
calls it shuiguo yuyan ∈ᵰ䇁㿔 ‘fruit language’.20 When we hear the word
“banana” or “apple,” Cao notes, we can picture an image in our minds.
When we hear “fruit” we don’t know what to picture. To a Party official
charged with delivering policy, the ambiguity of fruit language is pre-
cisely its virtue. If an official says “fruits are good,” and it turns out that a
higher-up decides that bananas are bad, the official can say “I meant ap-
ples.” Fruit language preserves an official’s options and might even save
his or her career. For example, in the days before the Beijing Massacre in
June 1989, two groups of leaders, headed by Li Peng and Zhao Ziyang,
respectively, differed over the fateful question of whether to order a mili-
tary crackdown. Luo Gan, a high official who had recently been minister
of labor, approached both the Zhao and Li camps with a bit of fruit lan-
guage: “I am certain justice will prevail,” Luo said.21 That made it possible
for Luo, once the massacre was over and officially “correct,” to claim that
he had been siding with Li Peng. But had matters gone the other way, he
could have been on the other side.
Liu Binyan has pointed out other examples in which phrases are abstract
enough that they can accommodate different or even opposite interpreta-
tions. In Liu’s examples the beneficiary is not an individual person, as in
19. Orwell, “Politics and the English Language,” p. 167.

20. Cao Changqing ᳍䭋䴦, “Yuyan baoli: Jiquan tongzhi de weishe liliang” 䁲㿔ᲈ࡯:
Ὁ⃞㍅⊏ⱘ࿕ᜥ࡯䞣 [Linguistic violence: The power of intimidation in authoritarian
rule], Zhongguo zhi chun Ё೟П᯹ [China spring], January 1992, p. 49.
21. Reported to me orally by Zhang Zuhua, June 18, 2008, in Princeton, New Jersey.
Politics 247
the Luo Gan example, but a political group—usually part or all of the
Communist Party. In such cases the point of abstract language is to pre-
serve a veneer of unity over controversies that remain unresolved beneath
the surface. Liu’s examples are from 1991,22 when Party policy was jianchi
sixiang yuanze, jixu gaige kaifang മᣕಯ乍ॳ߭, 㒻㓁ᬍ䴽ᓔᬒ ‘persist with
the Four Basic Principles and continue with reform and opening’. The
two parallel phrases, which represented countervailing impulses despite
their syllabic balance, presented an overall meaning that allowed radical
ambiguity. People on one side could say it meant weile jianchi sixiang jiben
yuanze, women yiding yao gaige kaifang Ўњമᣕಯ乍෎ᴀॳ߭, ៥Ӏϔᅮ㽕
ᬍ䴽ᓔᬒ ‘in order to continue with the Four Basic Principles, reform and
opening absolutely have to be pursued’, while people on the other side
could say gaige kaifang jue buneng weifan sixiang jiben yuanze ᬍ䴽ᓔᬒ‫އ‬ϡ
㛑䖱ডಯ乍෎ᴀॳ߭ ‘reform and opening must never be allowed to vio-
late the Four Basic Principles’. Another of Liu’s examples is the policy
watchword gaohuo guoying dazhongxing qiye shi yige zhengzhi wenti ᧲⌏೑㧹
໻ЁൟӕϮᰃϔϾᬓ⊏䯂乬 ‘to bring large and medium-sized state enter-
prises to life is a political question’. People with different interests used this
phrase under fundamentally different assumptions about what the key
“question of politics” was. Liu writes:
One group said that it meant “if you don’t solve this problem, the
economy will go nowhere, and if that happens the people will oppose
you and your socialism will be smashed, as it was in Eastern Europe and
the Soviet Union.” The other group said it meant “it’s fine to enliven
state enterprises, but current political arrangements must not be
changed, which means the ‘political core functions’ of Party secre-
taries must stay in place and not be taken over by factory managers.”23
The ambiguity of this kind of carefully crafted, overarching phrase had,

in Liu’s view, a number of advantages for officials. For high officials, the
phrases had the advantage of allowing one to stand above two contradictory
22. Liu Binyan ࡝䊧䲕, “Yi jiu jiu yi: Zhongguo yin zhuan duo yun” ϔббϔ: Ё೟
䱄䔝໮䳆 [1991: The skies in China beginning to clear], Mingbao yuekan [Ming Pao
monthly], no. 12, 1991, p. 79.
23. Ibid.
positions and appear to represent a sort of transcendent unity that sub-

sumed both. To others, who were lower in the power structure, the ambi-
guity allowed room to squeeze their own viewpoints into phrases that had
already been anointed from above as “correct.” But the purest examples of
the advantages of abstraction, for officials, are the ways the two- or three-
syllable terms that Cao Changqing calls “fruit terms” can generate an air
of authority even while their meanings remain vague.
Abstract nouns that carry an “air of scientific impartiality,” as Orwell
put it, are easy to spot in Communist political jargon: xingshi ᔶ࢓ ‘situa-
tion’, jumian ሔ䴶 ‘condition’, qingkuang ᚙ‫‘ މ‬circumstances’, cuoshi ᥾ᮑ
‘measures’, dongxiang ࡼ৥ ‘trends’, and many others. Although most such
terms are nouns, other parts of speech are affected as well—both in the
abstraction and the accompanying air of authority. Verbs like jiayi ࡴҹ
‘add’ and jinxing䖯㸠 ‘carry out’ can be prefi xed to other two-syllable verbs
(and in the process turning those verbs into nouns) to yield examples like
jiayi kaolü ࡴҹ㗗㰥 ‘add consideration’, jiayi fenxi ࡴҹߚᵤ ‘add analysis’,
and jiayi taolun ࡴҹ䅼䆎 ‘add discussion’; or jinxing fangwen 䖯㸠䆓䯂
‘carry out visitation’, jinxing caice 䖯㸠⣰⌟ ‘carry out estimation’, or jinxing
gaige 䖯㸠ᬍ䴽 ‘carry out reform’. 24 Jiayi and jinxing almost never contrib-
ute any useful meaning: jinxing taolun, jiayi taolun, and just plain taolun
almost always amount to the same thing, except for the airs. Tiyan ԧ偠,
literally “personally experience,” is similar. Normally one “lives” (shenghuo
⫳⌏), but to tiyan shenghuo ‘personally experience living’ is more formal
and political; it pretends to an experience that is somehow more scientific,
more correct.
Adverbs like xiangdang Ⳍᔧ ‘fairly, considerably’ and adjectival modi-
fiers like yiding de ϔᅮⱘ ‘definite’ are similarly useful in political language
because, like the nouns and verbs just mentioned, they claim authority but
lack any content that can be pinned down. If an official wants to give pub-
lic credit to someone but has little idea what, if anything, the person has
actually accomplished, he can say that comrade so-and-so’s work is xiang-
dang hao Ⳍᔧད ‘considerably good’ or that he zuo chu le yiding de gong-
24. See James H-Y. Tai, Syntactic and Stylistic Changes in Modern Standard Chinese in
the People’s Republic of China since 1949 (Washington, D.C.: United States Information
Agency, 1977), p. 7.
Politics 249
xian ‫ߎخ‬њϔᅮⱘ䋵⤂ ‘has made a defi nite contribution’. Some political

modifiers originate in standard meanings but then take on a spiffy veneer
when they turn political. Eventually, the spiff can overshadow the origi-
nal meaning. For example, guangda de ᑓ໻ⱘ ‘broad’ is used in Commu-
nist jargon only to modify politically correct groups: guangda de qunzhong
ᑓ໻ⱘ㕸ӫ ‘broad masses’, guangda de ganbu ᑓ໻ⱘᑆ䚼 ‘broad (masses
of) cadres’, and so on.25 The actual size of a guangda group is not as im-
portant as its political pedigree. An assembly of imperialist running dogs,
no matter how large, could never be guangda de zougou ᑓ໻ⱘ䍄⢫ ‘broad
running dogs’. In short, much of the original meaning of guangda leaks
away, leaving it, like the verb jiayi or the noun xingshi, primarily a vehicle
for airs.
Pretentious nouns, verbs, and other parts of speech can combine to pro-
duce sentences that seem all the more irrefutable just because everything
fits together on the same plane. The resulting verbiage, as Orwell has put
it, can make even “lies sound truthful and murder respectable.” During
the Cultural Revolution, people were terrorized, torn from their families,
even killed— events that a plain-speaking person might call jiapo renwang
ᆊ⸈Ҏѵ ‘literally, family-broken-person-perished’ or hunfei posan 儖亲儘ᬷ
‘literally, soul-flies-ghost-disperses’. But official language, in describing
the same events, could seal off all the pain, pull people to a distance
from it, and apply layers of varnish with phrases like caiqu cuoshi jinxing
zhengdun 䞛প᥾ᮑ䖯㸠ᭈ乓 ‘adopt measures to carry out rectification’.
Cao Changqing calls the result qiti yuyan ⇨ԧ䇁㿔 ‘gaseous language’,
which he characterizes as language that “says everything and yet says
nothing.”26
At the same time that some words, elevated to an abstract level, lost
their concrete meanings and began to drift within the official fog, others
came to moor on meanings that were indeed useful but sometimes differ-
ent from what they had originally been. Guangda ‘broad’, cited above, is
one kind of illustration; it came to mean something like “politically ami-
cable” as much as it meant “broad.” A more complete shift of meaning is
visible in a word like fandong ডࡼ ‘reactionary’. The literal sense of this
25. Ibid., p. 100.

26. Cao Changqing, “Yuyan baoli,” p. 49.
term is “oppose movement.” During the 1930s and 1940s the Commu-
nists were on the move, and it made sense for them to refer to their op-
ponents, the Kuomintang and others, as “opposing movement.” But after
the installation of a nationwide bureaucracy in the 1950s, the matter of
who was moving and who was blocking movement underwent fundamen-
tal change. By the end of the 1970s, one of the favorite targets of popular
“scar literature” was the unmoving and unmovable Communist bureau-
crat: whatever you said, the bureaucrat had a way of smiling and spinning
jargon—but doing nothing (and then sometimes, behind your back, ar-
ranging to punish you for having spoken up). Yet while the social role of
Communists had largely turned around, the word fandong ‘reactionary’
persisted as a standard term of abuse. Fandong still meant “opposing the
Communists,” but that now meant, ironically, that people who wanted
their society to move could be officially labeled as “opposing movement.”
In a similar way, the political terms zuo Ꮊ ‘left’ and you ে ‘right’
drifted away from their original meanings during the Mao years and began
to take on new and sometimes paradoxical senses. In the 1940s, when the
Communist movement stood in opposition to an authoritarian Kuomin-
tang regime and was publicly associated with the interests of society’s
downtrodden, its designation as “left” made sense. But in the 1950s, espe-
cially with the Anti-Rightist Movement of 1957, a difficult ambiguity
arose. The critics of the Party who were labeled “rightist” in 1957 stood
for things like free speech, a free press, fairness, the interests of the
downtrodden, and other notions that fifteen years earlier had been called
“left.” Meanwhile, the label “left,” in the 1950s and 1960s, continued to
refer to orthodox followers of the Party line. By the late 1970s, after Mao
died, this doctrinaire group had become known as baoshou ֱᅜ ‘conserva-
tive’. Their critics also called them jizuopai ᵕᎺ⌒ ‘ultraleftists’— and so
it happened that “ultraleft” and “conservative” became synonyms. In the
late 1980s, when the xinzuopai ᮄᎺ⌒ ‘new left’ appeared, its members
had to face the awkward question of what the “new left” had to do with
the “old left.” The new left’s inspiration arose in part out of academic lib-
eralism in the West, and the word “liberal,” in this Western sense, might
have been a good label for them. But, unfortunately, the term “liberal,”
which in literal translation would be ziyoupai 㞾⬅⌒, was out of the ques-
tion, because ziyoupai was already in use in China to refer to critics of the
regime whom the old left was still calling “rightist.” It will likely take
Politics 251
some time— and probably will not be possible before the end of Commu-
nist rule in China—to completely untangle this semantic knot.
The official language includes a number of characteristic metaphors.
Some are “dead metaphors” so well established that people no longer
think of them as metaphorical. Some fit together in patterns and consti-
tute “conceptual metaphors” in the sense that I discussed in the previous
chapter. The “stage metaphors” we considered are an example. An official
who takes a post may be said to dengtai ⱏৄ ‘ascend the stage’ and later,
leaving power, xiatai ϟৄ ‘come off stage’; political support from behind
the scenes is houtai ৢৄ ‘backstage’ support; an opponent can chang fan-
diao ଅড䇗; and so on. Stage metaphors are deeply rooted in the Chinese
language and were in wide use well before the Communist period. They
appear not only in language from official sources, but, at least as often, in
the informal talk of daily life. When protesting students in the spring of
1989 called for the ouster of Premier Li Peng, they chanted Li Peng xia tai
ᴢ吣ϟৄ ‘Get off stage, Li Peng!’
But if stage metaphors are deeply rooted, certain other metaphors in
the official language of contemporary China appear to have grown di-
rectly out of the Communist movement. Many scholars have observed,
for example, the prominence of military metaphors in recent decades, and
it is generally agreed that these arose from guerilla war terminology, which
had infused the Communist movement during the 1940s and was trans-
ferred to the project of building a new society in the 1950s. When “strug-
gle” was no longer literally warfare but the “struggle to build socialism,” a
whole constellation of new metaphors was born. T. A. Hsia has traced
how production became zhandou ៬᭫ ‘fighting a battle’ and a production
worker a zhanshi ៬຿ ‘fighter’; a workplace became a zhanxian ៬㒓 ‘battle
line’ and a person’s place within it a gangwei ቫԡ ‘sentry post’.27 Did this
happen only because Mao and his comrades could not imagine a different
vocabulary? Probably not, in the view of many scholars. It is more likely
that the militarization of language was part of a deliberate strategy. Ji
Fengyuan, who has made this argument, plausibly advances two ratio-
nales for it as “to teach [people] to subordinate themselves utterly to the
leaders” and “to transfer the urgency . . . of wartime struggles” to present
27. Hsia, Metaphor, quoted in Ji Fengyuan, Linguistic Engineering, pp. 87– 88.

concerns.28 Since Mao’s passing, the frequency of war metaphors in offi-

cial publications has declined, but they remain latent in the official reper-
toire and resurface from time to time, especially when China’s rulers see
a need to stimulate nationalist sentiment. When Tibetans rioted in March
2008, for example, an editorial in the Communist Party’s Tibet Daily an-
nounced a ni-si- wo-huo de diwo douzheng Դ⅏៥⌏ⱘᬠ៥᭫ѝ ‘fight to the
death between enemies’, in which the goal was to “resolutely counterat-
tack, thoroughly vanquish the reckless effrontery of the enemy forces, and
seize total victory in this battle.”29 That the opponent in this battle was
the Dalai Lama was not a reason to alter the metaphors; the reason for using
them was to stimulate and channel the allegiances of Chinese readers.
From the arena of pure politics, military metaphors have spread into
official language in most other spheres, including science, social trends,
and literature and art. In all cases, the prominence of military metaphors
seems to be greater when political sensitivities are higher. During the
Mao years and for several years thereafter, poems and short stories could
become tools of battle. A story might be well written, or especially illumi-
nating— or praiseworthy in other ways—but as soon as it broached politi-
cal comment, it could become a “weapon” in a “struggle” on a literary
“front.” When post-Mao scar stories began to appear in 1977 and 1978,
one critic praised them for fahui le duanpian xiaoshuo zhenchabing de zhan-
dou zuoyong থ᣹њⷁ㆛ᇣ䇈պᆳ݉ⱘ៬᭫԰⫼ ‘utilizing the battle func-
tion of short stories as advance scouts’.30 At a conference on the politics of
28. Ji Fengyuan, Linguistic Engineering, p. 89. There are obvious parallels to totali-
tarian language elsewhere. Henry Friedlander shows how Nazism “took technological
terms and applied them to human beings”; “The Manipulation of Language,” in
Henry Friedlander and Sybil Milton, eds., The Holocaust: Ideology, Bureaucracy, and
Genocide (Millwood, N.J.: Kraus International, 1980), p. 108 and elsewhere.
29. In Chinese: Jianjue huiji, chedi ba didui shili de xiaozhang qiyan da xiaqu, duoqu
zhei chang douzheng de quanmian shengli മ‫އ‬ಲߏ, ᕏᑩᡞᬠᇍ࢓࡯ⱘ౷ᓴ⇨✄ᠧϟএ,
༎প䖭എ᭫ѝⱘܼ䴶㚰߽, in “Zhenfeng xiangdui jianjue huiji, quebao Xizang zizhiqu
shehui wending” 䩜䫟Ⳍᇍമ‫އ‬ಲߏ⹂ֱ㽓㮣㞾⊏ऎ⼒Ӯ〇ᅮ [Focus squarely and
counterattack resolutely to ensure social stability in the Tibet Autonomous Region],
editorial, Xizang ribao [Tibet daily], March 17, 2008.
30. Liu Xicheng ߬䫵䆮, “Tan dangqian duanpian xiaoshuo chuangzuozhong de jige
wenti” 䇜ᔧࠡⷁ㆛ᇣ䇈߯԰Ёⱘ޴Ͼ䯂乬 [On a few questions in the creation of con-
temporary short stories], Guangming Daily, March 30, 1979.
Politics 253
comedians’ dialogues (xiangsheng Ⳍໄ) in 1980, an official sought to de-

fend the sarcasm of certain comic lines by saying they xiang bishou yiyang
cixiang diren ‫ࣩڣ‬佪ϔḋࠎ৥ᬠҎ ‘are like daggers pointed at the enemy’.31
In later times, after China’s economy began to flourish, military meta-
phors carried over into the commercial realm. In this context, they largely
shed their political connotations but still carried a special zing. For ex-
ample, in 2007 a television commercial for a brand of cell phones touted
the devices as shoujizhong de zhandouji ᠟ᴎЁⱘ៬᭫ᴎ ‘the fighter planes
of cell phones’.32 Here a qiyan rhythm (see Chapter 1) garnishes the mili-
tary metaphor with added stylishness. Such pizzazz apparently diverted
attention from the question of whether one would really want a fighter
plane in one’s pocket.
Many writers have noted how military metaphors in the official lan-
guage of mainland China eventually seeped into daily-life usage as well.
That such usage stems specifically from the Communist experience seems
to be attested by the fact that Chinese speakers from Taiwan, Hong
Kong, and elsewhere can fi nd it strange. To refer to the “education battle
front” sounds odd in Taiwan or in New York’s Chinatown, but jiaoyu zhan-
xian ᬭ㚆៬㒓 raises no eyebrows inside China. On a Beijing bus in 1994,
I heard a little boy tell his mother he had to pee. His mother explained to
him that Uncle Busdriver could not just stop in the middle of traffic for
him. He would have to hold on. Ni jianchi yi xia! Դമᣕϔϟ ‘persevere a
bit!’ she ordered him, using a phrase that I doubt would have occurred to
a mother in Taiwan, who more likely would have said something like ni
ren yi ren ba Դᖡϔᖡ৻? ‘can you hold on a bit?’. We need not assume that
the Beijing mother’s feelings toward her son were really more militaristic
than those of the imaginary Taiwan mother with whom I am comparing
her. But her idiom, and its roots, are different. In 2007 a law student from
Wuhan published an article in Southern Weekend decrying the prevalence
of “violent slogans” in Chinese society. Her concern was less with meta-
phorical violence in official language than with how it had rubbed off and
31. Xiangsheng chuangzuo zuotanhui bangongshi Ⳍໄ߯԰ᑻ䇜Ӯࡲ݀ᅸ, ed.,

“Xiangsheng chuangzuo zuotanhui jianbao” Ⳍໄ߯԰ᑻ䇜Ӯㅔ᡹ [Bulletin on the
conference on creating comedians’ dialogues] (n.p., 1980), day 1, p. 4.
32. I am grateful to Mao Sheng for the example.
been absorbed into common use by ordinary people. Why, for example,
would a sign against garbage dumping go so far as to read “the dumping
of garbage is strictly prohibited, on pain of death to your entire family!”
(yanjin dao laji, fouze quanjia si guangguang! Ϲ⽕‫צ‬ൗഒ৺ܼ߭ᆊ⅏‫?)!ܝܝ‬33
Cao Changqing has argued that the pervasiveness of violent language in
contemporary Chinese establishes an underlying “atmosphere of fear” that
brings “psychological pressure” to people and eventually becomes accepted
as a normal part of daily life.34
Another kind of metaphor that became prevalent in official Communist
language, especially in the realm of literature and art, was medical. This
family of metaphors originated in the Yan’an years, lasted through the
Mao years, and remained common through most of the Deng era. In a
1957 speech Mao referred to certain works of fiction as ducao ↦㤝 ‘poi-
sonous weeds’,35 and the term spread quickly during the Anti-Rightist
Movement that unfolded later that year. The opposite of poisonous weeds
were “correct” literary works. Zhengque ℷ⹂ ‘correct’ was a technical
term in all ideological matters, and correct literary works, in the medical
metaphor, were jiankang ‫ع‬ᒋ ‘healthful’, where “health” could refer to
the well-being either of an individual reader or the overall society that
readers inhabited. “Incorrect” works— everything from dissidence to
“hooligan” culture to pornography—were bujiankang ϡ‫ع‬ᒋ ‘unhealth-
ful’. In the late 1970s, after Mao’s death, shanghen Ӹ⮩ ‘scar’ joined this
family of metaphors. The term was borrowed from the title of Lu Xin-
hua’s story “Shanghen,” but was also understood as referring broadly to
literary works that looked back at the Cultural Revolution and served a
healing function for readers and society. Writers who probed deeply were
said to “dissect” ( jiepou 㾷ࠪ) society, using their pens as scalpels. Terms
like “poisonous weeds” had two lives: they rang of Gang-of-Four extrem-
ism, which was now officially under attack, yet persisted in both formal
and informal language. In 1980, in a major policy speech, Hu Yaobang
33. Li Fenglinᴢ޸ᵫ, “ ‘Baoli biaoyu’ keyi xiu yi” “ᲈ࡯ᷛ䇁”ৃҹӥ⶷ [Let us be

done with “violent slogans”], Nanfang zhoumo [Southern weekend], June 21, 2007.
35. Mao Zedong, “Speech at the Communist Party’s National Conference on Pro-
paganda Work,” in Selected Readings from the Works of Mao Tsetung (Beijing: Foreign
Languages Press, 1971), p. 496.
Politics 255
said: “comrades in literature and art are sensitive, so I don’t use the words
‘poisonous weeds’ very much. But poisonous weeds do, in fact, objectively
exist.”36 A year later, in another major speech, Hu again insisted that lit-
erature and art contained “unhealthful [bujiankang] and negative things
that hurt [you hai yu ᳝ᆇѢ] the people.”37
It is worth standing back for a moment to ask why the abstractions,
decentered meanings, and special metaphors of official language arise in
the first place. Why do states and their leaders prefer such things to plain,
concrete language? Reminding the reader that officialese is a widespread
phenomenon, and that by commenting on Communist Chinese society I
do not mean to suggest that the same conditions cannot be found else-
where, I would like to probe the reasons, which appear to be several.
One of the most basic reasons seems to be that the associations of “ab-
stract” with “high” and of “high” with “good” are deeply embedded in the
conceptual metaphors of Chinese, English, and many other human lan-
guages. As we have seen in Chapter 2, both Chinese and English use the
metaphor “up is good”: gaoshang 催ᇮ ‘high’ character is good character, a
“high-minded” person does not take the “low road,” and so on. Both lan-
guages also use the metaphor “abstract is up”: a more general level of analy-
sis is a “higher” level, a position of broader authority within a bureaucracy
is a “higher” position, and so on. Both of these metaphors are human con-
ventions, not natural facts, but they are sufficiently well established in our
minds to condition the way we think. And political authorities naturally
like the combination because it reflects well on their own “high” position.
The implicit claim of superiority in abstract language can transfer to the
speaker, supplying him or her with pomp.
A second advantage of abstract language is its utility in supplying sylla-
bles when one actually has little notion of what one is talking about, that
is, where the abstract level is in fact the only level that can be made to be
coherent. This advantage explains why words like “concrete,” “individual,”
and “praxis” are so useful in political language. They indirectly acknowledge
36. Hu Yaobang, Zai juben chuangzuo zuotanhuishang de jianghua.

37. Hu Yaobang 㚵㗔䙺, “Zai Lu Xun dansheng yibai zhounian jinian dahuishang
de jianghua ೼剕䖙䆲⫳ϔⱒ਼ᑈ㑾ᗉ໻ӮϞⱘ䆆䆱 [Talk at the memorial assembly for
the hundredth anniversary of Lu Xun’s birth], Guangming Daily, September 26, 1981.
the need to speak of nuts and bolts even though the speaker or writer can-
not provide them. They are a way to be concrete in an abstract way. Doris
Lessing has observed that phrases like “concrete steps,” “contradictions,”
and “interpenetration of opposites,” are useful in political language be-
cause they “fill up as much space as possible without actually saying any-
thing” while exuding an air of authority.38 They can cover up ignorance— or
fuzzy understanding at best—with something that sounds like wisdom.
A third benefit of abstract language can be to associate the speaker with
political trends and styles. Cao Changqing writes that “people use their
familiarity with the official language—in speaking, writing articles, ‘re-
vealing their thoughts’ at meetings, and preparing reports—to demon-
strate how they ‘stick tightly’ to the line of the times and show ‘political
depth.’ ”39 A general benefit of stylishness in any context is to provide the
security that comes from group participation; if I am stylish in the right
way, I am exempt from criticism by others who follow the same style. In
the authoritarian line-following that Cao Changqing analyzes, an impor-
tant additional security comes from hiding beneath what people above me
are saying. If my expression is “on message,” then I cannot be criticized
unless people “higher” than I am are also criticized. In order to be “cor-
rect,” I do not need to know anything about nuts or bolts. I need only to
manipulate verbiage correctly.40
Ignorance or vague ideas may be the most common weaknesses that
official language covers over, but there are others, and some are more in-
sidious. When Orwell speaks of making “murder respectable,” he has no-
ticed the utility of official language as euphemism. In such cases, the user
of abstract language might be quite clear on the nuts and bolts that lie
below and turn to abstract language precisely in order to cover them up.
For example, twentieth-century authoritarian regimes of several kinds
have used “cleansing” or “purification” as metaphors for systematic killing.
Hitler’s Säuberrungsaktion ‘cleansing action’, Mao Zedong’s qingli jieji
38. Doris Lessing, “Questions You Should Never Ask a Writer,” op-ed, New York
Times, October 13, 2007, p. A15.
40. As an academic, I feel a bit guilty in making the points about official language in
the preceding three paragraphs, since all three— and others— apply to pomp and ab-
straction in academic jargon as well.
Politics 257
duiwu ⏙⧚䰊㑻䯳ӡ ‘cleansing the class ranks’, Pol Pot’s “purification” of

Cambodian society, Slobodan Milosevic’s “ethnic cleansing” of Kosovo,
and others all draw from essentially the same ghastly metaphor. These
phrases now burn in our consciousness; they stand out, and we notice them.
But euphemism of a more pedestrian sort often goes unnoticed, even
though it surrounds us every day. To pick a representative example at ran-
dom, when unnamed Chinese people were accused of plotting to protest
at the Asian Games in Beijing in 1990, the Chinese government’s Minis-
try of Public Security noted that police were going to deal with them as
“international terrorists.” Would they be arrested? Jailed? Tortured?
Killed? The announcement said this: Zhongguo yi caiqu cuoshi jinxing daji
Ё೑Ꮖ䞛প᥾ᮑ䖯㸠ᠧߏ ‘China has already adopted measures to carry
out attacks [on them]’.41 Just what measures, doing exactly what, were not
things the Ministry wished to reveal.
This example illustrates another way official language uses euphemism.
The authorities, not wanting to acknowledge that their police, or secret
police, or military police have done something, say that “China” did it. In
the decades of Communist Party rule in China, it has become standard in
the official language to use the word “China” to refer to the views or
actions of the state, or of the Party, or even just an elite level within the
Party. Examples that refer to the views of “China” on climate change, or
anticorruption measures, or the Dalai Lama—views that statistically rep-
resent only a part, sometimes not a very large part, of the Chinese
populace— are easy to fi nd. Young people are trained to equate ai guo ⠅೑
‘love of country—patriotism’ with ai dang ⠅‫‘ ܮ‬love of Party’. A word like
zuzhi 㒘㒛 ‘organization’, originally abstract and colorless, took on a con-
notation of fear and gravity during China’s Mao years because of what it
meant at the concrete level; the “organization department” in a “work-
unit” controlled not just your career but your entire fate in life. In Hitler’s
Germany the Organization and in Pol Pot’s Cambodia the Angkar, “Orga-
nization,” used the abstraction in similar ways.42
41. Quoted in “Shi ri yao wen” क᮹㽕䯏 (Main news of the last ten days), in Xinwen
ziyou daobao ᮄ䯏㞾⬅ᇐ᡹ [News freedom herald, April 30, 1990, p. 1.
42. Henry Friedlander, “Manipulation of Language,” p. 106; and Someth May,
Cambodian Witness: the Autobiography of Someth May, ed. James Fenton (London: Fa-
ber, 1986), p. 149.
Below the elite who ran the “orga nization” in China were the renmin
Ҏ⇥ ‘people’, but this, too, was an abstract term whose practical uses served
the interests of the elite. In his incisive article “ ‘Non-people’ in the People’s
Republic of China,” Michael Schoenhals has shown how only some of the
people are renmin Ҏ⇥ ‘people’. Schoenhals quotes Central Committee
member Bo Yibo in 1981: “ ‘People’ includes workers, peasants, the urban
poor, intellectuals, etc., but certainly not landlords and comprador bour-
geois elements.”43 And even among the workers, peasants, and other fa-
vored categories, only those who are docile qualify as “people.” A worker
who seriously misbehaves can become a “bad element,” which is a separate
category from “people.” “No wonder,” Schoenhals concludes, “the Party
loves the People. Here is a semantic entity that by defi nition is incapable
of rebelling.”44 In Party parlance, the word qunzhong 㕸ӫ ‘masses’ is simi-
lar, but has been used a bit differently from renmin. Qunzhong shares with
renmin the connotation of “from below”: opinions from below, reactions
from below, and so on. But there are a few additional levels of meaning in
a phrase like qunzhong yijian 㕸ӫᛣ㾕 ‘mass opinion’. Formally, it is sup-
posed to mean the opinions of the majority in society. But these opinions
are hard to know, and are often unknown to the very people who use the
term. Within bureaucracies, qunzhong yijian means, practically speaking,
“opinion that the next lowest level in the bureaucracy claims to be the views
of everybody below them.” Bureaucrats at a given level are supposed to “re-
flect mass opinion” to the level above them; practically speaking, however,
qunzhong yijian can be the opinion of the bureaucrats themselves. Since the
people above them have little access to the actual views of the people below
them, bureaucrats in the middle can, and do, present their own views as
qunzhong yijian.
Terms that refer to democratic ideals are often welcomed in China’s
official language, because they sound good, even if they in fact are used in
ways that indirectly support authoritarianism. For example, a policy to-
ward the press that is kuansong ᆑᵒ ‘relaxed’ is referred to as “democratic”
43. Michael Schoenhals, “ ‘Non-people’ in the People’s Republic of China: A

Chronicle of Terminological Ambiguity,” Indiana Working Papers on Language and
Politics in Modern China, no. 4 (July 1994), p. 3.
44. Ibid., p. 16.
Politics 259
even though the fact of relaxation simultaneously implies the existence of

a control mechanism that could easily tighten any time authorities wished
it to. State control, be it tight or relaxed, is something different from de-
mocracy, but the label is used anyway. The dissident astrophysicist Fang
Lizhi, writing in 1987, objected to official language that described “people’s
representatives” (ostensibly a democratic term) who came to “inspect”
(guancha 㾖ᆳ) the university where he was vice president. People’s repre-
sentatives, he argued, are supposed to be listening to our opinions and
passing them upward, not coming to inspect us. “We are accustomed to
this term [guancha],” Fang wrote, but in terms of democratic theory “it is
entirely incorrect.”45
Fang was unusually astute; most users of the official language do not
notice such implications. But in other ways—in ways the official language
itself encourages— ordinary people could become attuned to fi ne distinc-
tions of political usage. Victims of the 1957 Anti-Rightist Movement, for
example, were said to “wear rightist caps” (dai youpai maozi ᠈ে⌒ᐑᄤ).
To get the label removed was to get the cap off your head, but there were
two ways to do this, zhai maozi ᨬᐑᄤ ‘remove cap’ and pingfan ᑇড ‘over-
turn [label]’, and the difference between them was tremendous. Zhai maozi
‘remove cap’, which happened to some rightists in the 1960s, meant that
you were no longer a rightist but a “former rightist.” Still a stigma followed
you, rather the way the label “former sex offender” does in U.S. society.
The implication that you had once been evil was indelible. Pingfan, which
happened to many people in the late 1970s, implied that your original
capping was a mistake. You never should have been capped in the first
place. Your reprieve now, at least in theory, was complete. Similarly fi ne-
tuned distinctions could apply to a person’s lishi ग़৆ ‘history’. If you made
political mistakes in the past, but then looked at them squarely, confessed,
showed contrition, and so on, your personal file might record the notation
lishi qingchu ग़৆⏙Ἦ ‘history is clear’. That was good, but not as good as
a file that said lishi qingbai ग़৆⏙ⱑ ‘history is pure’, which meant that
45. He Lilang ԩ╻ᳫ, “Fan zi neibu pipan cailiao: Fang Lizhi Liu Binyan yanlun”
ড䌘‫ݙ‬䚼ᡍ߸ᴤ᭭: ᮍࢅП߬ᆒ䲕㿔䆎 [Classified antibourgeois criticism material: The
words of Fang Lizhi and Liu Binyan], Jiushiniandai бकᑈҷ [Nineties], no. 6 (1987),
p. 39.
there never had even been a problem that had to be cleared up. If a politi-
cally sensitive appointment were at stake at the organization department
in one’s work unit, the difference between a lishi that was qingbai and one
that was qingchu would be crucial. People knew this, and so it became im-
portant never to “make a mistake,” even of the correctable kind, and never
even to be accused of making a mistake. This is one reason why bureau-
cratic inertia became so heavy within China’s socialist system.
While officials used the official language in these ways, ordinary peo-
ple sometimes borrowed its devices to use it “from the other end,” as it
were. A worker, instead of saying wo shi gongren ៥ᰃᎹҎ ‘I am a worker’
could choose to say, in more abstract political language, wo shi gongren jieji
៥ᰃᎹҎ䰊㑻 ‘I am [of] the working class’ and by doing so lay claim to a
status that could make a difference in a plea for some kind of access or
privilege.46 We will look further at such matters in the section “The Pop-
ular Response to the Language Game.”
2. Grammar and rhythm. Premodern official Chinese used rhythms, em-

ployed clichés, and sometimes obeyed set forms, but these patterns were
importantly different from the grammar rules, borrowed largely from
Western language, that contemporary official Chinese observes. Premod-
ern Chinese could omit subjects and elide meanings in a variety of flexible
ways, and punctuation, if any, was used more to mark rhythms or empha-
ses than to clarify conceptual relations. By contrast, in the official language
of contemporary China, sentence structure obeys an orga nization that
Western-style grammar imposes, and punctuation is used to keep that kind
of structure clear.
The officially published works of Mao Zedong are a good example.
Mao’s sentences are long and sometimes convoluted, but are impeccably
grammatical and carefully punctuated. We know from several sources
that Mao’s oral language was not like this. It was colloquial, earthy, and
certainly not “Western” in the manner of his officially published works.
There is irony, of course, in the Party’s decision to bring Western gram-
mar into Mao’s thought. The assumption of those who engineered this
conversion—mainly Hu Qiaomu and others at Yan’an—seems to have been
46. I am indebted for the example to Chen Ping, letter to author, May 16, 1992.
Politics 261
that Western grammar makes something seem more clear, more scientific-
sounding, more authoritative, or in any case, better. Yet it was also the
grammar of people who were considered, at another level, imperialists and
the enemy. How Hu Qiaomu and others handled this irony, or whether
they even noticed it, is not well understood.
The effects of Western grammar are especially noticeable in the ways
in which terms in the official Chinese language change parts of speech to
match Western-language usages. Nouns can become adjectives, for ex-
ample. In the early twentieth century, the word yingxiong 㣅䲘 ‘hero’ was
a noun whose corresponding adjective was yingyong 㣅࢛ ‘heroic’: Qiu Jin
was a nüyingxiong ཇ㣅䲘 ‘heroine’, but one spoke of a yingyong de Qiu Jin
㣅࢛ⱘ⾟⩒ ‘heroic Qiu Jin’. But at some point during the Cultural Revo-
lution years, yingxiong slipped, in Maoist language, to become an adjec-
tive, and then there could be yingxiong de renwu 㣅䲘ⱘҎ⠽ ‘heroic char-
acters’. Lexical slips of this kind, from nouns to adjectives, sometimes
became not only standard but especially “correct,” so that they were used
even in the presentation of Chinese language to foreigners. A standard
textbook of Chinese for foreigners, published in 1971, was called Jichu
Hanyu ෎⸔∝䇁 (literally, “foundation” Chinese).47
In the mid-1970s, James H-Y. Tai did a systematic survey of this kind of
category-shift in the grammar of China’s official language during the
Communist era.48 In addition to nouns becoming adjectives, Tai found
verbs becoming nouns. Tigao ᦤ催 ‘raise’, which in modern Chinese before
the Communist era was a verb, could now be a noun in a sentence like sheyuan-
men shenghuo you le hen da de tigao ⼒ਬӀ⫳⌏᳝њᕜ໻ⱘᦤ催 ‘the lives of
commune members had [i.e., saw] major raises’. The commune members
could also make hen da de jinbu ᕜ໻ⱘ䖯ℹ ‘big progress’, where jinbu,
originally a verb, is used as a noun, and their example could dedao quanguo
de puji ᕫࠄܼ೑ⱘ᱂ঞ ‘gain nationwide dissemination’, where puji
‘spread’, another verb, again is used as a noun.49 Western languages
commonly use the same kind of category switch, of course. In English,
47. Jichu Hanyu ෎⸔∝䇁, formal English title Elementary Chinese (Beijing: Shangwu
yinshuguan).
48. Tai, Syntactic and Stylistic Changes.
49. Ibid., pp. 46–47.
“progress” can be a noun or verb with only an accent change to mark

the difference. But in English is it generally not the case that the category
switch corresponds to a borderline between officialese and ordinary
language.50
Tai also notes how adjectives (also called “stative verbs”) imitate Western-
style adverbs in Communist-era officialese. Genghao ᳈ད ‘better’ plays
this role in a phrase like genghao de wei renmin fuwu ᳈དഄЎҎ⇥᳡ࡵ
‘serve the people better’, in which the particle di ഄ is inserted to convert
genghao from adjective to adverb. The same is true of genggao ᳈催 ‘higher’
in a sentence like women yao genggao de juqi geming de qizhi ៥Ӏ㽕᳈催ഄВ
䍋䴽ੑⱘ᮫Ᏸ ‘we must raise the revolutionary banner even higher’.51 Tai
is right that these usages are distinctive to the official language. No teacher
in the daily life of a third-grade classroom would tell a child ni dei genggao
de ju ni de shou Դᕫ᳈催ഄВԴⱘ᠟ ‘you must in-a-higher-manner raise
your hand’.
Even more commonly, and again with a strong flavor of officialese, ad-
jectives can turn into transitive verbs. Huopo ⌏⋐ ‘lively’ becomes a tran-
sitive verb in a statement like zhei jian shiqing huopo le ta de gongzuo ganjin
䖭ӊџᚙ⌏⋐њҪⱘᎹ԰ᑆࢆ ‘this matter enlivened his working spirit’.
The same is true of fengfu Єᆠ ‘lively’ in tamen fengfu le xiangcun wenhua
huodong de neirong ҪӀЄᆠњеᴥ᭛࣪⌏ࡼⱘ‫ݙ‬ᆍ ‘they enriched the con-
tent of rural cultural activities’.52
Nouns can turn into transitive verbs in the official language. In stress-
ing the need for “stability” in Tibet in the spring of 2008, the Tibet Daily
used guanxi ݇㋏ ‘relation’ as a verb to warn that xizang wending guanxi
quanguo wending 㽓㮣〇ᅮ݇㋏ܼ೑〇ᅮ ‘Tibetan stability relates to sta-
bility in the whole country’.53
James Tai makes the astute observation that Communist official lan-
guage tends to use the “aspect marker” le њ more than it is used in ordinary
50. Generally but not always. For a time in the 1990s it was stylish in American of-
ficialese, for example, to use interface, originally a noun, as a verb: “we interfaced for
twenty minutes,” etc. Such usage, as in the Chinese cases, serves as a marker that “this
is official language”; ordinary people in daily life speak more plainly.
51. Tai, Syntactic and Stylistic Changes, pp. 39, 41.
52. Ibid., pp. 43, 45.
53. “Zhenfeng xiangdui jianjue huiji, quebao Xizang zizhiqu shehui wending.”
Politics 263
language. One of Tai’s examples is ta shoudao le shisheng he pinxiazhongnong

de zanyang ҪফࠄњᏜ⫳੠䋿ϟЁ‫ⱘݰ‬䌲ᡀ ‘he received praise from the
teachers and students and the poor and lower-middle peasants’.54 Tai does
not offer an explanation of this use of le, but— since so many other features
of official-language grammar are borrowed from Western grammar—it
may be that this usage is born of an impulse to reflect Western-style past
tense. Chinese does not use tense, and therefore has no tense markers. But
if one were looking for a way to reflect a past tense, this minor abuse of le
would be one way to do it. Whether or not this speculation of mine is cor-
rect, there is no doubt that this sort of use of le was common in official
language by the late 1980s. News reports of meetings where officials gave
speeches commonly used phrases like XXX shizhang jiang le hua Ꮦ䭓䆆њ
䆱, YYY fushizhang ye jiang le hua ࡃᏖ䭓г䆆њ䆱, ZZZ tongzhi ye jiang le
hua ৠᖫг䆆њ䆱 ‘Mayor XXX spoke, Deputy Mayor YYY also spoke,
and so did Comrade ZZZ’. Le in such phrases, like other usages derived
from Western-style grammar, carries a subtle assertion of pomp.
Tai goes on to observe that in tandem with the rise of le in the official
language, the corresponding aspect marker zhe ⴔ ‘in the process of’ also
became more common. Tai lists examples such as tamen zhijian youzhe fei-
chang miqie de guanxi ҪӀП䯈᳝ⴔ䴲ᐌᆚߛⱘ݇㋏, literally something
like “between them there exists an extremely close relationship.”55 Again
Tai is right to point out the subtle abnormality of official language. In or-
dinary Chinese the same thought would probably just be tamen de guanxi
feichang miqie ҪӀⱘ݇㋏䴲ᐌᆚߛ ‘their relationship is extremely close’,
or something like that.
In general, the more intensely the official language focused on telling
people what to do, the more structurally rigid it became. To illustrate, let
us revisit the example from the Introduction, of the Cultural Revolution
phrase weida de guangrong de zhengque de Zhongguo gongchandang ӳ໻ⱘ‫ܝ‬
㤷ⱘℷ⹂ⱘЁ೑݅ѻ‫‘ ܮ‬the great, glorious, correct Communist Party of
China’. During the tense times when that slogan was used, you could, if
you chose, refer to the Party without using any adjectives at all; but if you
did choose to use adjectives in public contexts, then “great,” “glorious,”
54. Tai, Syntactic and Stylistic Changes, p. 76. I have changed the translation slightly.
55. Ibid., p. 76.
and “correct” were the three that you used— and if you used all three,
they had to be in that order. If you were to mix the order, and refer, for
example, to the “glorious, great, correct” Party, the “correct, glorious,
great” Party, or the “great, correct, glorious” Party, you would sound less
than great, not very glorious, or—most dangerously—not correct. If your
wrong word order were perceived as intentional, your phrase could even
smack of subversion and mean very serious trouble for you. The “correct-
ness” of a set phrase inhered very much in its form, not just its content.
The more politically charged the phrase was, the more form mattered.
As we saw in Chapter 1, form in official language was sometimes rein-
forced by rhythms, which could make phrases easier to memorize and
enhance the sense that they are inevitably “right.” This sense of rightness
could survive even when concrete content was hardly clear. It is hard to
say, for example, exactly what a person was supposed to do when following
the rhythmical Cultural Revolution slogan linghun shenchu gan geming
♉儖⏅໘ᑆ䴽ੑ ‘make revolution in the depths of your soul.’ Make revolu-
tion deep inside myself? What exactly am I supposed to do? Rebel?
Against what? Obey? Obey what? But whatever the phrase was supposed
to mean, one point was clear: it was correct. And here is one reason why
this sort of rigid and rhythmical language has been useful to authoritari-
ans. It says, in essence, “stop thinking and just accept official correct-
ness.” Young Red Guards during the Cultural Revolution could head off
to “struggle,” sometimes even to their deaths, under the spell of rhythmic
and spellbinding— but essentially contentless— phrases such as xia ding
juexin, bupa xisheng, paichu wannan, qu zhengqu shengli! ϟᅮ‫އ‬ᖗ, ϡᗩ⡎⡆,
ᥦ䰸ϛ䲒, এѝপ㚰߽ ‘firm your resolve, fear no sacrifice, cast aside every
obstacle, go forth and seize victory!’ Or even more bluntly: yi bupa ku, er
bupa si ϔϡᗩ㢺, Ѡϡᗩ⅏ ‘fi rst, don’t fear hardship; second, don’t fear
death’.
This last example exhibits not only rhythm but repetition, which, in com-
bination with rhythm, can enhance the implicit sense that “this is right,
there is no need to think about it.” Other examples from the Mao era, also
cited in Chapter 1, are daming, dafang, dazibao ໻号໻ᬒ໻ᄫ᡹ ‘big out-
cry, big release, big-character posters’, in which the repetition of da . . .
da . . . da . . . hammers away as if to proclaim that there is no way that the
thrust of this phrase can or should be stopped; and loushang louxia, diandeng
Politics 265
dianhua ὐϞὐϟ⬉♃⬉䆱 ‘electric lights and telephones upstairs and

down’, where the parallel repetition of lou and dian also somehow creates
a sense that only a fool would gainsay the rightness of what is said. Al-
though less common in the Deng, Jiang, and Hu years than it was under
Mao, hammer-like repetition has continued to be used in political rheto-
ric, especially when the topic is high priority. In March 2008, in response
to the Tibetan uprising in Lhasa, a political editorial in the Tibet Daily
arrayed a set of seven consecutive four-character phrases: tigao jingti, caliang
yanjing, mingbian shifei, qizhi xianming, tuanjie yizhi, tongzhou gongji, tongchou
dikai . . . ᦤ催䄺ᚩ, ᪺҂ⴐ⴯, ᯢ䕼ᰃ䴲, ᮫Ᏸ剰ᯢ, ಶ㒧ϔ㟈, ৠ㟳݅⌢,
ৠқᬠᗒ . . . ‘raise vigilance, keep the eyes open, distinguish clearly be-
tween right and wrong, take a clear-cut stand, come together in unity, help
one another, hate the enemy together . . .’ This barrage was followed by
three six-character phrases and then three longer phrases each of which
began with baowei . . . ֱि . . . ‘safeguard . . .’56 The extremely important
political topic apparently called for unusual attention to linguistic craft.
Repetition in official rhetoric strengthens the demand on the listener
to accept the correctness of a phrase. It adds to what rhythm alone achieves
in this regard. But why should this be? Why should the repetition of words
or phrases lend them more authority? John Frankenstein has offered the
interesting hypothesis that repetitive rhythmic phrases tend to “infan-
tilize” the listener. The phrase (cited in the Introduction) gaogao xingxing
chucheng zou 催催݈݈ߎජ䍄 ‘happy-happy out-of-city go—have a happy
trip leaving the city’ has, Frankenstein points out, a childlike flavor. Da-
ming, dafang, dazibao ໻号໻ᬒ໻ᄫ᡹ ’big outcry, big release, big-character
posters’ carries it, too, although to a lesser degree. Children’s language,
especially in Chinese, uses a lot of repetition: mama ཛྷཛྷ ‘Mama’, baba
⠌⠌ ‘Papa’, baobao ᅱᅱ ‘precious (little one)’, guaiguai ЪЪ ‘well-behaved
(little one)’, and so on. Might it be that such language draws the minds of
listeners partway back to childhood, to a time when they were more ac-
cepting of paternalistic authority?57 Is this part of why an authoritarian
government finds such language useful?
57. John Frankenstein, email message to author, September 15, 2002.
As a broad generalization, Chinese-language political slogans from

outside the People’s Republic are less childlike than Beijing’s slogans. Po-
litical slogans in Taiwan tend to draw more on classical Chinese. For ex-
ample in 1996, when the Taiwan leader Li Denghui wanted to advise en-
trepreneurs from Taiwan on how to invest on the mainland, he announced
a slogan of jieji yong ren ៦ᗹ⫼ᖡ ‘be not hasty, use patience’. Mainland
slogans also sometimes use classical language, but it is worth noting that
when they do, they are slogans directed at other members of the ruling
elite, not at the (infantilized) “masses.” When Deng Xiaoping wanted to
set a strategy in foreign policy during China’s economic rise in the 1990s,
he advanced the classical-language phrase tao guang yang hui ䷀‫ݏܝ‬᰺
‘hide capacity and await the time’. This was a phrase that others in leader-
ship positions needed to hear; it did not need mass consumption.
The use of numbers in Communist-era slogans has achieved effects
that are similar to those of repetition. Many scholars have noted this
practice, and examples are legion. In 1928 Mao Zedong laid out san da jilü
ba xiang zhuyi ϝ໻㑾ᕟܿ乍⊼ᛣ ‘three great disciplines and eight points
for attention’ for the People’s Liberation Army. In the early 1950s, urban
China was consumed by the san fan ϝড ‘three antis’ and wu fan Ѩড ‘five
antis’ movements. In the Cultural Revolution, the population was divided
into the hong wu lei 㑶Ѩ㉏ ‘five red categories’ and hei wu lei 咥Ѩ㉏ ‘five
black categories’. Oracular wisdom descended from Mao as yi fen wei er
ϔߚЎѠ ‘one divides into two’. Stalin’s (and Mao’s, after his death) errors
and credits were summed up as sanqikai ϝϗᓔ ‘30 percent versus 70 per-
cent’. Deng Xiaoping’s new direction from the late 1970s was called sige
xiandaihua ಯϾ⦄ҷ࣪ ‘the Four Modernizations’, but Deng insisted as
well on the sige jiben jianchi ಯϾ෎ᴀമᣕ ‘Four Basic Principles’. A polite-
ness campaign called wu jiang si mei Ѩ䆆ಯ㕢 ‘five pay-attentions and four
beautifuls’ arose from the Deng era as well. Jiang Zemin, in his time,
added san ge daibiao ϝϾҷ㸼 ‘three represents’, and later Hu Jintao spon-
sored ba rong ba chi ܿ㤷ܿ㘏 ‘eight honors and eight shames’. These are
only a few of many examples. What do the numbers contribute? What ex-
plains their perennial attraction for the people who devise the slogans? It is
a special characteristic of Communist Chinese language. Other authoritar-
ian systems have used numbers, too, and so does Chinese culture in other
spheres. But not as much.
Politics 267
Might the use of numbers, like the use of repetition, be a way of infan-
tilizing the people to whom slogans are addressed? At a minimum, the
numbers seem to be saying, “here is a list of the things to focus on, and we
are numbering them to help you be sure none are overlooked.” Cao
Changqing has noted two other functions of numbers in slogans. One is
that, as with other lexical items in the official language, they are a bit
“gaseous” in the sense that they draw attention away from content and
toward the form of phrases— and form alone, as we have seen, can deliver
a sense of correctness. The other contribution of numbers is what Cao
calls gaikuoli ὖᣀ࡯ ‘completeness of coverage’.58 If I tell you, in official
cadence, that there are “eight honors and eight shames,” the implication is
that these sixteen points pretty well sum up everything. There is nothing
else you really need to know, or to ask. If you have grasped the “five pay-
attentions and four beautifuls,” then you are ready to be a fully polite per-
son; there is no need to go (or think) further.
3. Moral weight. Official language in contemporary China regularly as-

sumes a strong position of moral rectitude. Certain words, grammatical
patterns, and rhythms carry a strong connotation that “this is right.” It is
another feature that tends to stifle independent thought, because who will
argue with something that is automatically right? The automaticity of the
implication depends on the official context. Words and phrases that have
ordinary meanings in ordinary contexts take on a connotation of moral
rightness when used officially. We saw this briefly in the example of guangda
qunzhong ᑓ໻㕸ӫ ‘broad masses’, where guangda clearly takes on the con-
notation of “good.” We saw it also, in the Introduction, with the example
of zui ᳔ ‘most’, which consistently had positive connotations during the
late Mao years in phrases like zui zui zui weida de Mao zhuxi ᳔᳔᳔ӳ໻ⱘ
↯Џᐁ ‘the most, most, most great Chairman Mao’. Phrases with nega-
tive connotations, also with early roots in the Communist movement, in-
clude yixiaocuo ϔᇣ᪂ ‘a small bunch’,59 and ᵕᇥ᭄ ‘tiny minority’.

59. The phrase was used in a xiangsheng “comedian’s dialogue” piece, Zhi laohu 㒌
㗕㰢 [Paper tiger], published in 1950. See Xi Xiangyuan ᐁ佭䖰 and Sun Yukui ᄭ⥝
༢, eds., Zhi laohu (Beijing: Sanlian shudian, 1950), p. 26.
Even rhythms can carry connotations of right and wrong. In Chapter 1

we saw how wuyan and qiyan rhythms, in many kinds of official, commer-
cial, and daily-life contexts, can carry subtle suggestions that the ideas
they help to communicate are somehow more natural, authoritative, or
exalted than the same ideas would be if expressed without the rhythms. In
political language, especially in high-pressure political contexts, this kind
of effect is strong and easy to observe. Consider again this example, cited
in the preceding section: xia ding juexin, bupa xisheng, paichu wannan, qu
zhengqu shengli! ϟᅮ‫އ‬ᖗ, ϡᗩ⡎⡆, ᥦ䰸ϛ䲒, এѝপ㚰߽! ‘firm your re-
solve, fear no sacrifice, cast aside every obstacle, go forth and seize victory!’
The phrase has a strong and clear rhythm: TAH, TAH, ta-TAH! . . . TAH,
TAH, ta-TAH! . . . TAH, TAH, ta-TAH! . . . ta-TAH-ta-TAH! TAH! The
words themselves, of course, also have strong positive meanings. Can we
separate these positive meanings from the contribution of the rhythm?
Does the rhythm, by itself, have a positive connotation? I believe that it
does, and that we can easily see this by doing another experiment with
word substitution. What happens if we insert words that have negative
connotations while keeping the same rhythm? What if someone said: pai-
huai, panghuang! bimian xisheng! jiawei taopao! chiuh jieshou shibai! ᕬᕞᕋ
ᕼ! 䙓‫ܡ‬⡎⡆! །ሒ䗗䎥! এ᥹ফ༅䋹! ‘dither and vacillate, avoid sacrifice,
tuck tail and run, go and accept defeat!’ The expression sounds funny—
not just odd, but literally funny, and the reason it generates humor is that
the austere, positive implications of the rhythm are incongruous with the
negative content of the words.
There is a well-rooted custom in Chinese culture of choosing names for
children that express parental hopes. Names have often mentioned moral
virtues such as honesty, sincerity, and learning, or, especially for boys,
strength or bravery, and for girls, purity and refi nement. At the height of
Maoism, some of these traditional notions were jettisoned as “feudal” and
politics often entered the process of coosing names. Children born in the
late 1960s and early 1970s received names like Xiangdong ৥ϰ ‘toward
the East’ or ‘toward [Mao Ze]dong’, Xuegong ᄺᎹ ‘learn from workers’,
and even Miezi ♁䌘 ‘annihilate capitalism’.60 There are anecdotal accounts
of experiments with keeping to the left on roadways (since left is inherently
60. For more examples see Ji Fengyuan, Linguistic Engineering, pp. 153–154.

Politics 269
better than right), and having a red light (red is good) mean go, not stop.
The color red was powerful enough that it could turn the word hong 㑶
‘red’ into a transitive verb, as when socialist flowers hong bian renjian 㑶䘡
Ҏ䯈 ‘redden the entire world’.61 Red served as the opposite, morally speak-
ing, for both black and white. A baizhuan ⱑϧ ‘[politically] white expert’
was evil compared to a red expert, but so were any of the hei wu lei 咥Ѩ㉏
‘five [politically] black categories’ of landlord, counterrevolutionary, rich
peasant, rightist, or bad element.62
On the whole, animals have not fared too well within the moral land-
scape of the official language of Communist China. Some examples of
negative connotations— of dogs, for example—were inherited from pre-
Communist times. But the general trend during the Mao years turned
more sharply against animals. Donkeys, crabs, lizards, fl ies and insects,
tigers and leopards, and other species of real animals joined with cow
ghosts and snake spirits, dev ils, demons, ghosts, monsters, vampires, and
other imaginary beasts in a menagerie whose abiding unity, despite its
superficial diversity, was that all were used metaphorically to denigrate
one or another kind of human being. With or without modifiers, animal
terms were meant and understood as negative.
Terms that have turned intrinsically positive in the official language
include kexue ⾥ᄺ ‘science, scientific’, whose uses Michael Schoenhals has
analyzed.63 During the Mao years kexue sometimes meant little more or
less than “good” or “politically correct.” A sentence like ni zheige kanfa bu
kexue Դ䖭Ͼⳟ⊩ϡ⾥ᄺ ‘this view of yours is unscientific’ did not have to
refer to a point of science or to any claim that the scientific method had
not been properly applied. It could just be a general way of saying “I don’t
like your view.” The term wenhua dageming ᭛࣪໻䴽ੑ ‘Great Cultural
61. Hsin-cheng Chuang, The Great Proletarian Cultural Revolution: A Terminological

Study (Berkeley: Center for Chinese Studies, 1967), p. 8, cited in Ji Fengyuan, Linguis-
tic Engineering, p. 190.
62. See also Elizabeth J. Perry and Li Xun, “Revolutionary Rudeness: The Lan-
guage of Red Guards and Rebel Workers in China’s Cultural Revolution,” Indiana
East Asian Working Paper Series on Language and Politics in Modern China, no. 2
( July 1993), pp. 9–10.
63. Michael Schoenhals, Doing Things with Words in Chinese Politics: Five Studies
(Berkeley: Center for Chinese Studies, 1992), p. 9.
Revolution’ is another illuminating example because it had strong positive

connotations during the last years of Mao; but then, after Mao, the Deng
Xiaoping regime sought to purge the phrase of its positive connotations
even while continuing to use it. An order went out to China’s media that,
beginning in August 1980, the term should bear quotation marks: not Cul-
tural Revolution but “Cultural Revolution.” Here was a way the positive
connotations that one leadership group added to a term could be peeled off
by another group.
4. Goal orientation. The official language of Communist China has a ten-

dency to stress goals, often in absence of comment on how to achieve the
goals. The tendency is visible at several linguistic levels: from the struc-
ture of general policies down to the level of slogans and even to single
words. It is also evident in many spheres of life, ranging from military
orders to literary policy.
For example, at the level of general policies, after the Third Plenum of
the Eleventh Congress of the Communist Party of China in 1978, China’s
writers were exhorted to jiefang sixiang 㾷ᬒᗱᛇ ‘liberate [their] thought’.
They should not xin you yuji ᖗ᳝ԭᚌ ‘harbor residual fears’ because of
what happened to them and others during the Cultural Revolution. At the
same time, though, they were cautioned to observe sige jiben jianchi ಯϾ෎
ᴀമᣕ ‘Four Basic Principles’: the leadership of the Communist Party,
Marxism-Leninism-Mao-Zedong-thought, the dictatorship of the prole-
tariat, and the socialist system. These contradictory guidelines created a
problem for the writing process. How, in actually writing a story or
poem, could one follow both sets of instructions? It seemed that one was
being told to go and stop at the same time.64 The official language had set
forth goals, but the problem of how to reach the goals was left to individ-
ual writers. Similarly, but in a very different context, on June 3, 1989, sol-
diers of the Chinese army in Beijing were ordered to clear Tiananmen
Square of protesting students “resolutely,” with “absolutely no delays . . .
by 6 a.m. [on June 4]”; at the same time, there was to be “no bloodshed
64. This was, in fact, a standard tool in the project of inducing self-censorship in
writers, and we will consider it further in the section below “How the Game Is Played.”
Politics 271
within Tiananmen Square— period.”65 Soldiers were given rifles and ma-
chine guns. How, exactly, were the twin goals of fi nishing on time and
ensuring no bloodshed to be achieved? This was left to the ones receiving
the orders. The official message was just “Get it done.”
At the word level, the project of stating a goal without specifying a
means is well reflected by the widespread use in mainland China of the
“dummy” verb gao ᧲, a word that in isolation is almost impossible to trans-
late because it usually means literally nothing except “do,” “pursue,” or
“bring about” the noun that follows it. All of the freight of a gao phrase is
carried by the verb’s direct object. Gao has been especially useful in offi-
cial language because its dummy-verb nature is perfect for officials who
want to say “get [something] done” without committing themselves about
just what to do. An order to gao shengchan ᧲⫳ѻ ‘do production’ tells a
worker simply to produce more output— one way or another. Slogans like
gao shehuizhuyi ᧲⼒ӮЏН ‘pursue socialism’ and quanxin quanyi gao sihua
ܼᖗܼᛣ᧲ಯ࣪ ‘pursue the Four Modernizations with all your heart and
mind’ all lay emphasis on the desired results, even though, in these two
examples, the results themselves are expressed only vaguely. Eventually,
gao spread from the official language into many other spheres of life. By
the end of the Mao years, what a janitor did was gao weisheng ᧲ि⫳ ‘do
sanitation’, a physicist would gao wuli ᧲⠽⧚ ‘do physics’, a young person
gao duixiang ᧲ᇍ䈵 ‘pursue a [romantic] partner’, an artist gao meishu ᧲㕢ᴃ
‘do art’, and so on.
Gao is also used in Taiwan and overseas Chinese communities, but not
nearly as much as on the mainland. The usage has roots in regional dialects
(notably of Hunan, the native province of a number of Communist leaders)
and was sometimes used as a euphemism in sexually suggestive phrases
such as gao nüren ᧲ཇҎ ‘mess with women’ and luangao nannü guanxi
х᧲⬋ཇ݇㋏ ‘be promiscuous in sexual relations’. But by the middle of
the twentieth century, gao had become a mainstream term within the Com-
munist movement and was neutral in connotation and flexible in usage. In
65. Zhang Liang, The Tiananmen Papers: The Chinese Leadership’s Decision to Use Force
against Their Own People, ed. Andrew J. Nathan and Perry Link (New York: Public
Affairs, 2001), p. 370.
1951 Lao She, the master of Beijing brogue who had returned to China
from the United States in order to support the new Communist society,
used gao in phrases like sixiang ruoshi mei gao tong . . . ᗱᛇ㢹ᰃ≵᧲䗮 . . .
‘if you don’t get your thinking clear . . .’66 In the summer of 1980, during
a coordinated government effort to get writers to “liberate thought,” meet-
ings and headlines commonly used the phrase ba wenyi gao huo ᡞ᭛㡎᧲⌏
‘bring literature and art to life’. In an internal speech on literature and art
policy, Hu Yaobang used gao more often than any verbs other than shi ᰃ
‘is’ and you ᳝ ‘have’. Urging that there be national guidelines on what top-
ics to write about, Hu said (I will leave the elusive gao untranslated here):
First, we must help writers to gao planned design. Cultural depart-

ments, the National Artists Association and the Writers Associations
at all levels should do what they can to gao a few conferences in this
regard. Of course, for an artist to gao creativity is different from gao-
ing manufacturing . . . so we must not oblige literary artists to gao
collective design in a mechanical way.67
The snippet shows how pervasive gao had become in official language,
especially in contexts where officials were speaking casually. The conve-
nience of gao as a dummy verb for telling people to just “get it done” con-
tinued for many more years. In the early 1990s, when Chinese students
overseas were denouncing the June Fourth massacre, personnel in Chi-
nese consulates received instructions to find ways to discredit those who
spoke out most vociferously. Exactly how to discredit them—by rumors
of corruption, sexual misconduct, or whatever— did not matter, as long as
the discrediting worked. The instruction was simply to gao chou ᧲㟁
‘cause [them] to stink’.
The overwhelming importance of the goal in this kind of usage tends
to leave moot the questions not only of means but also of the actor. “Get it
done” can leave ambiguous the question who should get it done. This as-
66. Lao She, “Wenyi zuojia ye yao zengchan jieyue” ᭛㡎԰ᆊг㽕๲ѻ㡖㑺 [Writ-
ers must be productive and frugal, too], in Shuoshuo changchang 䇈䇈ଅଅ [Speaking
and singing] 4, no. 6 (1951), p. 9.
67. Hu Yaobang, Zai juben chuangzuo zuotanhuishang de jianghua.
Politics 273
pect of “goal orientation” is visible in the fact that nearly all Chinese politi-
cal slogans grammatically are free-floating predicates. They are clauses
that could take subjects but do not: dadao sirenbang ᠧ‫צ‬ಯҎᐂ ‘down with
the Gang of Four’, quanxin quanyi gao sihua ܼᖗܼᛣ᧲ಯ࣪ ‘pursue the
Four Modernizations with all your heart and mind’, and buyao xin you yuji,
xiangqian kan ϡ㽕ᖗ᳝ԭᚌ, ৥ࠡⳟ ‘do not harbor residual fears, look to
the future’ are all examples. Slogans in English, in contrast, do sometimes
name the subjects, for example, “Hell, no! We won’t go!” during Ameri-
can protests of the Vietnam War. But even when English-language ex-
amples do not state a subject, they differ in one crucial way from most
slogans in Chinese. Compare, for example, “down with the Gang of Four”
to dadao sirenbang. The English phrase has no subject and grammatically
cannot accommodate one. You cannot say “I down with the Gang of Four”
or “You down with the Gang of Four.” But the Chinese phrase, the free-
floating predicate, can do this. You can say wo dadao sirenbang ‘I knock down
the Gang of Four’, ni dadao sirenbang ‘You knock down the Gang of Four’,
or zanmen dajia yikuair dadao sirenbang અӀ໻ᆊϔഫ‫ܓ‬ᠧ‫צ‬ಯҎᐂ ‘all of us
together knock down the Gang of Four’. With no subject stated, Chinese
slogans are not exactly imperative and not exactly descriptive, either.
They mean something like “would that it be that [predicate].”
Yet Chinese slogans as free-floating predicates that can take subjects, if
one wants, have the flexibility (as English-language slogans usually do
not) to turn themselves into imperatives by implying subjects. The phrase
tiba nianqing ganbu ᦤᢨᑈ䕏ᑆ䚼 ‘promote young cadres’, which has been
used several times since the Mao era, is an example, as is jianjue fandui
tiaoji shangfang മ‫އ‬ডᇍ䏇㑻Ϟ䆓 ‘resolutely oppose the jumping of levels
in petitioning’, which was used in the Hu Jintao era to try to prevent
people who were petitioning the government from approaching higher
levels without going first through lower levels. Such slogans are gram-
matically the same as dadao sirenbang, yet in their political contexts their
imperative mood emerges more clearly. They carry less a sense of “would
that it be that . . .” and more a sense of “you should . . .”
Why, in this kind of case where the subject is fairly clear, does the of-
ficial language still prefer subjectless slogans? Why not just name the
subject and say “You must promote young cadres,” “You must prevent level-
jumping,” and so on? Hu Ping has observed that the reason for omission
of the subject can be a sense of embarrassment at the fact that one is

giving orders. Officials in authoritarian governments—perhaps in Chi-
na’s more than others—tend to feel better when they pretend they are not
being authoritarian. They seem to like to say “Let’s have X done, but let’s
not say that I am telling you to do it.” The point is especially clear in cases
where the message itself is potentially embarrassing. When Deng Xiao-
ping said in 1985 yibufen ren xian fuqilai ϔ䚼ߚҎ‫ܜ‬ᆠ䍋ᴹ ‘a part of the
people [may] get rich first’,68 Chinese people watched as, unsurprisingly,
well-connected Party members became that special “part” who got rich
first. Yibufen ren xian fuqilai became a pleasant-sounding catchphrase, but
it was also an order to lower-level officials to allow disparate enrichment
to take place. When the latter point needed to be made unmistakably clear
(even at the cost of some embarrassment), the word rang 䅽 ‘let’ was ap-
pended: (rang) yibufen ren xian fuqilai ‘(let) a part of the people get rich first’.
5. “Fit” as a kind of truth. In his incisive book Doing Things with Words in
Chinese Politics, Michael Schoenhals points out the special power and im-
portance of tifa ᦤ⊩ ‘ways of putting things’ in the official Chinese lan-
guage of the Communist era. He quotes an authoritative 1978 statement
in the People’s Daily that tifa “are very serious matters that must be re-
solved scientifically. . . . Where the tifa is off by one millimeter, the the-
ory will be wrong by a thousand kilometers.”69 Among the illustrations
Schoenhals offers is one from a 1965 speech by the vice president of the
Central Party School to a group of students. It is all right, the official said,
to refer to China’s current society as a you jieji de shehui ᳝䰊㑻ⱘ⼒Ӯ ‘a
society that has classes’ but not a jieji shehui 䰊㑻⼒Ӯ ‘class society’. Schoen-
hals comments that “here the distinction between socialism on the one
hand and capitalism, feudalism, and slave society on the other rested on
the presence or absence of a de.”70 Even when the issues at stake were less
weighty, an assumption prevailed that key political phrases had to be just
68. “Deng Xiaoping: rang yibufenren fuqilai” 䙧ᇣᑇ˖䅽ϔ䚼ߚҎ‫ܜ‬ᆠ䍋ᴹ [Let

part of the people get rich fi rst], http://news.xinhuanet.com /newscenter/2005-01/16
/content _2467918.htm, accessed June 29, 2012.
69. Schoenhals, Doing Things with Words in Chinese Politics, p. 7. I have altered the
translation slightly.
70. Ibid.
Politics 275
so— and that it was, somehow, a serious error if they were not. During the
brief interregnum of Hua Guofeng after the death of Mao Zedong, it was
“correct” to refer to weida de Mao zhuxi ӳ໻ⱘ↯Џᐁ ‘the great Chair-
man Mao’, jing’ai de Zhou zongli ᭀ⠅ⱘ਼ᘏ⧚ ‘the beloved Premier Zhou’,
and yingming de Hua zhuxi 㣅ᯢⱘढЏᐁ ‘the wise Chairman Hua’. The
three rulers had to be mentioned in that order if all three were mentioned;
moreover, the three adjectives that had been selected to match “scientifi-
cally” the nature of the men they referred to could not be switched around
without incurring the defi nite sense that the speaker had made a mistake.
Tifa became an extension of grammar.
Schoenhals calls the rigidity of tifa a “form of power.” This simple claim
has greater profundity than may appear on the surface. In an obvious sense,
it means that holders of power can insist that people say certain things in
certain ways and expect that, over time, thought and behavior will follow.
Aiguo, aidang, airenmin ⠅೑⠅‫ܮ‬⠅Ҏ⇥ ‘love the country, love the Party,
love the people’ has lilt, repetition, fixity, and a sense of rightness that sinks
in after a while. The power of tifa can sometimes be used to turn things
entirely upside-down. During the spring of 1989, the demonstrations at
Tiananmen were called a xuechao ᄺ╂ ‘student protest’ or a minzhu yun-
dong ⇥Џ䖤ࡼ ‘democracy movement’ up until the moment when they
were suppressed. Then, almost overnight, government media began to call
the same events a fan’geming baoluan ড䴽ੑᲈх ‘counterrevolutionary
riot’. It was a formulation most people would not have dreamed of using
only a week earlier, but now it had become a correct tifa, and people had to
begin using it, in public contexts, on pain of punishment for being “incor-
rect.” The phrase remained in place and entered textbooks and the media,
and twenty years later many people in a younger generation of Chinese
knew no better than to accept fan’geming baoluan as what had happened.
But tifa are “forms of power” not just because brutish power can stand
behind them and force their acceptance. They have an intrinsic power,
too, when they cut off alternative ways of thinking and limit the concep-
tual horizons of the people who adopt them. This phenomenon is observ-
able in many kinds of ritualized language, be it in politics, religion, song, or
elsewhere. Maurice Bloch, after studying rituals of the Merina people of
Madagascar, gained insights that, without much alteration, can apply to the
use of tifa in Communist China:
In a highly formalised or ritualised political situation there seems no

way whereby authority can be challenged except by a total refusal to
use the accepted form which is compulsory for this type of occasion,
i.e., a total refusal of all political conventions.71
Since very few people can fi nd the will for “total refusal,” formalized au-
thority comes to be accepted. Its “ceremonial trappings,” Bloch writes,
“seem to catch the actors so that they are unable to resist the demands
made upon them.”72 If authoritarian language provided for a range of al-
ternatives for how to express something, then that language still might
not be so constricting (from the point of view of users) or such a lever of
power (from the point of view of authority). But Bloch’s key point is pre-
cisely that a range of alternatives is denied. He calls the language of po-
litical or religious ritual an “impoverished” language in the sense that
“many of the options at all levels of language are abandoned so that choice
of form, of style, of words and of syntax is less than in ordinary language”
and concludes that “as soon as you have accepted a form of speaking in an
appropriate way you have begun to give up at a bewilderingly rapid rate
the very potential for communication.”73
The tradition of giving fi xed names to things in Chinese political lan-
guage has a long pedigree. Many have noted how Confucius, in the Ana-
lects, offered instruction on governance to a newly ascended ruler by say-
ing zheng zhe zheng ye ᬓ㗙ℷг ‘to govern is to rectify [names]’.74 And
Xunzi (in Burton Watson’s translation) held that:
When the ruler’s accomplishments are long lasting and his under-
takings are brought to completion, this is the height of a good gov-
ernment. All of this is the result of being careful to see that men stick
to the names which have been agreed upon.75
71. Maurice Bloch, “Symbols, Song, Dance and Features of Articulation or Is Reli-
gion an Extreme Form of Traditional Authority?,” Archives Européennes de Sociologie 15,
no. 1 (1974), p. 59.
72. Ibid., pp. 59– 60.
73. Ibid., pp. 60 and 61.
74. Lunyu 䂪䁲 [The analects], Yan Yuan 12 买⏞㄀कѠ.
75. Xunzi: Basic Writings, trans. Burton Watson (New York: Columbia University
Press, 2003), p. 145.
Politics 277
On the basis of bits of text such as these, as well as official histories in in-
tervening years, scholars have observed that the role of the official nam-
ing of things in Chinese tradition has been prescriptive as much as de-
scriptive; it is right that things be put a certain way, and be referred to in
the same way. I noted in Chapter 2 that, within some of the strains of
ancient Chinese thought, the modern notion of “true” is not as useful a
way to measure the worth of utterances as are notions of dang ⭊ ‘fitting’
or ke ৃ ‘admissible’ or ‘appropriate’.76
The power-engineering of the official language of contemporary China
has even stronger roots in modern authoritarianism, especially Soviet
Leninism. There are, however, some important differences between an-
cient and modern notions of politically correct “fit.” Perhaps most funda-
mental among these is how the two kinds of language relate to truth or
falsity of fact. In ancient China, when a statement was dang ‘fitting’, its
truth could be part of what made it fitting. Xunzi says, for example, fei wo
er dang zhe wu shi ye 䴲៥㗠⭊㗙਒᏿г ‘the person who criticizes me ap-
propriately is my teacher’.77 Here we assume that what “my teacher” says
in criticizing me is factually true, but that truth alone is only part of what
makes the criticism “fitting.” It is fitting because it is factually true and
designed to help me. In short, factual truth cooperates in the project of
making a statement dang. In modern authoritarianism, by contrast, the
project of establishing certain tifa as “fitting” is often done in rivalry with
factual truth, or even in outright opposition to it.
For example, imagine yourself as a Chinese citizen during the Hua
Guofeng interregnum in the late 1970s. If you were to say, in an official
context, weida de Hua zhuxi ӳ໻ⱘढЏᐁ ‘the great Chairman Hua’,
your statement could be criticized as incorrect, and the grounds for call-
ing it incorrect would not be truth or falsity in a factual sense. Hua
Guofeng might or might not have been “great,” however one chose to
defi ne that word, but that would not be the point, and you would not
be criticized in those terms. You would be criticized with a sentence
like zhe wenti bushi neme tifa 䖭䯂乬ϡᰃ䙷Мᦤ⊩ ‘that is not the way
this matter is put’, and the reason would be that you had produced an
76. See section “How Do Metaphors Work in Ordinary Language?,” in Chapter 2.

77. Xunzi: Xiushen 㤔ᄤ: ׂ䑿 (Xunzi: Cultivating oneself); see Watson, Xunzi:
Basic Writings, p, 24.
inappropriate match of adjective with leader’s name. Yet your “incor-

rectness” would be just as clear as if truth versus falsity had been the
criterion. “The notion of true or false,” as Bloch observes, generalizing
over a number of cases, “has been eliminated by the way the proposition
has been put.”78
In short, “fit” and “truth” have become parallel entities in China’s mod-
ern authoritarian language. Both are standards for measuring the affirm-
ability of statements, and both are positive. Each is opposite, but in a dif-
ferent way, to cuowu 䫭䇃 ‘mistaken’. This bifurcation of fit and truth
distinguishes modern official Chinese from its ancient counterpart, in
which standards like dang or ke could stand as the lone measure of a state-
ment’s affirmability. The fact that the modern official language has two
standards, not one, has produced a range of problems for ordinary lan-
guage users. I have noted Liu Binyan’s perplexity in 1958 over how to
handle “two kinds of truth.”79 (Liu, who knew Russian, was fond of the
“old joke” in Moscow that the only thing worse than living without truth
was living with two of them.)80 The resulting “language game” that citi-
zens of the People’s Republic of China became obliged to engage with
will occupy most of the remainder of this chapter.
The Language Game
During China’s Mao years, and especially after 1957, Chinese people be-
came more and more accustomed to constructing their public assertions
by asking not just “do they reflect reality?” but also “do they fit politi-
cally?” As the bifurcation of language spread more and more widely
through society, a second order of reality—an image of “the official ver-
sion of things”— seemed to take on a life of its own whenever topics with
political implications were being discussed. This second order of reality
was not idle puffery. It could have real consequences in the world, and in
78. Bloch, “Symbols,” p. 66.

79. Liu Binyan, “Listen Carefully to the Voice of the People,” p. 31.
80. A version of the joke, attributed to Alexander Fadeyev during the post-Stalin
thaw in the Soviet Union, is cited in Harold Swayze, Political Control of Literature in the
USSR, 1946–1959 (Cambridge, Mass.: Harvard University Press, 1962), p. 253.
Politics 279
that sense was itself very real. You had to deal with it in order to get cer-
tain things done, and people became accustomed to this fact. The British
scholar William Jenner, who lived in Beijing during the late Mao years,
wrote in 1978 that “in daily life our Chinese friends can cope perfectly
well with the distinction between what actually happens and the required
formulations that keep things functioning.”81 Even if you opted to avoid
the official language (which could mean having to settle for not getting
certain things done), you still could be forced to use it when others used it
against you. Some people got very good at manipulating the official lan-
guage, and benefited; others were less talented and could suffer.
Ji Fengyuan, writing about the Mao years, when political combat some-
times got heavy, has written that “revolutionary language became little
more than a weapon in a low-grade civil war.” 82 Words and phrases were
tools—verbal chisels, clubs, and levers, as it were. Ji writes of high Mao-
ism, when this kind of language use reached an extreme. For other, less
extreme times, a better metaphor for official language use is a game, espe-
cially an intellectual game like chess. The metaphor is apt for several
reasons: in both cases there are rules, tactics, and goals; in both cases
skills can be honed and a person usually gets better with experience. On
one point the analogy does not work well: games are normally entered
into freely, but China’s language game, especially during the years of high
Maoism, was not optional. You had to play. If you withdrew into silence,
the silence itself was viewed as a political position—indeed, almost al-
ways, an “incorrect” one. One had to “perform” a correct “appearance.” A
few years after Mao’s death, a young farmer told Andrew Nathan in inter-
view: “I don’t know what socialism is anymore; I only know how to talk
about it.”83 Since those years the language game has receded somewhat
from daily life, but it has never entirely disappeared and has remained
important, especially on politically sensitive topics, well into the twenty-
first century.
81. W. J. F. Jenner, “Is a Modern Chinese Literature Possible?,” in Wolfgang Kubin
and Rudolph G. Wagner, eds., Essays in Modern Chinese Literature and Literary Criti-
cism: Papers of the Berlin Conference 1978 (Bochum: Brockmeyer, 1982), p. 213.
82. Ji Fengyuan, Linguistic Engineering, p. 293.
83. Andrew J. Nathan, Chinese Democracy (New York: Knopf, 1985), pp. 175, 235.
Mark Salzman records an illuminating story from the mid-1980s about

what happened when he, a foreign teacher on a Chinese campus, killed a
rat as part of a general rat extermination campaign. People who brought
in dead rats were given rewards, and Salzman lined up for his. There was
a problem, though, because officially foreigners were not supposed to
know that there were rats in China. That information was “internal.” The
fact that a foreigner had killed a rat was true in the real world but could
not be true in the parallel world of the language game. Campus officials,
facing a dilemma, decided to deny Salzman his reward. So long as there
was no reward, they could avoid acknowledging that a foreigner knew
about rats. Salzman later asked one of his students if all of this was not a bit
silly. “Oh, of course it is very silly,” he records the student as answering.
“But the comrades in the office, like anyone else, would rather do some-
thing silly than something stupid.”84 To violate the language game in the
name of common sense would be obtuse.
In this story, campus officials were playing the language game defen-
sively. In cases where a person decides to be proactive in the game, the
reason is usually some kind of expediency: Will my move get something
done? Will it serve my interests (or those of my family, or my group)? I
mentioned in the Introduction the example of my friend in Guangzhou
who wanted a bigger apartment for his family and went to his Party leader
in 1979 with the question “Do you think we can concretize Party Cen-
tral’s policy on intellectuals?” His goal was a bigger apartment, and his
language-move was crafted with that goal in mind. It was a clear case of
the language game having consequences in the real world.
In other cases, causality between the real world and the language game
flows in the opposite direction. People sometimes make real-world changes
in order to shift the terms of play in the language game. For example,
during the prodemocracy demonstrations at Tiananmen in the spring of
1989, Fang Lizhi, the famous astrophysicist who two years earlier had
been a hero to prodemocracy students in Hefei, Anhui Province, and who
in 1989 was living on the outskirts of Beijing, deliberately stayed away
from the prodemocracy demonstrations in the city’s center. Why? Lack of
interest? No. Lack of sympathy? No. Only because he did not want the
84. Salzman, Iron and Silk, p. 201.

Politics 281
government to be able to say (even though it would have been false, and
even though government leaders would likely have known it to be false)
that he was an instigator of the demonstrations. Fang had recently been
expelled from the Communist Party under the label “bourgeois liberal.”
That label was a poor fit with reality but had been effective in the regime’s
language game. Fang knew that any association he might have with the
1989 demonstrators could allow the regime to transfer the label from him
to the demonstrators. In short, he foresaw a possible language move on
the government’s side and altered his real-world behavior in order to ob-
viate it. So he stayed away from Tiananmen Square.
There is one respect in which utterances in the official language game
are more reliable than ordinary assertions about truth and falsity. In ordi-
nary language, a statement might or might not be true, and if a speaker
means deliberately to deceive a listener, the statement is called a lie. But in
the language game, where the standard is not “Is it true or not?” but
“Does it serve the speaker’s interests or not?” the counterpart of a “lie” is
not possible. The person who hears a statement can be 100 percent sure
that “this speaker believes that this statement serves his or her interests.”
It might not be easy to figure out exactly what the specific interest is, but
one can be sure that the speaker is aiming at it. It also can happen that a
speaker miscalculates in the choice of which words will actually serve his or
her interests. But that, too, does not change the principle that the speaker
believes that the words will be expedient. To experienced players of the
language game, this “impossibility of a lie” is a useful fact. One can use it
to analyze what a speaker is doing with words.
Let us look at some more examples. When Tibetans protested in the
streets of Lhasa and elsewhere in March 2008, and when Uighurs did
likewise in Urumchi and elsewhere in July 2009, why did the players of
the language game within China’s media bureaucracies say that the dem-
onstrations had been “planned and instigated abroad”? Because they be-
lieved this to be true? Did they really believe that the Dalai Lama and
Rebiya Kadeer were pulling strings from distant places? This seems
highly unlikely. They were adequately informed on what was happening
in these places and would not have subscribed to such a far-fetched theory
in the ordinary-talk world of “is it true or not?” It is quite possible that
they did not even give much thought to questions of true versus false.
Their duty in their work was not to act as observers of events but to pur-
sue results within the flow of those events. Their goals included discredit-
ing the protesters by making them seem to be puppets; intimidating the
Dalai Lama and Rebiya Kadeer into not speaking out from overseas (by
saying, in essence, “by speaking out you only confi rm what we are saying
about you”); stimulating Chinese nationalism and presenting the Commu-
nist Party of China as its champion; and distracting the attention of people
inside China from problems of corruption, inequality, pollution, and so on
where the Party’s image was under strain. To choose as tifa that Kadeer was
the “mastermind” of the Uighur protests or that the Dalai Lama is “a jackal
clad in Buddhist monk’s robes” disregards plausibility,85 but plausibility was
not the standard the fashioners of such phrases were using. The standard
was “what will work?” and by that standard the tifa seemed right.
The distinction between truth and utility as standards for evaluating
statements in the language game struck me most clearly in 2001, in a con-
versation that I had with the distinguished Chinese writer Su Xiaokang
not long after the publication of The Tiananmen Papers, which I had
helped to edit.86 In February 2001, the Chinese compiler of those papers,
who used the pseudonym Zhang Liang, was able, through friends who
were highly placed in the Chinese government, to get access to the record
of a meeting of the Politburo of the Communist Party of China that was
held on February 8, 2001.87 At that meeting Jiang Zemin, China’s presi-
dent at the time, addressed the matter of the publication of The Tiananmen
Papers and said, according to the terse notes on the meeting that reached
me, that this leak of state secrets was the largest in the history of the Party
and showed that people of ill will, who wish to overthrow the Communist
Party and the socialist system, were hiding in high places inside the Party.
The record also shows Jiang saying that the leaking of the papers was a
naked sale of state secrets to Americans and that whoever did it has lost
qualification to be Chinese. Farther down the list— at point 5—the sum-
mary reads: “Andrew Nathan and the other two so-called American editors
85. China Central Television, July 7, 2009; Edward Wong, “Dalai Lama’s Defeat to
Be a Holiday in Tibet,” New York Times, January 19, 2009.
86. See note 65.
87. Zhang Liang email to Andrew Nathan, February 28, 2001.
Politics 283
[i.e., Perry Link and Orville Schell] are by no means pure and indepen-
dent scholars; they wear the overcoats of scholars in order to use scholarly
status while they work for the CIA.”
This statement interested me and raised a question. I have never worked
for the CIA, or any branch of the U.S. government, but here was the top
leader of the world’s largest government saying that I did. I wondered
what, in Jiang’s own mind, he might have been thinking he was doing
with his words. It seemed a rare chance to try to figure out how official
language is used at the highest levels of the Communist Party of China.
If Jiang’s words had been broadcast to all of the Chinese people, then I
could have understood them as an attempt to manipulate public opinion.
But that was not the case. Jiang was addressing a small group of his fellow
rulers at the top, in a context where, I imagined, there should have been
no barrier— at least for a question such as this one—to presenting the
truth as well as one could. So I began to wonder: Did Jiang himself believe
what he had said? If so, how had he formed his opinion? By listening to
intelligence briefings? Could it be that intelligence reports to the leader
of the largest nation on earth are as inaccurate as this? That seemed
frightening. But had Jiang perhaps not believed what he had said to the
Politburo? Would he use manipulative language on this question in ad-
dressing even his highest-level colleagues? That, too, seemed a bit fright-
ening. I decided to poll a number of my Chinese friends, including writers
and former officials who had had extensive experience with China’s offi-
cial language. “When Jiang Zemin told the Politburo that I work for the
CIA,” I asked, “how do you think it felt to him in his own mind? Did he
think he was telling the truth? Or doing something else?”
The poll results split almost evenly. Some said Jiang likely felt that he
was telling the truth. The CIA does have a record of undercover activity
around the world, and the general lore that surrounds the CIA in the sub-
culture of Chinese Communism exaggerates this record in flamboyant
ways. The Chinese intelligence ser vices might or might not have had
good information on specific people like Nathan, Link, or Schell, but the
people in those ser vices also have incentives, especially in sensitive cases,
to tell top leaders what they want to hear. Jiang could well have heard a
briefing that said Nathan, Link, and Schell were only wearing the outer
clothing of scholars.
On the other side, just as many in my poll felt that Jiang likely had not
believed the things he was saying. On politically dangerous questions it
was important for leaders, even among their top-level colleagues (who,
after all, are also their potential rivals for power) to adopt politically un-
assailable positions. The pose of blaming the CIA was one that, what-
ever else happened, no one could characterize as “soft.” Ning zuo er bu you
ᅕᎺ㗠ϡে ‘better [to err to] the left than the right’ was a reliable maxim
for hazardous times, and its value outweighed less important questions
about the accuracy of this or that detail.
In all, though, Su Xiaokang gave me the answer that taught me the
most. He began by chiding me for a wrongheaded question. Jiang Zemin,
in that situation, doesn’t even ask the question you are asking, Su said. It
doesn’t occur to him. He doesn’t care whether the words are “true or not”;
he only wants them to work. Is he telling the truth to his colleagues, or
deceiving them? Neither, said Su. In Jiang’s own view, he is merely sup-
plying to his colleagues the words that he thinks will be useful in getting
done what needs to be done. Does he expect his words to be believed?
That’s the wrong question, said Su. He expects his words to be put to use.88
If Su is right, then the record of this Politburo meeting might be
viewed an example of how tifa originate. Starting from a high level (al-
though not necessarily as high as the Politburo), tifa spread downward and
outward to the rest of government and society, where they turn into stan-
dard tools in the playing of the official language game. They might seem
rigid or awkward at first, but with repeated use and the passage of time,
people learn to accept them as normal and to use them in a variety of situ-
ations. In 1980 a literary scholar at Zhongshan University in Guangzhou
told me about his complaints about the slogan xingwu miezi ݈᮴♁䌘
‘promote the proletarian and annihilate the bourgeois’ that Hua Guofeng
had unveiled in 1976 when he was Party chairman. The scholar said he
had no problem with the theory of xingwu miezi. The Marxist ideal was
all right with him. The problem was that this four-character phrase be-
came a nasty little club in the hands of the local stewards of political
88. Based on a conversation with Su Xiaokang, spring 2001, Princeton Junction,

N.J. I have omitted quotation marks because this is a summary, not a transcript, of
what Su said.
Politics 285
correctness—Party secretaries, chiefs of Organization Departments, and

the like. These people did not understand much Marxism themselves, and
were not inclined to go read about it either. But they needed tools in order to
do what they saw as their job, which was to guan ren ㅵҎ ‘manage people’.
If a student wore bell-bottom trousers, or hair that was too long, or if a
professor taught from an unapproved book, it would be too much to go
read Marx in order to know what to do, but with a convenient little phrase
like xingwu miezi an official could get moving.
With the decline of the work-unit system in recent decades, the use of
official phrases to control people’s daily lives has also receded consider-
ably. But it has not disappeared. Special phrases in the official language
live on within bureaucratic subcultures; careers have been built around
terminology and the ideas that they enshrine, and people who live within
the subcultures cling to the vocabulary of language games in part because
their livelihoods are invested in doing so. In analyzing the repression in
Tibet in March 2008, Phuntsog Wanggyal, a Tibetan who had spent years
as a high Party official in Tibet, observed that the doctrine of fan fenlie
ডߚ㺖 ‘anti-splittism’ had deep roots in the bureaucracies that were
charged with Tibetan affairs. The very term had taken on a sort of tran-
scendent status. Careers had been invested in the idea, and no one would
think of questioning it. The Dalai Lama could explicitly say that he did not
favor fenlie ‘splitting’, but still the banner of fan fenlie had to be held high.
Too many people were already known for it, good at it, and paid to do it.
Other tifa this group depended on to justify its work included daoluan fenzi
ᤷхߚᄤ ‘trouble-making elements’ and didui waiguo shili ᬠᇍ໪೑࢓࡯
‘hostile foreign forces’. Such phrases were useful as explanations for any-
thing that might go wrong; they were also enduring justifications for the
vigilance that “anti-splittist” officials were providing to the motherland.89
Harsh terminology such as this does not dominate in the official lan-
guage, even on a “sensitive” topic like Tibet, but it is always available in
abeyance, ready to emerge as needed. A good example of “as needed” use
of formulaic language is the written confession. A ubiquitous device in the
Chinese police system, the confession is a way to establish guilt and justify
89. Interview with Wang Lixiong, March 2008; see also Wang Lixiong, “The Cry
of Tibet,” Wall Street Journal op-ed, March 28, 2008.
punishment; but it is also a tool in the language game. It is a way of draw-

ing a person into the language-world and thus (if it works) into the
thought-world of the political authority. I recall my own single experience
of writing a confession for the Chinese police. This happened after I
failed, in the fall of 1988, to register the Beijing Office of the Committee
on Scholarly Communication with China (CSCC) as a “business” within
the prescribed thirty-day period for doing so. The CSCC was not, in fact,
a business, which was the root of my original misunderstanding with the
police; learning to call it a business was part of my initiation into this par-
ticular corner of the language game. After several hours of negotiation
the officers and I settled on a confession document, which I signed, ad-
mitting that I had failed to register within thirty days. But I remember
being surprised at how much back-and-forth the police seemed to want
during the negotiation. At several points one of them would say some-
thing like “you broke rule X because you did not do Y by day Z!”— and
then look at me, eyebrows raised, as if to ask, “What do you have to say?
Do you want to rebut anything? Explain anything?” I reflected that an
American policeman, in a similar situation, would just write out a ticket,
hand it to me, and be off. But for the Chinese police, to get me involved in
the language game, and to see me perform properly within it, seemed to
be an important goal in itself. They seemed to assume that if the accused
uses language in the right way, step 1 has been achieved in bringing his or
her thought into line.90
It is not only an accused, or ordinary people in general, who can make
“mistakes,” of course. Officials can, too, and the penalties for misspeaking
can be costly to a bureaucratic career. For this reason the language game
has developed a range of vocabulary whose functions are to buffer officials
from mistakes. These vocabulary items are not official terms, not formally
approved tifa. They are natural outgrowths of the language game, invented
by its users, that eventually become widespread and standard. In the 1980s,
for example, a bureaucrat who heard a request for something— and who
90. I have written up this episode in “Jingguan qiyu: Liang ci sange yiwai” 䄺ᅬ༛
䘛: ܽ⃵ϝ‫ן‬ᛣ໪ [Strange encounters with police: Three surprises, twice], in Ban-
yang Suibi ञ⋟䱼ㄚ [Notes of a semiforeigner] (Taipei: Sanmin chubanshe, 1999)
pp. 68–72.
Politics 287
did not want to make a mistake—would often say, instead of “yes,” wenti
buda 䯂乬ϡ໻ ‘there’s no big problem’. Instead of “no,” he or she might
say youxie kunnan ᳝ѯೄ䲒 ‘there are some difficulties’, and instead of
“maybe,” might use yanjiu yanjiu ⷨおⷨお ‘we’ll study it’ or kaolü kaolü
㗗㰥㗗㰥 ‘we’ll think it over’. The hallmark of such terms is their vague-
ness; they combine easy fluency with the impossibility of being pinned
down. Essentially defensive, they are born of the dilemma of how to satisfy
(or mollify, or perhaps just cause to go away) the person making the request,
and, at the same time, avoid committing to a position that might later prove
to have been “incorrect.”
One of my own personal experiences with this kind of vocabulary
came in 1979 when I was living at Zhongshan University in Guangzhou.
I wanted to go into the city to buy socks. The Foreign Affairs Section at
the university was in the habit of providing a car whenever I went off
campus, but I felt that a car was a bit excessive for a mere sock-buying
expedition, so suggested that I ride my bicycle instead. The official at the
Foreign Affairs Section was affable and said wenti buda. This told me that,
from her point of view, there was “no big problem.” But what did that
mean? Did she approve or not? I later asked some Chinese friends, and
they said that I should take it as a yes. But it was a yes that, from the point
of view of the Foreign Affairs official, also said “I am not officially on re-
cord as saying yes.” If something had happened to me on that bicycle trip,
the official could have been in trouble if she had given explicit approval.
Later, when I heard officials answer requests with youxie kunnan, I came
to understand that this really meant no, but they preferred the vaguer
phrase because it allowed them to avoid the embarrassment of saying no
bluntly. Then I came to understand that the “maybe” responses—yanjiu
yanjiu or kaolü kaolü—were generally used to buy time in which the offi-
cial could hope that the request would simply go away or, if it did not,
could ask for guidance from above. Sensitive questions could get passed
successively to higher levels, and clogging near the top could mean that
some questions never actually reached any desk for decision. Comedians
and cartoonists satirized this pattern at considerable length in the 1980s,
often playing on the near-pun between yanjiu ⷨお ‘study’ and yanjiu ⚳䜦
‘cigarettes and liquor’, items that were useful as small bribes to get things
moving.
Vagueness has several uses in the official language. We have seen how
abstract vocabulary can create a special sense of authority. In this section,
we have seen how vagueness can be useful in avoiding bureaucratic re-
sponsibility. And we will see below how vagueness is useful in inducing
self-censorship by obliging people to guess what a vaguely stated prohibi-
tion might mean. Note in passing, though, that while vagueness in the
official language has all of these uses, the preference of ordinary people in
daily-life language is often in the other direction, toward the concrete.
Popular sayings sometimes borrow the form of official phrases and fill
them with very practical, concrete content. For example the phrase hai lu
kong ⍋䰚ぎ ‘sea, land, air’, which is an abbreviation for “navy, army, and
air force” in the official language, during the Cultural Revolution was a
popular nickname for homemade sandals cut from the rubber of worn-out
tires. Why were they called “sea, land, and air”? Apparently because they
could withstand any conditions, including water, dust, and sun. In any
case, later, when policies of “reform and opening” in the late 1970s sud-
denly made it all right to admit to material desires and overseas connec-
tions, the same phrase, hai lu kong, became a popular catchphrase for what
a young woman could ask as a condition of marriage: hai meant haiwai
guanxi ⍋໪݇㋏ ‘overseas connections’; lu was a two-speaker or four-
speaker luyinji ᔩ(䰚)䷇ᴎ ‘(stereo) recorder’, and kong was kongtiao ぎ䇗
‘air conditioning’. Other popular phrases imitated the practice in official
language of numbering things (three antis, Four Modernizations, etc.). As
an alternative to demanding hai lu kong, prospective brides in the late 1970s
could ask for the sanzhuan yixiang ϝ䕀ϔડ ‘three turns and one sound’,
meaning a watch, a bicycle, and a sewing machine (three things that turn)
plus a (sound-making) recorder.
Part of the popular enjoyment of such phrases is that they evoke the
abstraction and pomp of official phrases but bring them down to earth by
inserting concrete and practical content. In 1987, Deng Xiaoping sought a
linguistic watchword that would promote reform and opening, and con-
tinued economic growth, but also maintain Leninist political control.
These were conflicting goals, and the task of unifying them in a single
phrase was not easy. Eventually, Deng announced a formula of yige
zhongxin, liangge jiben dian ϔϾЁᖗ, ϸϾ෎ᴀ⚍ ‘one central concern and
two basic points’. The “central concern” was economic development. The
Politics 289
two basic points were (1) the Four Basic Principles (the socialist system,
dictatorship of the proletariat, Party rule, and Marxism-Leninism-Mao-
Zedong-thought), and (2) the policy of “reform and opening.” From the
Party’s point of view, this may have been an excellent phrase. But it was
far too abstract for ordinary people, who satirized it by making it concrete
in irreverent ways. A website for students returning from overseas said
a person’s “central concern should be health” ( yi jiankang wei zhongxin
ҹ‫ع‬ᒋЎЁᖗ) and two basic points should be to be xiaosa yidian ┛⋦ϔ⚍
‘on the cool and easy side’ and hutu yidian ㊞⍖ϔ⚍ ‘a bit muddle-headed’.91
A joke on the oral grapevine was even less respectful; it said the “one cen-
tral concern plus two points for attention” was only “a bikini policy.” Jiang
Zemin’s formula of “the three represents” (sange daibiao ϝϾҷ㸼) was
satirized, as we have seen, as “three [people] wearing wristwatches” (sange
daibiao ϝϾ᠈㸼), and later Hu Jintao’s “harmony” (hexie ੠䇤) was turned
into a “river crab” (hexie ⊇㷍).
If part of the popular response to China’s official language has been to
stand apart from it and even to satirize it, another side of the response— as
we have seen in several examples—has been to accommodate it, manipu-
late it, and “play the game.” This raises an important question we consid-
ered in Chapter 2 in connection with the “Whorf hypothesis”: how much
do habits of language use induce habits of thought? The question has
worried a wide variety of contemporary Chinese thinkers and observers
of China.
For me the question first arose during a visit to China in May 1973. In
those days foreigners were so rare on the streets of Chinese cities that my
group and I nearly always attracted a long train of curious onlookers, es-
pecially children, who followed us in polite silence whenever we walked
on the sidewalks. At one point, on a street in Xi’an, I turned around to see
if I could talk to some of the children in the crowd. A group of about eight
or ten, who appeared to be perhaps ten to twelve years old, formed a sort
of semicircle and showed rapt attention. I asked if they had seen foreign-
ers before, and they shook their heads. I asked what they wanted to be
when they grew up. Nobody answered. One boy let me catch his eye, so I
asked him directly, “What would you like to be when you grow up?” He
91. www.haiguinet.com /forum /viewtopic.php?t=1010621, accessed June 29, 2012.

took a step forward, stood erect, and said, with verve, wo yao dao zui jianku
de difang wei renmin fuwu! ៥㽕ࠄ᳔㡄㢺ⱘഄᮍЎҎ⇥᳡ࡵ! ‘I want to go
to the most stressful place to serve the people!’ I said “Great!” and turned
to another child. “What about you? What would you like to be when you
grow up?” The child answered, in the same tone of voice, wo yao dao zui
jianku de difang wei renmin fuwu! I asked two or three more, and each said
wo yao dao zui jianku de difang wei renmin fuwu! The first child had clearly
set an example—in what, with a foreigner watching, must have felt to
them like a performance situation. The fact that each child pronounced
the sentence in exactly the same way, without the variation of a single syl-
lable, suggested that the phrase itself had likely been taught somewhere,
perhaps in school or at a Young Pioneers meeting.
It left me wondering about what was going on in their minds. How
much did the linguistic formula that they had repeated correspond to
their inner thoughts? Had they used it to cover up various other ideas
about, perhaps, being a Party secretary, a teacher, doctor, or fi reman?
Or had the “right answer” of dao zui jianku de difang wei renmin fuwu!
temporarily— or maybe not so temporarily?— obliterated such other
thoughts? The question, which of course applies to adults as well as to
children (if not necessarily to the same degree), is how the playing of the
language game comes to affect the ways a person thinks. Ritualized lan-
guage, according to Maurice Bloch, by its very nature “both excludes ex-
planation and hides this exclusion.”92 When Bloch notes that “you cannot
argue with a song,”93 he means that songs and other ritualized language
not only tell you what to think, but, even more importantly, cut off your
ability to look squarely at things that you otherwise should want to look at.
In cases where ritualized language is backed by a strong political authority,
any inclination even to try to argue, or to look elsewhere, is powerfully
discouraged.
In looking at situations of repression in the modern world, many writ-
ers have noted that when a dominant ideological language takes hold, it
tends to pervade a culture as a whole, including its victims. Henry Fried-
lander, surveying the language of the Nazi Holocaust, finds that “not only
the perpetrators, but also the victims spoke the language of Nazi totali-
92. Bloch, “Symbols,” p. 67.

93. Ibid., p. 71.
Politics 291
tarianism,” including euphemisms for transporting, sorting, and killing

of victims. “Nazi language was a prison language,” and “both jailers and
convicts spoke it.”94 Bloke Modisane, recalling his youth in South Africa,
records instances of slipping into apartheid language— sometimes with-
out meaning to, and at other times quite deliberately, as he wished to
manipulate it to his advantage.95 Someth May shows similar examples of
adaption to, and manipulation of, the official language during Pol Pot’s
devastation of Cambodia.96 Anna Wierzbicka finds an “unofficial, under-
ground language of antipropaganda” in Soviet-dominated Poland.97 Sarah
Cook fi nds ordinary people playing a “language game” in Syria in the
early twenty-first century.98 Other examples are not hard to fi nd.
Still we can ask: how much do patterns of thinking change because of
language-game use? How much are people’s worldviews affected? We
know at least that the whole system of “thought reform” in China has al-
ways assumed that there are effects. Many veterans of the system, too,
have noticed them. The philosopher Liu Xiaogan has this to say:
The hot iron of “struggle worship” has left its mark upon the lan-
guage habits, patterns of thinking, and ways of behavior of the Chi-
nese people. In the lexicon of mainland Chinese, “struggle,” “revolu-
tion,” “rebellion,” “resolute,” and “thoroughgoing” are unequivocally
positive in their connotations; they symbolize nobility and glory. On
the other hand words like “retreat,” “negotiate,” “compromise,” “tol-
erance,” “gradualism,” and “improve,” have at times all been broadly
rejected as terms with negative connotations, terms that symbolize
shame and dishonor.99
94. Friedlander, “Manipulation of Language,” p. 111 and p. 113, n. 28.

95. Bloke Modisane, Blame Me on History (London: Thames and Hudson, 1963),
pp. 41, 59.
96. Someth May, Cambodian Witness: The Autobiography of Someth May, ed. James
Fenton (London: Faber, 1986), pp. 146–151 and 175.
97. Anna Wierzbicka, “Antitotalitarian Language in Poland: Some Mechanisms of
Linguistic Self-Defense,” Language in Society, no. 19 (1990), pp. 1–59.
98. Sarah Cook, “The Language Game: Words of Power and the Power of Words
in Syria and China.” unpublished manuscript, 2010.
99. Liu Xiaogan ߬ュᬶ, “Mao Zedong yu douzheng chongbai lungang” ↯⋑ϰϢ
᭫ѝዛᢰ䆎㒆 [Mao Zedong and the worship of struggle: An outline], unpublished
paper, 1996, p. 20.
The official language game of the Mao era, in Liu’s view, forged the lin-
guistic tools—the words, phrases, syntactic patterns—in which everyone
in the culture talked and thought.
In Hu Fayun’s novel So It [email protected], an elderly writer who was
persecuted cruelly during the Mao era worries about how Chinese mem-
ory and cultural expression—including even his own, despite his con-
scious wishes—have been colored by Maoist politics. He tells friends how
he recently found himself humming a tune and then felt shock to realize
that the lyrics were “lifting my head, I gaze at the Big Dipper; in my
heart, I long for Mao Zedong.”100 He then reflects:
Even our own most personal emotional memories are soaked in an

all-encompassing, all-pervading ideological culture. . . . Within a few
decades, they took from us our ability to express suffering and sorrow.
They took our ability to express love. What they gave us instead were
fraudulent stand-ins. . . . Even today we do not have an authentic, un-
tainted cultural vehicle with which to record our lives. . . . Other
countries have it. Even the poorest and most backward countries have
it. But the country with the largest population and the longest history
on earth does not have it—you have to admit there is something hor-
rifying about this. The long-term effect on our nation’s psyche is still
something that no one can measure.101
Other writers have observed how hard it can be to get “outside” the mind-
set of a political language game even when one wants to. Ji Fengyuan notes
that Jung Chang, the well-known author of Wild Swans: Three Daughters of
China and Mao: The Unknown Story,102 viscerally detested Mao Zedong’s
Cultural Revolution as early as 1966 but could not conceive the notion of
100. Hu Fayun 㚵থѥ, [email protected] བ⛝ @sars.come [So it [email protected]], ch.

25. This passage was censored from the version of the novel published in Beijing, but
has been restored in the translation by A. C. Clark, Such Is This [email protected]
(Dobbs Ferry, N.Y.: Ragged Banner, 2011), pp. 131–132.
101. Hu Fayun 㚵থѥ, [email protected] བ⛝ @sars.come [So it [email protected]]
(Beijing: Zhongguo guoji guangbo chubanshe, 2006), pp. 75–76.
102. Jung Chang, Wild Swans: Three Daughters of China (London: Flamingo, 1992);
and Mao: The Unknown Story (New York: Knopf, 2005).
Politics 293
blaming Mao himself until, in 1974, she happened across a piece of writ-
ing that came from outside the Maoist modes of expression.103 Similarly,
the literary critic Li Tuo, in analyzing the life-course of the famous mod-
ern writer Ding Ling, who absorbed the outlook of Maoist language at
midcareer and apparently could not extricate herself from it even as she
suffered two decades of banishment by the regime that it supported, com-
ments in summary that “once a person enters within a certain discourse,
to exit it becomes extremely difficult.”104
A number of Chinese writers since the 1980s have recognized this prob-
lem and have sought, as it were, to “climb out” of official language and its
worldview. This is not easy, because the act of climbing out of a language
still requires a language, and one must use the language one has. The effort
can become something like trying to climb out of one’s skin. In the 1980s,
for example, Li Tuo tried to abandon zui ᳔ ‘most’ in his personal language
use just because the term had been so abused during the Mao years (in
phrases like zui zui zui zhengque ᳔᳔᳔ℷ⹂ ‘most, most, most correct’ . . . ;
as mentioned ealier). But it was not easy to keep his promise to himself. Li
found himself saying, inadvertently, things like wo zui buxihuan zhei zhong
yuyan ៥᳔ϡ୰⃶䖭⾡䇁㿔 ‘I can’t stand this kind of language’.
One method for trying to exit official language has been to dive into
another. Li Tuo has held that the true breakthrough in language— and
therefore thought—in Chinese literature of the immediate post-Mao
years was “misty poetry” (menglong shi ᳺ㚻䆫) more than the more nu-
merous and widely read works of “scar” literature. Others have pointed
out that one reason why Zhong Acheng’s story “The Chess Master”
seemed so fresh and new when it appeared in 1984 was that its language
was fundamentally different from Maoist style.105 It avoided not only Mao-
ist political usage but much of the Western-influenced vocabulary and
grammar that had come into Chinese since late-Qing times. In its phrase
103. Ji Fengyuan, Linguistic Engineering, p. 290.

104. Li Tuo ᴢ䰔, “Ding Ling bujiandan: Maotizhixia zhishifenzi zai huayu
shengchanzhong de fuza jiaose” ϕ⦆ϡㅔऩ: ↯ԧࠊϟⶹ䆚ߚᄤ೼䆱䇁⫳ѻЁⱘ໡ᴖ㾦㡆
[Ding Ling is remarkable: The complex role of intellectuals in the production of dis-
course under the Maoist system], Jintian (Today), no. 3 (1993), p. 236.
105. Zhong Acheng 䬎䰓ජ, “Qiwang” ẟ⥟ [The chess master], Shanghai wenxue
[Shanghai literature], July 1984, pp. 15–35.
structure it resembled the simulated oral storytelling style of Ming-Qing

fiction, a style that may have come easily to Zhong because “The Chess
Master” and others of his stories had, in fact, originated in his oral story-
telling to friends. Other critics have observed that one attraction of mod-
ernist language to Chinese writers in the 1980s (Mo Yan, Can Xue, Ma
Yuan, Ge Fei, and others) has been, in part, its utility in breaking free from
political and social conventions of standard written Chinese.106 Ha Jin
chooses to write in English, not his native Chinese, for similar reasons.
During the protest demonstrations of the spring of 1989, the issue of
whether and how to climb out of official language generated controversy.
Intellectuals who drafted the May 16 Declaration, a sharply worded cri-
tique of how the government was handling the protests, differed over
what kind of language the declaration should use. Should the statement
include words like tongzhi ৠᖫ ‘comrade’, shehuizhuyi zuguo ⼒ӮЏН⼪೑
‘socialist motherland’, jiefanghou 㾷ᬒৢ ‘after Liberation’, and sanzhong
quanhui ϝЁܼӮ ‘third Party plenum’? To some degree these were ques-
tions about how much to identify the authors of the declaration with the
Party-state. But they were also, more deeply, questions about whether use
of official terminology draws one into a certain mind-set.
That such drawing-in can happen is clear from some of the language of
overseas dissident groups in the years following the June Fourth massa-
cre. Now driven into exile, where they were more free to state things as
they wished, some of their phrases continued to resemble officialese: wei
Zhongguo minzhu er nuli fendou ЎЁ೑⇥Џ㗠ࡾ࡯༟᭫ ‘strive hard for Chi-
nese democracy’ xianshen yu weida shiming ⤂䑿Ѣӳ໻Փੑ ‘devote oneself
to the great mission’, and so on. As I noted in the Introduction, in the sum-
mer of 1989 when refugees met to form a “Chinese Democratic Front,” one
faction argued that the use of preexisting officialese for the slogans they
were writing would make them more effective, while the opposing faction
felt that any use of such language whatsoever was too compromising.
In sum, the question of how much a political language game shapes a
person’s thought appears to have no easy answer. There can be little doubt
106. For example, Yang Xiaobin ᴼᇣⒼ, “Fan yuyan: Xianfeng wenxue de xingshi
wenti” ড䇁㿔: ‫ܜ‬䫟᭛ᄺⱘᔶᓣ䯂乬 [Anti-language: The question of form in avant-
garde literature], Wenlunbao [Hebei], July 25, 1988.
Politics 295
that it does contribute to the closing of minds, at least sometimes and to

some extent. But it is also true that people resist language games, con-
sciously and perhaps unconsciously as well. Moreover, because the games
themselves can change, people sometimes just outlast them, or at least parts
of them. Perhaps easier to analyze and to illustrate are questions about the
ways people who have been involved with the contemporary Chinese lan-
guage game have tried to cope with it or to exploit it to their own advan-
tage. I turn next to those questions. I divide them into the two broad cat-
egories of how the game is played by (1) the rulers, who on the whole have
been the originators of the game, and (2) the ruled, who have needed to
adapt and respond to the game in various ways. In making this division I
do not mean to suggest that every tactic within the game neatly falls on
one or the other side of the divide. The push and pull of the interaction is
more complex than that; my division is only a matter of organizational
convenience.
How the Game Is Played: From the Side of the Rulers
When we turn to analyze the ways people in the Chinese government

play the language game, there is a certain danger of focusing on parts and
not seeing the whole. The regime’s tactics have comprised a wide variety
and have operated at several levels: in the connotations of words, in pres-
suring individual people, in political meetings within work-units, in society-
wide campaigns, and so on. Each topic deserves attention, but to look at
them one by one runs the risk of overlooking perhaps the most important
point of all, which is to see how pervasive the whole enterprise is. Its parts
work in concert, and the whole is stronger than the sum of the parts. It is not
easy to encompass the whole picture at once, but Cao Changqing, in the
opening paragraph of his essay “Linguistic Violence: The Power of Intimi-
dation in Authoritarian Rule,” has done as well as anyone:
The most distinctive feature in the ideological project of the Com-

munist Party of China is its language. The Party mobilizes not only
its propaganda organs but all state organs to spread and implement
its language system, the purpose of which is to influence, mold and
manipulate the language and thought processes of ordinary people in

order to control how they think about certain questions and what
they say about them. In this cause the Party has mustered the print
and electronic media, book publishing, “political study” meetings,
meetings that are supposed to be non-political, “confidential chats”
with leaders, written “reflections on experience” and “reports on per-
sonal thinking,” speeches at big assemblies and discussions within
small groups, and other devices whose number and variety are con-
stantly renewed with changing times, all with the goal of forcing
people into a language system where they learn how to speak in con-
formity with the requirements of the Party: what to say, what not to
say, how to put things, and what vocabulary to use.107
The breadth of scope of the regime’s linguistic engineering is matched

by the persistence with which its key tenets are pressed. When a political
point matters, its assertion in the official language is unremitting. Zhou
Liming, who was a vice consul in charge of propaganda at the Chinese
Consulate in San Francisco until he defected after June 4, 1989, spoke as
an insider as he explained, in a speech in 1991, how sheer repetition was a
deliberate part of government strategy. As part of his explanation for how
people could come to accept the whitewash of a massacre, Zhou cited
Joseph Goebbels’s famous line “If you tell a lie big enough and keep re-
peating it, people will eventually come to believe it.”108
In a study of the Anti-Rightist Movement of 1957, Michael Schoenhals
compiled a list of terms that were used to drive home, through repetition,
the labels that indicated the ways “rightists” were evil. These included
“rightist hard-core element,” “old-line rightist,” “rightist pathbreaker,”
“vicious rightist ‘counselor,’ ” “rightist careerist,” “utterly evil rightist
element,” “rightist wolf in sheep’s clothing,” “sinister and ruthless rightist
element,” and others. Schoenhals found these phrases in official newspa-
pers that were published between July 27 and September 4, 1957. From the
same sources, he compiled another list, this one of phrases that described

108. “Communist Propaganda: The Art of Deceiving People,” China Forum News-
letter 1, no. 1 (January 1991), p. 4.
Politics 297
rightists with the word fan ড ‘anti-’. The word had appeared in abundance,
apparently in order to make it clear that rightists, whatever their descriptions,
were united in their opposition to the Party, the people, and Communism.
Schoenhals lists these terms: “utterly arrogant and utterly reactionary
and utterly despicable anti-Party element,” “anti- Communist and anti-
People conspiratorial activist,” “anti-Party buffoon-gang accomplice,”
“anti-Party careerist, traitor, and spy turned anti- Communist vanguard,”
“anti-Communist specialist,” “anti-Party ‘eulogist,’ ” “anti-Communist
‘valiant general,’ ” “anti- Communist black gang strategist,” “old-line anti-
Communist,” “anti-Party clique ‘military counselor,’ ” “anti-Communist
‘rocket gun,’ ” and “rightist element oozing anti- Communist toxin from
the depth of the soul.” 109 Later in the Mao years this sort of compulsive
political language reached even into dictionaries. Ji Fengyuan has com-
pared dictionaries published at different times in contemporary China and
finds that the end of the Mao era was a high point in the politicization of
dictionary examples. A 1976 English- Chinese dictionary illustrates the
verb “live,” for example, with the sentence “without the Party and Chair-
man Mao I could not have lived to see today’s happiness”; “attribute” is
illustrated by the sentence “we attribute all our successes to the wise lead-
ership of the Communist Party of China”; and “wherever” gets the exam-
ple “we will go wherever the Party directs us.”110
The true power of this kind of repetitive, insistent, and pervasive politi-
cal terminology shows itself not in things like dictionary examples, where
an editor has a right, after all, to pick examples arbitrarily. Strong-arm
language is most impressive when it contradicts a person’s own perceptions
and own memory— and wins anyway. For example, Beijing citizens ap-
pearing on state-run television in the days following June 4, 1989, used
“counterrevolutionary riot” to describe events that until two days ear-
lier everyone had been calling a “democracy movement” or “student
protests”— events that were likely to have been extremely vivid in their
own memories. We can only imagine the anxiety and inner turmoil that
109. Michael Schoenhals, “ ‘Non-people’ in the People’s Republic of China: A

Chronicle of Terminological Ambiguity,” Indiana Working Papers on Language and
Politics in Modern China, no. 4 (July 1994), p. 13.
110. Ji Fengyuan, Linguistic Engineering, pp. 233–234.
must have accompanied the mismatch between memory and politically

correct usage. Orwell dramatizes this problem in his famous episode in
1984 in which Winston Smith, attached to a torture machine, is asked
how many fingers are the four that are held before his face. Four? “And if
the party says that it is not four but five—then how many?”111 After stum-
bling, Winston fi nally gets the right answer: “Five! Five! Five!” A similar
dramatization lives in the lore of Chinese tradition in the episode, re-
corded by Sima Qian in The Records of History (Shiji ৆㿬), in which a
powerful imperial official tested the loyalty of courtiers by presenting a
deer to the second emperor of Qin and asking that the assembled court-
iers acknowledge it to be a horse. When the exercise was over, the official
found ways to put to death all who had called a deer a deer.112
In the official language of contemporary China, this sort of “strong-
arm lie” technique draws some of its power from the mutual reinforce-
ment of interlocking parts. Medium-sized lies fit together in ways that al-
low a big lie to take shape. Not only, for example, did unarmed protesters
in 1989 suddenly turn into “rioters” and “hooligans.” They were also—
despite their large numbers— a “tiny minority.” Worried for the future of
their own country, they took actions that were “instigated by foreigners.”
Daring to tell the truth, they were “spreading rumors.” And so on. These
various linguistic tags were not matters of accident. For a long time, ever
since the 1950s, it had been part of the job of the Communist Party’s De-
partment of Propaganda to keep media workers aware of exactly which
terms to use in characterizing politically sensitive people and events.
Labels change, and are updated, as the political needs of the Party— as
viewed by its top leaders— go through stages. Michael Schoenhals has
edited a collection of instructions that were sent to news editors in the
year 1979 through a publication called Xuanchuan dongtai ᅷӴࡼᗕ (Pro-
paganda trends). Topics such as the following are addressed, with reasons
for the specific word choices noted: “Henceforth, no longer refer to Dalai’s
Renegade Clique,” “Do Not Accuse the Viet namese of Lack of Gratitude,”
111. George Orwell, 1984, pt. 3, ch. 2.

112. Sima Qian ৌ侀䙋, Shiji ৆㿬 [Records of history] Qinshihuang benji ⾺ྟⱛᴀ
㋔. The passage is translated by Yang Hsien-yi and Gladys Yang in Szuma Chien, Selec-
tions from Records of the Historian (Peking: Foreign Languages Press, 1979), pp. 192–196.
Politics 299
“Phase Out Educated Youth,” “Use Appeasement with Care,” “We suggest
Achieve Wealth Through Diligence,” and “Criteria for Writing about the
Seamy Side of Socialism.”113
Such guidelines are ephemeral. They enter and exit political grace ac-
cording to the regime’s needs of the day. But certain others are peren-
nial. They sink in, begin to seem normal, and eventually become so well
established that people overlook the fact that originally they were rooted
in political design. An example is the word renmin Ҏ⇥ ‘people’, which
is nearly ubiquitous in the names of government-related things. The names
of the currency, the post office, the media, many buildings, and nearly
every government office at every level include the word renmin. The
original purpose of using the term was to claim that these Party-controlled
institutions “belong to the people”; after decades of use, however, the
word renmin became so routine as to have, in practice, no content at all.
By the early twenty-fi rst century, only a bookish person might notice
any irony when the People’s Armed Police suppress a protest by the
people.
The term lingdao 乚ᇐ ‘leader(ship)’ may be an even better example. In
their original meanings, the two components of this phrase, ling and dao,
are both verbs that describe the leading of a person who follows willingly,
even gratefully. A paradigmatic example of ling is what an usher does in a
church or a theater; dao, originally, is something like what a tour guide
does. Do these activities resemble what a lingdao does within China’s po-
litical system? Hardly. Overwhelmingly, people in China follow their ling-
dao because they have to, not because they choose to. Words like “author-
ity,” “ruler,” or “boss” would be more literally descriptive of what a lingdao
does, but if one puts such words directly into Chinese—using terms like
quanwei ᴗ࿕, dangju ᔧሔ, or laoban 㗕ᵓ—the rhetorical flavor changes
immediately. The lingdao themselves view such terms as “hostile to the
socialist system.” They prefer the euphemism, which saves their face, be-
cause face is one root of their power. A term like lingdao has the added ad-
vantage that it makes any objector seem, at least at the rhetorical level, to
be wayward or ungrateful.
113. Michael Schoenhals, ed. and trans., Selections from Propaganda Trends, an Or-
gan of the CCP Central Propaganda Department (Armonk, N.Y.: M. E. Sharpe, 1992).
Do words like renmin or lingdao actually affect the ways people think
and behave, or have they become—like the “dead metaphors” we consid-
ered in Chapter 2— so ordinary that they are simply standard tags? In
many cases, no doubt, they are standard tags. The People’s Daily is just the
name of a newspaper. Still, their persistence in fossil form does ground
concepts in a certain way, subtly softening the image of authoritarian rule.
The fresh, unfossilized political usages of the kind that appear in Xuan-
chuan dongtai no doubt have more immediate effects. The Soviet theoreti-
cian L. O. Resnikov, in summarizing the views of Lenin and Stalin on this
general question, wrote that “language [is] a powerful tool which can be
used to affect thoughts, feelings, and especially behaviour” and that this is
something that “Marxism teaches us.”114 This Soviet approach to language
as a tool in thought-engineering arrived in China primarily in the 1950s,
when it merged with assumptions that were deep in Chinese tradition
about speech as moral performance. The result, for China’s rulers, was an
enduring confidence that government-prescribed words could indeed pro-
duce “correct” behavior in citizens.
In her book Linguistic Engineering, Ji Fengyuan shows— at least for the
extreme case of the late Mao years in China—that this confidence was
justified. Government-sponsored terminology, applied with both force
and subtlety, did have substantial effects. Ji draws on theoretical literature
as well as China studies to sort out some of the psychological mechanisms
that went into the “engineering” of thought. She offers a useful summary,115
which includes the following. Mere exposure to terms makes a significant
difference; in addition, a validity effect takes hold when a person is aware of
being part of a large group (i.e., when everyone accepts something, then I
tend to as well). Group pressure is brought more intensely into focus in po-
litical study sessions where “correct” oral performance is required. Personal
models are important in showing people how to speak and behave correctly,
and self-perception theory can explain how going through the outward mo-
tions of believing something can lead a person into actually believing it.
Higher-order conditioning can explain how the well-known connotations of
114. Cited in J. W. Young, Totalitarian Language: Orwell’s Newspeak and Its Nazi and
Communist Antecedents (Charlottesville: University Press of Virginia, 1991), p. 211.
115. Ji Fengyuan, Linguistic Engineering, pp. 27–37.
Politics 301
certain words (like the negative connotations of a word like “poverty” or

“disease”) can spread, through repeated association, to political terms (like
“capitalism”).
“Higher-order conditioning” has a good example in the 1983 campaign
against “spiritual pollution.” The targets of this campaign were foreign
influences— everything from AIDS to rock music to ideas about human
rights. People naturally had differing opinions about whether some of
these things are good or bad, but the blanket label wuran ∵ᶧ ‘pollution’
was hard to argue against. Negative connotations “leaked” from the label
to all of the items that it covered, and anyone defending “pollution” in any
form started from a position of obvious weakness. In a parallel manner,
positive connotations could also leak from one item to another. For ex-
ample, aiguo ⠅೑ ‘patriotism’ has been overwhelmingly positive in Chi-
nese since late-Qing times, but in the Communist era it has routinely
been paired with aidang ⠅‫‘ ܮ‬love the Party’ to form aidang aiguo and has
even gone on to form the mellifluous qiyan phrase aidang aiguo airenmin
⠅‫ܮ‬⠅೑⠅Ҏ⇥ ‘love the Party, love the country, love the people’. The
phrase rolls off the tongue and gives the air, by its rhythm and its repeti-
tion, of being inexorably right. It is hard for any person using it to stop
and say, “Wait! Why is a political party heading up this phrase?” The
phrase arrives as a whole package.
It can be a problem, for the Party’s language engineers, that politically
correct phrases sometimes gain momentum on their own. The Propa-
ganda Department sometimes acknowledges this problem when it issues
orders— such as those published in Propaganda Trends—to stop using certain
phrases. When phrases sound good, and feel good, and users get used to
thinking in the terms that they set, it is not always easy to apply the brakes.
In the late 1970s, for example, the term fanshi pai ޵ᰃ⌒ ‘whateverists’—
which was used to refer to doctrinaire Maoists because they insisted on
“whatever” Mao said—turned out to be so appealing to people who were
fed up with Maoists that it continued in popular use even after officials,
embarrassed at the term’s spurt in popularity, had called for its end.
The problem of the momentum of phrases also appears when top-level
policy takes a sharp turn and language use cannot turn as quickly. During
the Cultural Revolution, for example, it was common to say that the people
needed to “struggle” against “anti-Party” people like “revisionists” and
“capitalist-roaders” who were “opposed to socialism,” aimed to “usurp

power,” and so on. Then, in October 1976, the political landscape changed
abruptly when leaders of the Cultural Revolution, now relabeled the “Gang
of Four,” were arrested. Writers and editors across China were called on
to denounce the Gang of Four, but what language were they to use? Ji
Fengyuan quotes from an English-language textbook that apparently was
hastily pulled together in Henan Province in 1977. Two young people are
in dialogue:
A: The Gang of Four Anti-Party clique wanted to usurp Party and state
power and restore capitalism. We Red Guards, never allow them!
B: That’s right. The struggle against the Gang of Four is a life-and-death

struggle between the proletariat and the bourgeoisie, between socialism
and capitalism, and between Marxism and revisionism.116
It would be harsh to conclude that the English teachers who constructed

this dialogue, who after all may have been terrified at the time, could not
think except in jargon. They may have been well aware that to say that the
former leaders of the Cultural Revolution are now, suddenly, representa-
tives of the bourgeoisie does not make any sense; they may have opted
(and who could blame them?) to play their language game in a way that
seemed to maximize their possibilities for safety. Still, the poverty of
their options in vocabulary was a problem. They apparently felt that their
only reliable store of negatively charged terms lay inside the same old bag
that they had long been using— and now were using anachronistically.
In a broader sense, the entire corpus of post-Mao “scar” literature has
sometimes been criticized for aiming to expose the excesses of the Mao
years but doing so in language and literary forms that in many ways in-
herited Maoist style.117 There is considerable merit in this criticism. Fic-
tion written in the immediate post-Mao years tends, just as during the
Mao years, to be set in a good-and-evil world in which heroes and villains
vie over political right and wrong. The continuity is worth pondering,
116. Ibid., p. 307.

117. Li Tuo, for example, makes this criticism, in “Xiandai hanyu yu dangdai
wenxue” ⧒ҷ⓶䁲㟛⭊ҷ᭛ᅌ [Modern Chinese and contemporary literature], Xindi
wenxue [New land literature], no. 6, p. 40.
Politics 303
because more than just style and form are at stake; patterns of thinking—
conceptual categories and worldview— also seem to carry over, despite
claims that a watershed has passed. This kind of subtle but powerful con-
tinuity is part of what Li Tuo has uncovered in his careful study of Ding
Ling.
As an aside on the topic of the “momentum” of political phrases in Chi-
nese, it can be amusing to note how some sayings in Chinese Communist
history have taken on an iconic status, not only in China but around the
world, even though their origins are not what most people suppose. For
example, the phrase mozhe shitou guo he ᩌⴔ⷇༈䖛⊇ ‘cross the river by
feeling the rocks’ is routinely attributed to Deng Xiaoping, who felt that
China needed a gradual and practical approach to the uncharted task of
economic reform. Deng did hold these views, and did use that phrase, but
the political use of the phrase originated with Deng’s rival Chen Yun, fi rst
in 1950 and again in a major speech in 1980.118 Deng is also famous for the
phrase buguan bai mao hei mao, zhuo dao laoshu jiushi hao mao ϡㅵⱑ⣿咥⣿,
ᤝࠄ㗕哴ህᰃད⣿ ‘it matters not whether a cat is white or black; if it
catches mice it is a good cat’. The phrase is a famous emblem of Deng’s
pragmatism: don’t be bothered with ideological labels; if something
works, then let it work. It is far from clear, however, that that white cat was
not in fact brown. A Sichuan aphorism (Deng was from Sichuan) says huang
mao, hei mao, zhiyao zhuazhu laoshu jiushi hao mao 咘⣿咥⣿া㽕ᡧԣ㗕哴ህ
ᰃད⣿ ‘brown cat, black cat, if it catches mice it’s a good cat’. We also hear
via Zhuo Lin, Deng Xiaoping’s wife, that Deng was fond of Pu Songling’s
collection of stories Liaozhai zhiyi 㘞᭟ᖫᓖ (Strange tales of Liaozhai)
and liked to bring it with him while traveling; the section of the novel
called quguai 偅ᗾ (Expelling demons) contains the sentence 咘⣌咥⣌,
ᕫ哴㗙䲘 huang li hei li, deshuzhe xiong ‘brown cat, black cat, the rat-
catcher is the more powerful’.119 Another example is the phrase zhifu
guangrong 㟈ᆠ‫ܝ‬㤷, rendered in English as “to get rich is glorious,” which
is widely accepted as a hallmark of the Deng era. Scholars have not been
118. Chen Yun, “Zhongyang gongzuo huiyishang de jianghua Ё༂Ꮉ԰Ӯ䆂Ϟⱘ䆆

䆱 [Speech at the Central Work Committee], December 16, 1980.
119. Ma Ruifang 偀⨲㢇, Yangzhou xinwenwang ᡀᎲᮄ䯏㔥 [Yangzhou news net],
March 14, 2008; www.28gl.com /html /86/t-30786.html, accessed June 30, 2012.
able to establish whether Deng invented this phrase, or indeed whether he

even said it.
Iconic phrases from Mao Zedong have also grown reputations that ex-
ceed their origins. Mao is widely believed, for example, to have proclaimed
at Tiananmen on October 1, 1949: Zhongguo renmin zhan qilai le Ё೑Ҏ⇥
キ䍋ᴹњ ‘the Chinese people have stood up’. In fact, Mao opened his fa-
mous speech that day with the words Zhonghua renmin gongheguo zhong-
yang renmin zhengfu jintian chengli le ЁढҎ⇥݅੠೑Ё༂Ҏ⇥ᬓᑰҞ໽៤
ゟњ! ‘the central people’s government of the People’s Republic of China
today is established!’ The phrase about the people standing up is authen-
tic but was delivered ten days earlier, on September 21, at a meeting of the
Chinese People’s Political Consultative Conference. Mao is also well known
outside China for having said, apparently in 1968, “women hold up half the
sky,” but this phrase in English translation (and translations into other lan-
guages) oddly omits a key word. In Chinese, what Mao is quoted as having
said is funü neng ding banbiantian ཛཇ㛑乊ञ䖍໽ ‘women can hold up half
the sky’, not that they do in fact hold it up. It does not take a feminist to
recognize that Chinese peasant women, in their many kinds of hard work,
had already been holding up half or more of the sky for several centuries
before Mao made his comment. In asserting only that an equal contribu-
tion from women is possible, not actual, Mao was hardly the feminist that
many in the West took him to be.
Many of the ways the regime “plays the language game” have to do with
force. I have referred to the “forceful lie,” “psychological conditioning,” and
political terms that have “momentum.” All these might be called, broadly
speaking, the “push” side of the regime’s language engineering, but there
is another side, a “pull” side, that is at least as important. Here the effort is
to induce willing compliance from people, to cast the world in such a way
that it seems only right that the regime should be in charge, only right that
people should want to obey, and wrong, indeed peculiar, to consider going
astray.
The fundamental principle of the “pull” side of language engineering is
to place the Party, especially the highest levels of the Party, at the moral
center of the world, so that anyone who opposes it is automatically on the
defensive. One way this is done is through implications about the sizes of
groups. The official language makes it clear that a majority—the main-
Politics 305
stream of “the people”—is always on the side of the Party. The Party’s op-
ponents are therefore always a minority, and often a “tiny” one. Sometimes—
especially in times of emergency—this centrality of the Party is stated
baldly. After the 1989 Beijing massacre, for example, when the moral image
of the Party was at a nadir, the People’s Daily flatly announced a “socialist
education” campaign in which “patriotism means love of the socialist
People’s Republic of China. To carry out socialist education is to carry
out love of the Party.”120 In more normal times this kind of overt state-
ment about the lovability of the political center is avoided, though, be-
cause there is a certain loss of face in having to make the claim in such a
straightforward way. It is preferable to imply the moral centrality of the
Party.
Such implication is achieved in several ways. One, as noted, has been
the use of the word renmin ‘people’ in the names of innumerable institu-
tions and offices involving Party rule. The usages become so ordinary, so
accepted, that it seems axiomatic that the Party and the people are at one.
Any opponent of the Party-plus-people is automatically (1) a minority, (2)
removed from the center, and (3) morally inferior. These three notions
(minority status, displacement from the center, and moral inferiority), al-
though analytically separable, are assumed in the terms of the official
language to imply one another. A rival group to the mainstream one in
the top leadership, for example, can be called a jituan 䲚ಶ ‘clique’. In the
official language a jituan might be powerful, and thus dangerous, but it is
always decentered and can never be truly popular. Chiang Kai-shek, even
as president of China, headed a jituan. Liu Shaoqi and Lin Biao, in their
times, had jituan. Mao Zedong’s wife, Jiang Qing, was at the political and
moral center until Mao died, when suddenly she, too, headed a jituan.
People who are as fully decentered and wrong as a jituan, but not as
high-ranking or powerful, tend to be only fenzi ߚᄤ ‘elements’— a term
the Chinese Communists borrowed from Soviet-era Russian. There have
been dizhufenzi ഄЏߚᄤ ‘landlord elements’, youpaifenzi ে⌒ߚᄤ ‘right-
ist elements’, zichanjiejifenzi 䌘ѻ䰊㑻ߚᄤ ‘bourgeois elements’, bufafenzi
ϡ⊩ߚᄤ ‘illegal elements’, and many other kinds. A blanket category
called simply huaifenzi ണߚᄤ ‘bad elements’ has included everything
120. Renmin ribao Ҏ⇥᮹᡹ [People’s daily], June 2, 1991, p. 1.

from criminals to drug addicts to dissidents to gays. After the 1989 mas-
sacre, there were liusifenzi ݁ಯߚᄤ ‘June Fourth elements’. A small num-
ber of fenzi words do not carry heavy negative coloration. Zhishifenzi ⶹ䆚
ߚᄤ ‘intelligentsia’, which is a loan word from the Soviet Union, is prob-
ably the most obvious example.
We have noted how the terms yixiaocuo ϔᇣ᪂ ‘a small bunch’ and ji-
shaoshu ᵕᇥ᭄ ‘tiny minority’ have been used since the early years of the
People’s Republic to label groups that rulers wish to denigrate. These
terms have continued in use into the twenty-first century to refer to Ti-
betans, Uighurs, Falun Gong adherents, democracy advocates, and oth-
ers. In a Tibet Daily editorial on the disturbances in Lhasa in March 2008,
jishaoshu and fenzi were combined in the double-barreled negative label
jishaoshu bufafenzi ᵕᇥ᭄ϡ⊩ߚᄤ ‘tiny minority of illegal elements’.121
Examples such as this show the implicit connection in the official lan-
guage between the ideas of tiny, evil, and potent. The smallness of the op-
position reinforces the implication that it is evil, and if something is both
tiny and evil it must be fairly potent and therefore is dangerous. Note that
these ideas in their origins— outside the language game— do not have
these connotations. Normally, if things are as disparate as the word fenzi
suggests, as well as extremely few in number, they would be something like
needles in haystacks, or marbles in the ocean. There would be no reason to
get very upset about them. But in official rhetoric, a “tiny minority” can
be enough to warrant a call for hundreds of millions of people to tigao
jingti ᦤ催䄺ᚩ ‘raise vigilance’ and jianjue fandui മ‫އ‬ডᇍ ‘resolutely op-
pose’ the small group. Because of the connection between tiny, evil, and
potent, phrases such as yixiaocuo and jishaoshu have the paradoxical effect of
belittling and magnifying their object at the same time. But the paradox
does not matter in the playing of the language game. Indeed it helps, be-
cause the point of terms like yixiaocuo and jishaoshu is not, after all, to seri-
ously estimate the size of an opposition but to establish that it is evil and
dangerous. The implicit contradiction of “little group” and “big problem”
helps to do this because it shows how strong the evil little group can be.
Phrases like “small bunch” and “tiny minority” have the added advan-
tage of leaving plenty of room for the average citizen, at the receiving end
Politics 307
of the rhetoric, to choose to join the majority. If troublemakers are a small

minority, then the choice for you, the average citizen, is easy: come to the
center and join the mainstream. If you do, the safety of knowing that the
authorities approve is reinforced by safety of numbers. A person who joins
the mainstream does not have to defend the choice.
Official language can create the illusion of a mainstream where origi-
nally none existed. On any given subject, before the official language
game begins to address it, people might have a natural variety of views,
and many might have no view at all. Is long hair attractive? Is Falun Gong
good for one’s health? Is the Dalai Lama a sincere man? Without official
intervention, it is unlikely that Chinese people would have much consen-
sus on such questions. But once it becomes an official view that long hair
suggests spiritual pollution (as happened in 1983), that Falun Gong is an
evil cult (after 1999), or that the Dalai Lama is a splittist (several times,
but especially in 2008), then people everywhere in China have an incen-
tive to join in these views, and a “mainstream opinion” is manufactured.
This sort of opinion is artificial in its origins, but eventually it can turn into
a very real thing—if not “opinion” in the original sense, then at least recog-
nition of what the officially “correct” view is. In extreme cases, a manufac-
tured view can represent opinion that differs radically from what original
opinion was. In the late Mao years, for example, farmers in many parts of
China wanted to dismantle “people’s communes” and return to family farm-
ing. But as long as Mao was alive, the official view, accepted as mainstream,
was that people loved the people’s communes. In a society that lacked press
freedom, and where “incorrect” expression of any kind was suppressed in
public, it could happen that a large majority of people might in fact have a
consensus but have no way of knowing that this was so. If numbers alone
were what mattered, then, on the question of whether to dismantle the com-
munes, Mao and his lieutenants, the creators of the “mainstream view,” in
fact were the small minority.
The claim of moral correctness at high levels in the Party-state is fur-
ther enhanced in the official language by abstract vocabulary of the kind I
have analyzed above under “lexicon.” Party Central does not think things
over, it “applies consideration” ( jiayi kaolü ࡴҹ㗗㰥); unlike ordinary people,
who simply do things, Party Central “adopts measures” (caiqu cuoshi 䞛প
᥾ᮑ). On matters of yuanze ॳ߭ ‘principle’, the Party cannot but jianchi
മᣕ ‘be resolute’, and so on. Such vocabulary, consistently applied, gen-

erates a lofty, august atmosphere. It is the verbal counterpart of the physi-
cal arrangements in the great meeting halls where major Party meetings
take place: huge glossy red banners bearing gold-colored Chinese charac-
ters, long tables covered in red cloth, bearing neatly arrayed white tea-
cups, one before each chair, spaced as if by a machine. All of this, the
verbal and the nonverbal together, is paichang ᥦഎ ‘ceremonial extrava-
gance’. Paichang has deep roots in Chinese culture and has taken on an
especially exaggerated form in Communist China, due in part to borrow-
ings from Soviet culture. Paichang is not just for show; it contributes im-
portantly to face, prestige, and political power. To be able to present a
decorous display of language is an important way of claiming political
power and attracting respect for it.
The assumption is well established in Chinese culture that correct lan-
guage use, when properly internalized in a person, leads to personal “cul-
tivation,” which gives rise to appropriate moral behavior, which in turn
provides legitimate authority to rule. Such things as a person’s calligra-
phy, or the ability to write a good poem, have been seen as outward signs
of moral character. These assumptions, despite Maoist campaigns to “de-
stroy the four olds,” have survived well into the Communist period and
thrived even within Maoist culture itself.
Mao supplied his own calligraphy to many institutions, where it stood
as signs both of his political patronage and of the receiving institution’s
fealty to him. Recipients included the People’s Daily, the Guangming Daily,
and many provincial newspapers, as well as universities, publishing houses,
railway stations, and—as Richard Kraus points out in his fascinating book
about calligraphy and politics in the People’s Republic Brushes with
Power— even such obscure places as the mosquito nets at Fujian Teachers
College.122 In 1978 students at Zhongshan University wanted to start a
student-run literary magazine but were having trouble getting political
permission from campus authorities to do so. When the central literary
commissar Zhou Yang came to Guangzhou for a visit, the students went
122. Richard Curt Kraus, Brushes with Power: Modern Politics and the Chinese Art of
Calligraphy (Berkeley: University of California Press, 1981), p. ix.
Politics 309
to see him and asked if he would be so kind as to donate his calligraphy for
the front cover of their inaugural issue. Zhou did, and with that imprima-
tur the students were able to launch their magazine. Around the same time,
Marshal Ye Jianying donated his calligraphy for the front gate of Zhong-
shan University, an act that symbolized his broad endorsement of the
university as well as its loyalty to him.
The political symbolism of poetry has been important as well. As noted
in Chapter 1, Mao Zedong coveted his image as poet, and Zhou Enlai’s
admirers used countless poems to praise Zhou. In 1978 Chen Yi, a former
mayor of Shanghai and foreign minister of the People’s Republic, pub-
lished a collection of his poems in two thick volumes.123 (A significant
difference between Chinese political culture and that of the modern West
emerges if one imagines, just as a thought experiment, what the reaction
might be if a leading Western politician were to announce publication of
a collection of poems.) Since the 1980s, the publication of poetry and
donation of calligraphy by high-ranking Communists has declined, but
the underlying assumption that formal political language carries an air of
exaltation and is related to the personal power of leaders has remained
very much in force. Such language is still common at formal meetings and
in the media, where it continues, in its presentation and delivery, to un-
derscore the connection between proper language, morality, and political
legitimacy.
One consequence of this persisting tradition is that whenever a prob-
lem arises—whenever something appears that might detract from the
paichang that supports power—it is difficult to face the matter squarely.
To face it squarely, or to name it plainly, can entail a loss of face, might
damage the aura of exaltation, and such damage, in the terms of the po-
litical culture, might be detrimental to power. Problems are therefore
dealt with indirectly—by looking askance, using euphemisms, or the like.
For example, after Wei Jingsheng wrote his famous poster in 1978 call-
ing for democracy as a “fifth modernization,” he was arrested and sent to
prison not for the challenge that he had issued, direct address of which
123. Chen Yi 䰜↙, Chen Yi shici xuanzhu 䰜↙䆫䆡䗝⊼ [Annotated collection of the
poetry of Chen Yi] (Beijing: Beijing chubanshe, 1978).
would have embarrassed China’s rulers, but for “passing military secrets
to a foreigner.” Many others who have challenged the ruling authority by
raising questions about corruption, fraud, environmental damage, and
other such topics have been punished on charges that are beside the point,
and often false, but convenient to the rulers from a cosmetic point of view.
To argue over the real issues would detract from the aura that supports
power.
One of the most common of these charges-of-convenience has been
“revealing state secrets” (applied with no sense of the irony of the result-
ing association of things like corruption and fraud with state secrets). It is
one of the most feared charges, because it can carry heavy penalties. Other
such charges have been rumor mongering, various kinds of sexual misbe-
havior, and corruption. (He Qinglian has documented cases in which
journalists who expose corruption are charged with corruption.)124 The
blind lawyer and dissident Chen Guangcheng, who defended women who
were resisting forced abortion, was imprisoned in 2006 after charges of
assembling a crowd to disrupt traffic.125 The conflation of sexual misbe-
havior and political misbehavior, in which the former is stated when the
latter is meant, has been recurrent. In the 1983 campaign to “oppose spiri-
tual pollution,” long hair, premarital sex, drug use, and political dissidence
all counted as spiritual pollution. Since then, intermittent campaigns to
yan da Ϲᠧ ‘strike hard’ have aimed at “hooligan crime”—robbery, rape,
kidnapping, and so on—but have prominently included, especially in the
early 2000s, sao huang ᠿ咘 ‘sweeping up the yellow [i.e., pornography]’.
This expression has been used as a broad label for various things that au-
thorities identify as obscene, including dissident politics. It, too, is huang,
and can be swept away during a yan da campaign. In May 2009 the Minis-
try of Industry and Information Technology announced that every per-
sonal computer sold in China beginning July 1 of that year would need to
include “Green Dam Youth Escort” software designed to avoid “the poi-
124. See He’s accounts of the cases of Gao Qinrong and Zhang Chongbo, quoted in
Perry Link, review of He Qinglian, Zhongguo zhengfu ruhe kongzhi meiti Ё೑ᬓᑰབԩ
᥻ࠊၦԧ [How the Chinese government controls the media], New York Review of
Books 52, no. 3 (February 24, 2005), p. 37.
125. Philip P. Pan, “Chinese to Prosecute Peasant Who Resisted One- Child Pol-
icy,” Washington Post, July 8, 2006, p. A12.
Politics 311
soning of our youth’s minds by harmful information on the Internet.”126

The order characterized “harmful information” as disu zhi feng Ԣ֫П亢
‘vulgarity’, but Chinese Internet users were quick to complain that disap-
proved political opinion was covered as well.
The foregoing examples of euphemism might seem clumsy and simple-
minded, but their functions in repressive politics are far-reaching and
worth analyzing in more detail. In some cases, the point of a euphemism
is to disguise the identity of the person who causes others to suffer. This
can be as simple as the naming of a scapegoat. For example, when horrific
accounts of the Great Leap famine were reported to Mao and other top
Party leaders in the summer of 1960, and when it could not have been more
obvious that Mao and others at the top were responsible, Mao said: “many
landlords, rich peasants, counterrevolutionaries and bad elements have
seized political power and committed evil acts.”127 Other techniques of dis-
guise are more subtle. We have seen above, for example, how Chinese po-
litical slogans often omit subjects. In the yan da ‘strike hard’ campaigns of
the 1990s, exactly who was instructing the police to be severe? It was conve-
nient that the grammar of Chinese slogans allows this item to remain un-
specified. No one had to be named as possible perpetrator. (A similar effect
in political English is achieved by use of the passive voice: “It has been de-
cided that . . .”) Even when a doer is named, the connection to past wrong-
doing can remain fuzzy. In 1980, for example, China’s official media began
to list Kang Sheng, the much-feared head of the secret police under Mao, as
one more post-Mao villain, alongside Lin Biao and the Gang of Four. But
the references to Kang, even in blaming him, skirted the delicate question
of his role of running the secret police. Instead, the standard reference to
him was lilun quanwei ⧚䆎ᴗ࿕ ‘authority on theory’.
Political euphemisms, in China and elsewhere, are used not just to mask
perpetrators but, even more commonly, to obfuscate what they have done.
Orwell shows how this can be done through abstraction, or what he calls
“sheer cloudy vagueness”:
126. Human Rights in China, “Chinese Government Orders Computer Manufac-

turers to Pre-install Filtering Software,” www.hrichina.org/content /301, June 8,
2009, accessed June 30, 2012.
127. Yang Jisheng, Tombstone: The Great Chinese Famine, 1958–1962, translated by
Stacy Mosher and Guo Jian (New York: Farrar, Straus and Giroux, 2012), p. 61.
Defenceless villages are bombarded from the air, the inhabitants

driven out into the countryside, the cattle machine-gunned, the huts
set on fire with incendiary bullets: this is called pacifi cation. Millions
of peasants are robbed of their farms and sent trudging along the
roads with no more than they can carry: this is called transfer of popu-
lation or rectifi cation of frontiers. People are imprisoned for years with-
out trial, or shot in the back of the neck or sent to die of scurvy in
Arctic lumber camps: this is called elimination of unreliable elements.
Such phraseology is needed if one wants to name things without call-
ing up mental pictures of them.128
Nazi language avoided “calling up mental pictures” by using euphemistic

words like Aktion ‘action’. There were Grossaktion ‘big actions’, Kleinaktion
‘small actions’, Einzelaktion ‘individual actions’, Studentenaktion ‘student ac-
tions’, Allgemeine Befriedungsaktion ‘general pacification actions’, and others,
and they referred to things like roundups, incarcerations, deportations,
and murder.129 A similar euphemism in Mao’s China was jieji douzheng 䰊
㑻᭫ѝ ‘class struggle’, which became a dreaded phrase, especially at the
height of the Cultural Revolution. The word douzheng originally means
“struggle,” so it might not seem entirely a euphemism. But we should view
it that way, because the intensity and variety of the cruelty involved went
well beyond ordinary “struggle.” It included ransacking homes, burning
books, public humiliation, beatings, torture, eye-gougings, killings, de-
mands that families pay for the bullets that killed their loved ones, and
even— as Zheng Yi has carefully documented— such extreme atrocities as
the ritual eating of bits of flesh of murdered “class enemies.”130
The term jieji douzheng dwindled quickly from view during the post-
Mao years, but other well-known euphemisms lived on. Laodong jiaoyang
ࢇࡼᬭ‫‘ ݏ‬education through labor’, which lasted well into the twenty-
first century, was even more feared than laodong gaizao ࢇࡼᬍ䗴 ‘reform
through labor’. Both were euphemisms for labor camps, but “education
128. Orwell, “Politics and the English Language, p. 173.

129. Friedlander, “Manipulation of Language,” pp. 110, 113.
130. Zheng Yi 䜁㕽, Hongse jinianbei ㋙㡆㋔ᗉ⹥ [Red memorial] (Taipei: Huashi
wenhua gongsi, 1993), pp. 2–51.
Politics 313
through labor” was actually worse than “reform through labor” because it
was an administrative measure, not a criminal punishment—which meant
that it had no fi xed terms of ser vice. One received (was grateful for?) “re-
education” as long as authorities said one needed it. In 2009, when Chi-
nese government censors closed certain websites for the sake, they said, of
President Hu Jintao’s ideal of “harmony” (hexie ੠䇤), Chinese netizens
turned the euphemism on its head by inventing the use of hexie as a transi-
tive verb. To write that a website bei hexie 㹿੠䇤 ‘has been harmonized’
came to mean that authorities had closed it down.
Political euphemisms not only turn unpleasant things into
abstractions— or, as Orwell says, make them “cloudy”—but also prettify
them. “Education” sounds better than “forced labor,” and “harmony” of
course sounds better than “coercion.” I have noted how metaphors of
“cleaning” or “purifying” were attractive to several twentieth-century
authoritarian regimes as they carried out unspeakable acts of cruelty. It is
worth probing the reasons why the perpetrators of violence or repression
like to prettify. What exactly does this achieve? Superficial pleasantness is
one reason, of course, but there are others.
One is that even if a person believes that a certain violent act is right, the
same person is often aware that other people, observing, may not agree. In
these cases, prettification serves the purpose of deflecting criticism others
might bring, even if the doer, left to his or her own conscience, might not
think that euphemism is necessary. A chilling example of this sort of
mind-set is that of Nazi Special Ser vices militiamen who, as Simon Wie-
senthal has reported, felt no need for euphemism because no evidence of
their killing would survive to be observed: “we will destroy the evidence
together with you,” they said to prisoners. “And even if some proof should
remain and some of you survive, people will say that the events you de-
scribe are too monstrous to be believed.”131
In other cases, third-party observers are not relevant. The need for
prettification can arise wholly inside perpetrators of outrageous acts
when their consciences or moral upbringing confl icts with something
else—an order from above, pressure from peers, or aroused passions within
131. Simon Wiesenthal, The Murderers Are Among Us, quoted in Primo Levi, The
Drowned and the Saved (New York: Vintage, 1989), pp. 11–12.
themselves— and they are led to do things that their better selves would
rather not have done. In such cases, euphemism helps the perpetrator
whether or not other people are looking. Primo Levi offers a number of ex-
amples of this kind of inner conflict in people whom he observed running
Nazi death camps. “The person who has inflicted the wound,” Levi writes,
“pushes the memory deep down, to be rid of it, to alleviate the feeling of
guilt.”132 In the organization of the camps themselves, Levi shows, a num-
ber of arrangements made it possible that people would not have to look
squarely at what was happening: the Special Ser vices (SS) personnel, who
ran the crematoria, “were kept rigorously apart from the other prisoners
and the outside world”; the SS itself was staffed at the bottom levels by
Jews who themselves would soon be killed— an arrangement that allowed
the transfer of guilt for the act of killing away from those who ordered it
and onto people who were victims in any case, and who, in addition,
would soon be dead and therefore incapable of making a report.133
Another reason for degrading human beings before killing them, in
Levi’s analysis, was that it could subtly allow people on the killing side to
feel that they were killing things that were less than fully human. If the
inmates of death camps are called “vermin,” then killing them is killing
vermin and not, after all, murder. Should this be called “euphemism”? It
might seem an odd term, since euphemism is supposed to make something
sound good and calling people vermin does the opposite. But the essence of
euphemism is still at work here, because the function of the word “vermin”
is still to deflect a square look at what is happening. It makes it possible for
the perpetrator of a vile act to view the act as less vile. Azumo Shiro, a Japa-
nese soldier during World War II, noticed a similar utility of animal meta-
phors in recalling the mind-set he and his fellow soldiers had during the
Nanjing Massacre in China in 1937. They had gotten used to referring to
the Chinese as pigs, bugs, and other animals. Azumo remembers an in-
stance when they raped a Chinese woman and then shot her in the back as
she ran away. They killed her because Japanese army regulations forbade
rape (even though, on another level, the practice was tolerated and even
encouraged). The soldiers wanted to erase any evidence of the rape. “Per-
132. Levi, Drowned and the Saved, p. 24.

133. Ibid., pp. 51–52.
Politics 315
haps when we were raping her, we looked at her as a woman,” Azumo wrote,
“but when we killed her, we just thought of her as something like a pig.”134
That facile conversion within the mind—from human being to pig, in
order to cushion a perpetrator’s conscience—has a grisly parallel in Zheng
Yi’s account of the politically induced cannibalism that happened in Wu-
xuan County in Guangxi during the summer of 1968. There was no official
policy to promote cannibalism at the time, but, in the feverish competition
to demonstrate “final victory” over “class enemies,” activists sometimes
disemboweled their victims, removed their hearts and livers, and chal-
lenged villagers to participate in eating bits of the internal organs. Zheng
Yi shows how the ghastly political rituals that resulted led to psychologi-
cal conflicts for villagers who wanted to— or at least felt pressure to—
participate in the political campaign but at the same time could not face
the prospect of eating human flesh. Leaders in one village, according to
Zheng, found a middle way through this dilemma. They ordered that hu-
man flesh and pork be cut into equal-sized pieces and boiled together in
a large pot in the village square. They suspended the pot above eye level
while villagers passed by to receive one piece of meat each. All villagers
could then say “I have shown a firm class standpoint” but could also say,
perhaps only to themselves, “It is possible I have not eaten human flesh.”135
The village leaders had figured out a way to use ambiguity in order to
avoid looking squarely at something awful. Their technique resembles
that of fi ring squads in twentieth-century America that loaded all rifles
but one or two with blank cartridges, so that no man on the squad would
have to carry the burden of knowing for sure that he had killed some-
one. This sort of avoidance reflex apparently is a deeply rooted human
response.
In yet another version of the avoidance response, the doer of an unpleas-
ant deed can pretend “I am doing not exactly this, but something else.” In
China the device has many examples in the area of the political criticism
of literature. In 1979, for example, three young playwrights wrote the
play What If I Really Were? It features a young man who pretends to be the
134. Quoted in Iris Chang, The Rape of Nanking: The Forgotten Holocaust of World
War II (New York: Penguin, 1997), pp. 49–50. See also p. 218.
135. Zheng Yi, Hongse jinianbei, pp. 90– 91.
son of a high official in order to get a ticket to a theater show.136 His ruse
succeeds, but then, to his dismay, he finds he cannot extricate himself
from it. Others—most of them ambitious officials—take him at his word,
shower him with food, gifts, and invitations, introduce him to potential
girlfriends, and follow up their favors with requests for reciprocity. In its
context, the play was a bold exposé of official corruption. It was an im-
mediate hit in China, and its written text circulated widely underground
even after its stage performance was banned. Then, in 1980, in a number
of Propaganda Department meetings that were followed by austere arti-
cles in the press, the play was denounced.137 It was criticized for having
encouraged young people to sympathize with a deceitful impersonator, a
“swindler.” To audiences who had loved the play, the criticism was odd.
The whole point of the play had been not to encourage deceit, hypocrisy,
bribery, sycophancy, or string-pulling but to satirize and denounce those
very things. Audiences had perceived these vices not in the ne’er-do-well
young impersonator but in the menagerie of officialdom that surrounded
him, and that reminded them of officials they had encountered in their
own experience. Yet, viewed against its own goals, the focus of the official
criticism on the swindler was brilliant. It drew attention away from the
sore point—the play’s devastating comment on officialdom— and toward
an issue where officialdom had not only face but the upper hand. Sym-
pathize with a trickster? Socialist China may have its problems, but we need
socialist solutions, not bourgeois-individualist ones, said the Party-run
media. This was the lesson readers of the criticism were to take home—
even though in the play itself there was no suggestion that the protago-
nist’s little trick was any kind of solution, either for society or for himself.
At the play’s end (and the end of a literary work, in traditional Chinese
storytelling as well as in socialist realism, is where the “lesson” appears),
the young man is crushed.
136. Sha Yexin ≭৊ᮄ, Yao Mingde ྮᯢᖋ, and Li Shoucheng ᴢᅜ៤, “Jiaru wo shi
zhen de?” ‫؛‬བ៥ᰃⳳⱘ? [What if I really were?], Qishiniandai [Seventies] (Hong
Kong), no. 1 (1980), 76– 96.
137. Hu Yaobang led the criticism in Zai juben chuangzuo zuotanhuishang de jianghua.
The leading written criticism was by Chen Yong 䰜⍠, “Cong liangge juben kan wenyi
de zhenshixing he qingxiangxing” ҢϸϾ࠻ᴀⳟ᭛㡎ⱘⳳᅲᗻ੠ؒ৥ᗻ [The realism
and the tendencies of literature and art as seen in two plays], Renmin ribao Ҏ⇥᮹᡹
[People’s daily], March 19, 1980, p. 5.
Politics 317
The official criticism of What If I Really Were? implicitly warned Chi-

nese audiences that if you sympathize with a trickster then you, not just
the trickster, are a problem. You should reflect on your own thinking and
make appropriate changes before you get into more trouble. This general
way of inducing self-censorship has been fundamental in the Chinese
Communist movement ever since its years in Yan’an in the 1940s. Com-
pared to their Soviet predecessors, the Chinese Communists have used
censoring mechanisms that rely less on physical and bureaucratic controls
(where a censor blots out or changes what a writer writes) and more on
psychological engineering (where an authority induces a writer to con-
form to what the authority clearly wants). The self-censorship system is
grounded in the principle of fear. At times, especially during the Maoist
campaigns of the 1940s–1960s, this was an intense and immediate fear: a
person could be taunted, ostracized, raided at home, and in other ways
“struggled.” More normally, though, and especially in the decades since
Mao’s death, the fear that anchors self-censorship has not been a clear and
present sense of panic. It is more like a dull, well-entrenched leeriness that
people who deal with the censorship system have gotten used to and have
accepted as a normal part of their environment. In daily life it is so routine
as to go largely unnoticed, but its controlling power is impressive nonethe-
less. It operates almost in the manner of a traffic light: people control
themselves because of it, and a person who chooses to disobey seems not so
much courageous as stupid.
While fear anchors this system, linguistic vagueness plays a major role.
Abstract words name things one is supposed to avoid, even though they
are hard to defi ne: rightism, revisionism, ultraleftism, spiritual pollution,
bourgeois liberalization, splittism, anti-China tendencies, and many others.
The evaluative components of the meanings of such terms could not be
more clear: no one should dream of being a splittist, a purveyor of pollu-
tion, or the like. But the denotative meanings of such terms are always
vague. What exactly is “rightism”? Telling unapproved truths in print, as
exemplified by Liu Binyan and other famous “rightists”? Owning a degree
from the University of Chicago or having family members who live in In-
donesia, as other cases show? What about “spiritual pollution”? Exactly
what is it? Listening to Hong Kong music? Wearing long hair? (How
long? Am I safe at eight centimeters? Twelve?) It has always been impos-
sible to know such things, and it has always been a part of the deliberate
design of the control system that people must make their own guesses at
the answers. As soon as I am made to guess, the system has successfully
turned me into my own censor: if I’m not sure about the rule on long hair,
maybe I should be conservative and keep mine at six centimeters.
In addition to vagueness about the exact borderlines of what constitutes
an error, vagueness over the punishments that are applied if one commits
the error also creates anxiety, and this, too, strengthens self-censorship. If
I am judged to be guilty of spiritual pollution, will I be criticized at school?
Punished by a bad job assignment when I graduate? Or worse, perhaps,
be expelled? These questions were impossible to answer during the Anti-
Spiritual Pollution Campaign in 1983. It is also well known that punish-
ments are applied irregularly. People from well-connected families can get
away with more, on a wide range of issues, than ordinary people can. So
where— each person must ask— does my case fall?
It is worth trying to list the several ways linguistic vagueness serves
official purposes. Here are some:
1. Vague warnings frighten more people than precise warnings do. If I don’t
know what “spiritual pollution” or “bourgeois liberalization” is, then it could
be virtually anything; therefore, it could be what I am doing; therefore, I
pull back. (Result: many people begin to censor themselves, often unneces-
sarily.) If I could know exactly what the rules were, then I could tailor my
speech and behavior precisely to avoid mistakes, which would then leave me
relatively free and relaxed to speak and behave in all other areas. (Result:
many fewer people would pull back.) Clarity serves the purpose of the cen-
soring state only when it wants to silence a specific person or action; when it
wants to intimidate a large group, vagueness works much better.
2. A vague threat pressures an individual to curtail a wider range of activity
than a precise prohibition does. If I don’t know exactly what it is that the au-
thorities are prohibiting, then I have to guess for myself. If I am living in
fear of being punished for my possible mistake, then my imagination of
what topics could be risky might run far and wide, perhaps well beyond
anything that the authorities originally had in mind, or even knew about.
(They might not know, for example, that I listen to Australian radio— and
they might not even care; but I know, and fear that they might care, so I
cut back on my listening.)
3. Vague charges allow for arbitrary targeting. It is probably a rule of hu-
man nature that people who exercise arbitrary power like to disguise the
Politics 319
real reasons for their actions, because this allows people to look respect-
able even while doing whatever they want. In a culture like China’s, where
the leader’s “face” represents his morality, which in turn affects his politi-
cal legitimacy, the desirability of appearing to be moral is especially
strong. The need for such pretense only increases as a particular leader’s
moral behavior worsens. In this context, the availability of vague and
even self-contradictory guidelines can be very useful. When a guideline
says “long hair is spiritual pollution” at the same time that some people
do have long hair and are not bothered, it is the authority alone, in the
privacy of his or her own mind, who can decide whether or not to punish
a given person for violating the rule. The same space for arbitrary
power is opened when a rule says “internal- circulation materials must
not be made public” at the same time that many such materials are
openly available in bookstores. The authority can punish a given person
for whatever reason the authority wishes and still be able to point to an of-
ficial rule as justification for the punishment. China’s constitution itself
illustrates this useful flexibility. It provides that citizens have freedom
of speech, of assembly, and of the press. But its preamble also sets down
the inviolability of Communist Party rule, Marxism-Leninism-Mao-
Zedong-thought, the dictatorship of the proletariat, and the socialist
system. The huge space between these two contradictory poles (both of
which, by the way, are poor descriptions of the actual patterns of life in
China) gives leaders im mense room to be arbitrary while still claiming
to be legal.
4. Vague accusations are useful in eliciting information from detainees. When
a person is arrested or detained in a political case, police normally do not
explain to the accused what the accusations are. Formal charges are vague
or sometimes entirely absent. “You yourself know the reason,” the accused
is often told. It is then up to you, the accused, to “earn lenience” by “show-
ing sincerity”—which means opening up and telling everything that you
know. The police often begin by saying they already possess an exhaus-
tive amount of information on your activities and that the purpose of the
interrogation is not to get information but to measure your sincerity by
observing your confession. But this can well be a lie. Usually the point is
precisely to extract new information, which can then be used against ei-
ther you or someone else. Clarity about the original accusation would
obviously render this tactic useless.
5. Unclear or contradictory instructions can be used in order to shift responsi-

bility for mistakes away from leaders and onto the people who are receiving the
instructions. We have seen above two examples of this tactic: telling writers
in the late 1970s to “liberate their thought” but also to adhere to “Marxism-
Leninism-Mao-Zedong-thought,” and telling soldiers at Tiananmen in
June 1989 to clear the square by 6 a.m. “absolutely” but to do it without
bloodshed. In such cases, the authority figure can “have it both ways”: if
something goes wrong in either this direction or that, the giver of the order
can always avoid responsibility by pointing to the side of the instructions
that said the mistake should not have happened. The blame falls to the
person who was trying to carry out the contradictory orders. The efforts
of such a person to avoid both a Scylla and a Charibdys naturally intensify
self-monitoring and self-censorship.
6. Occasionally, in high-profile political cases, the purpose of vagueness is to veil
the identities of the people who are being attacked. In 1967, for example, the
official Chinese press began to use the phrase “the top person in authority
taking the capitalist road” to refer to Mao’s chief rival, Liu Shaoqi. In this
case, the point was not really to withhold information (it was no secret
who the “top person in authority” was) but to provide a layer of face-
saving indirection: to be naming him without naming him. Similarly in
1974, in a campaign to “Criticize Lin Biao and Confucius,” Mao and the
people close to him were using Confucius as a stand-in for Zhou Enlai,
who was their political target at that time. In the Confucius case, most
ordinary people did not know that Zhou was the behind-the-scenes tar-
get, and hence were left with the very considerable puzzle of why two
people as different as Lin Biao and Confucius were being lumped to-
gether. But the elite knew what was going on, and that, for the purposes
of Mao and his associates, was enough. To them it probably seemed just
an incidental bonus that people at all levels, regardless of what they
knew, began self-censorship of anything that might seem bad about
Confucius.
In sum, there are many ways language itself has been a tool of the Chi-
nese Communist Party’s efforts to maintain its control of Chinese soci-
ety. The use of both forceful, domineering language and vague, abstract,
or indirect language may seem in some ways an opposition, but there is
unity between these practices in their pursuit of the common goal of shap-
Politics 321
ing the ways people think, speak, and behave. Still, we must not claim too
much for language alone. Although it plays an important role— and a role
whose power often goes unnoticed—it is only part of a wider array of
techniques the Party uses to maintain its power. Those other techniques
include, for example, financial incentives (You want your business to
thrive? It might be best to cooperate) and appeals to patriotism (You want
to be proud of your country? Then be proud of “your” Party). During the
Mao era, when the commune and work-unit systems were dominant (the
work-unit system remained important through the 1980s as well), many
kinds of daily-life questions—housing, education, medical care, permis-
sion to travel, marry, even buy a bicycle or sewing machine— all depended
on one’s political credentials in the eyes of Party officials. By the early
twenty-first century these controls had largely disappeared, and nothing
so comprehensive has replaced them. But Party leaders have shown flexi-
bility in finding alternate means of applying pressure. One of these, for
example, is the use of China’s borders to exact compliance from citizens
who dare to criticize the government. The technique is used regardless of
the direction in which people want to cross the border. You want a pass-
port in order to travel abroad? Be obedient, or you won’t get it. You are an
exile who wants to come back to China for your mother’s funeral? Sign
the “confession” we have prepared for you, or better yet give us informa-
tion on overseas dissidents, and we will let you back in. And in the end,
behind all of the myriad ways “soft” techniques are used to shape people’s
incentives, there stands the police—uniformed police, plainclothes po-
lice, cyber police, quasi-military People’s Armed Police, and others, in-
cluding simply thugs for hire. This book is about language and cannot
address the full range of the Party’s tools and techniques of social control.
That would need another book, or several. My focus will now turn to
some of the ways citizens in the People’s Republic, including ordinary
people as well as protesters, have responded to the official language game.
How the Game Is Played: From the Side of the Ruled
On the whole, Chinese people have used adaptation more than resistance
in response to their government’s political use of language. They have
found ways to adjust to the language game, defend themselves within it,
and use it to advance their own interests. For people who stand on the
weaker side in a governing relationship, such an approach is only prudent.
The Chinese Communist Party’s special use of language for political
purposes originated during the 1930s and 1940s, before Communist ac-
cession to power in China as a whole in 1949. During the early 1950s,
many in the Chinese population were still largely unaware of the language
game. To most, slogans like “serve the people,” “oppose corruption,” and
“join cooperatives” had straightforward meanings that implied idealistic
purposes. There seemed no reason to feel distance from such terms or to
second-guess what they might mean. As we have seen earlier, it was only
in the late 1950s, with the Anti-Rightist Movement and the Great Leap
Forward, that the bifurcation between official language and ordinary lan-
guage became pervasive and began to impinge on daily life in ways that
obliged almost everyone to adjust to “two kinds of truth.” Now people
had to learn how, when speaking officialese, to set aside ordinary norms
of truth and falsity and make words “fit” prescriptions of what was
“correct.”
People began to use ordinary language to talk about how to handle the
official language: “You can just say X” or “If they say Y, you can say Z,”
and so on. In this way of speaking, X, Y, and Z, however high-sounding,
become mere tools. Victor Erlich noticed essentially the same mechanical
use of language in the Soviet Union, where it emerged a few decades be-
fore it came to China. Erlich writes that “what had been a rhetoric of
crude yet genuine ideological commitment becomes a threadbare ratio-
nale for crass personal materialism.”138 Statements can be true or not, far-
fetched or not—that sort of thing does not matter— so long as a job gets
done or a goal gets reached. A detailed example might best show how this
works. In 1988 my work with the Committee on Scholarly Communica-
tion with China involved arranging for a group of American scholars to
do fieldwork in Zouping County, Shandong Province, where a group of
local county officials was in charge of hosting them. This work allowed
138. Victor Erlich, “Post-Stalin Trends in Russian Literature,” Slavic Review 23, no.
3 (September 1964), p. 407. Erlich is commenting on “the Drozdovs,” a phrase made
famous by the character Drozdov in Vladimir Dudintsev’s novel Not by Bread Alone.
Politics 323
me some close-up views of how these officials handled their government’s

language game, and I later wrote this about it:139
As part of their compensation [for hosting the Americans], the offi-

cials wanted my assistance in purchasing an American Jeep. In order
to get a loan from Chinese banking authorities, they needed a formal
statement from my office that the Jeep would be for use by the Amer-
ican scholars in the rural county. When I pointed out that this was
not strictly true, they explained that the statement was only a “for-
mality,” and made it clear that it would be quite unfriendly of me not to
write the statement. So I did. Then a few weeks later, with their loan
secured, the same county officials came back asking for another for-
mal statement, this one certifying that the Jeep would not be used by
Americans. If the Jeep were for Americans, they explained, regula-
tions required that it be bought in foreign exchange currency rather
than Chinese currency. I decided to call the Bank of China to point
out the directly contradictory regulations and to ask for guidance.
This decision perplexed the rural officials. “Why do that?” they asked
indignantly. “Just write another statement. It doesn’t matter.” To them,
contradictory statements were a mere nuisance that should not impede
the task of acquiring a Jeep.
Eventually the officials did get their Jeep. Before I got around ei-
ther to calling the Bank of China or to writing a statement, they made
a separate appeal to my Chinese office assistant, who, acting on the
principle that I was a generally well-intentioned person who of course
would want to help, wrote a statement for me and affi xed the official
CSCC chop to it.
If this incident illustrates how people use ordinary language to talk

about the official language—how to manipulate it to get things done—the
reverse could also happen. People could use official talk to focus on things
that had been said informally. This happened often during the Mao era
139. Perry Link, Evening Chats in Beijing: Probing China’s Predicament (New York:
Norton, 1992), p. 187. Certain other examples below are also drawn from Evening
Chats in Beijing, pp. 181–190.
(and sometimes during the immediate post-Mao years as well) and could
be frightening. If person A at a formal political meeting reported that per-
son B had, in an unguarded moment, made politically incorrect statement
X, then statement X, originally not intended as an item in the language
game, entered the game willy-nilly. At the height of the Maoist frenzy in
the late 1960s, even something as informal as talking about one’s cat (mjo
⣿, a near homonym for Máo ↯) could, if overheard by a dim-witted or
mean-spirited neighbor, be brought to a political meeting, where it could
cause one great harm.140 By the end of the Mao era, an aphorism had arisen
in the unofficial culture: “Choose your personal friends outside of your
work-unit.” If you made friends and spoke informally with people at work,
you would always have to guard your words, even in informal contexts.
That was no way to relax. Even if you trusted your friends to keep confi-
dences, you would also have to trust that your friendship would never, for
any reason, lapse—because people have memories and could report you
later. That was no foundation for friendship.
On the other hand, the work-unit was the right place to learn how to
manipulate required terms and phrases. During political study sessions,
this kind of language use was, in any case, required, and to get good at it
was seen as a worthwhile skill. A person who was quick to sense a political
drift and to articulate it with a bit of eloquence could be said to zhudong
fucong Џࡼ᳡Ң ‘take an initiative in obeying’. In ordinary language such
a phrase might seem contradictory, because zhudong suggests an actor with
a mind of his or her own, while fucong suggests the opposite. But in a Mao-
era political meeting, zhudong fucong was a virtue. (Whether or not people
viewed it as a virtue outside the political meeting is another, more complex
question.) A similar example is jiji kaolong ⿃ᵕ䴴ᢶ ‘actively fall in [with the
Party line]’. In ordinary language, jiji suggests an activist and kaolong a fol-
lower. Jiji kaolong thus produces the odd notion of “lead in following,” but
in a formal political study session it was unquestionably a smart thing to do.
The value of being “active” or “taking an initiative” in such contexts
deserves some analysis. At bottom, following is the main point in a politi-
140. Cao Guanlong ᳍‫ݴ‬啭, “Mao” ⣿, in “San’ge jiaoshou” ϝϾᬭᥜ, Anhui wenxue
[Anhui literature] 1 (1980), pp. 17–31. Translated by John Berninghausen as “Cats,” in
Perry Link, ed., Roses and Thorns: The Second Blooming of the Hundred Flowers in Chinese
Fiction, 1979– 80 (Berkeley: University of California Press, 1984), pp. 123–130.
Politics 325
cal study session; why stress “initiative”? Such words show, fi rst of all,
energy, enthusiasm, and an example for others to imitate (i.e., follow).
More subtly, but at least as important, they underscore the fiction that the
activist is acting independently, guided by his or her own enthusiasm. In a
system that claims to be a “democratic” dictatorship within a “people’s”
government, it is important to preserve the fiction that the people are in-
deed the ones who come up with ideas such as “oppose revisionism,” “op-
pose bourgeois liberalization,” and “root out rightist tendencies to reverse
verdicts.” In Mao-era political meetings, it was common to refer to wo
geren de kanfa ៥ϾҎⱘⳟ⊩ ‘my personal opinion’ even though the phrase
was seldom used actually to lift the curtain on a person’s private thoughts.
It had two other uses: (1) to provide a buffer in case the opinion was criti-
cized (one could always say that my geren de kanfa was mistaken and now I
see that the Party’s view is correct), and (2) to maintain the democratic
fiction that all of us, in this Party-run political meeting, are expressing
our individual views. The general practice of using words to claim con-
formity to democracy has extended well beyond the Mao period. Official
language in the early twenty-first century continues in several ways to
rely on the pretense of pursuing alignment with majority opinion.
The foregoing are several examples of how people adapt to the official
language game. Some of the commonest ways they avoid or resist it de-
pend on the language “bifurcation” that I have noted in several contexts
earlier. To the extent that officialese and ordinary talk are different and
operate in separate spheres, each can maintain its own outlook, at least to
some extent. Alexander Yashin describes Russian farmers in the Soviet era
who were very good at manipulating official language during Party meet-
ings but then, in informal contexts—as if shedding a layer of clothing—
spoke in a very different mode and were full of lively complaints about the
officiousness of higher-ups.141 Similar behavior has been noted by many
Chinese writers, among them Gu Hua, who describes rural officials in
Hunan during the Cultural Revolution this way:
For years they had trained themselves to be two-faced, trained them-

selves in double-talk. . . . By day, they attended classes to “struggle
141. “Levers,” in Hugh McLean and Walter Vickery, eds., The Year of Protest, 1956: An
Anthology of Soviet Literary Materials (Westport, Conn.: Greenwood, 1956), pp. 193–210.
against the self and combat revisionism,” then they smuggled home
plexiglass and imported stainless steel to make a table-lamp or four-
poster double bed. During meetings they contrasted past bitterness
with present joys; but once home they griped because the commune’s
co-op had no kerosene, or even candles. They passed the day drink-
ing tea, smoking, reading the newspaper [all clichés for “official” be-
havior], and stressing the long-term importance of suppressing re-
visionism; the evenings, however, were spent jockeying for promotion
or fi nding jobs for their relatives or places in the army or in college
for their children.142
Even when speaking in a wholly informal context, like that of these

commune members during their evenings, the sliding back and forth
from unofficial to official language could be useful when special cir-
cumstances arose. The official language was generally more formal and
imposing, and could be useful if someone felt a need to make a claim of
authority or to issue a threat. A vivid example in my own memory is from
1980 in Guangzhou, when I saw a young woman from a politically well-
connected family enter an argument over whether liangbei ϸ‫‘ ס‬twofold’
in Chinese means “double” in English or properly should be viewed as
“triple” because yibei ϔ‫‘ ס‬onefold’ can also mean “double.” The debate
grew fierce, and then it grew personal. The well- connected young woman
suddenly decided to call her opponent a baizhuan ⱑϧ ‘[politically] white
expert’. A few years earlier, during the Cultural Revolution, this had
been a fearful label. It had serious implications. In 1980 it seemed a bit
out of place, but still, because of psychological associations that remained
alive in the minds of everyone present, and because of the political pedi-
gree of the speaker, it delivered a chill to the room. It had no intellectual
connection to the question of “double” versus “triple.” It was simply a
linguistic club that the young woman had decided to borrow for use in
combat.
The most common way of avoiding the official language game is to take
refuge in informal contexts where ordinary language is the norm. But
142. Gu Hua, “Futuling,” trans. Gladys Yang, in Pagoda Ridge and Other Stories (Bei-
jing: Panda Books, 1985), pp. 131–132.
Politics 327
people also achieve avoidance or resistance in many other more specific

ways. A review of some of them follows. I have chosen the examples in
part for their intrinsic interest and in part for what they reveal about the
nature of the language game.
One common tactic of resistance has been to play dumb. Who made a
certain anti-Party remark? One can say “I don’t know.” And where did you
hear it? “Forgot.” Who else is in your group? “What group?” And so on.
This tactic arose during the Mao era, and by the 1980s seems to have taken
on some fairly standard forms. One ready answer to the question “Where
did you hear it?” came to be “in the public toilet”—because no person could
be expected to identify a fellow toilet-user any more specifically than that.
In 1988 in Beijing I heard the following joke, which seems to be an em-
broidery on the standard “heard it in the toilet” line. It is a dialogue be-
tween an interrogator (I) and a citizen (C):
I: What did the person look like?

C: Don’t know.
I: How can you not know? What clothes was he wearing?
C: No clothes.
I: Hunh?
C: We were in a public bath.
I: I see. Well then, how tall was he?
C: Don’t know . . . we were sitting down in the water.
I: What did his face look like?
C: Hard to say. There was steam rising from the water.
The need to use such ploys has declined since the Mao era but has by no
means disappeared. Students used essentially the same tactic during the
crackdown after the Beijing Massacre in 1989, and signatories of Charter
08, the citizens’ manifesto for human rights and constitutional democracy
that appeared at the end of 2008, used it throughout 2009 to respond to
interrogation about their movement. In the early twenty-first century, the
Internet replaced the public bath or toilet as the safest place to say, un-
traceably, that one had heard something.
Although the “playing dumb” tactic takes the official language at its
face value and answers it at the same level, other methods make use of the
distance between official and unofficial language. We have seen Wu Zu-
xiang’s description of how intellectuals had learned, by the end of the
Mao era, to read the newspaper “upside down”. They understood that the
report of a heroic rescue of a few people in an earthquake, a fi re, or a mine
collapse should be read as a report that a much larger number of people
had likely perished. From an official positive X, one could reliably infer an
unofficial negative Y. The inference could work the other way around,
too: a negative X could be grounds to infer a positive Y. For example, when
Mao and his associates criticized Deng Xiaoping in 1976, they distributed
samples of Deng’s speeches in which he called for more intellectual free-
dom, more emphasis on economic growth, and less class struggle. People
were supposed to read these speeches and denounce Deng. And Deng was
indeed denounced—widely, loudly, and insincerely. Witnesses to the
events have attested that at least some of the denunciation was mere play-
acting, and that, in their own minds, some of the denouncers identified
with the very views they were denouncing.143 The campaign to denounce
Deng in fact was spreading his views. Essentially the same thing hap-
pened to the dissident astrophysicist Fang Lizhi after 1987, when he was
expelled from the Party as a “bourgeois liberal.” His speeches advocating
democracy and human rights were sent to universities as fanmian jiaocai ড
䴶ᬭᴤ ‘reverse teaching materials’—meaning materials to instruct stu-
dents in what not to think. Two years later, Fang concluded that this
teaching-materials campaign had spread his ideas to many more students
than he could ever have reached by himself. The attempt to use his speeches
“in reverse” had itself worked in reverse.
The double entendre that the official language makes possible and the
satire to which it opens the door have been exploited primarily by China’s
intellectuals. But others in society—farmers, workers, and others—have
also made use of double meanings. In the 1950s, for example, the Party
needed to recruit large numbers of soldiers. As an antidote to traditional
prejudices against soldiers, slogan-makers in the Party revised the tradi-
tional Chinese aphorism haotie bu da ding, hao nan bu dang bing ད䪕ϡᠧ
143. Jung Chang, Wild Swans: Three Daughters of China (London: Flamingo, 1992),
pp. 654– 656; Ji Fengyuan, Linguistic Engineering, p. 302.
Politics 329
䩝, ད⬋ϡᔧ݉ ‘good iron doesn’t go to make nails, good men don’t go to

be soldiers’ as hao nan cai dang bing ད⬋ᠡᔧ݉ ‘only good men go to be
soldiers’. Rural families across China were willing to accept this new po-
litically correct formulation, but many gave it an inversion of meaning. For
them, “good men” did not mean only “men who will make good soldiers”
(be obedient, willing to die, etc.). It could also mean men who found a way
to get out of the villages and into the cities. City life had long been viewed
as more attractive than rural life, but permits to live in cities were not easy
to get. Now, under the new slogan, joining the army was a way a “good
man” could get to a city, and maybe even establish a beachhead there for
the rest of the family.
In some cases, double meanings are used not just to play the system for
personal advantage but to engage in verbal combat with officialdom. In
1980, for example, Party officials held a meeting with practitioners of
xiangsheng comedians’ dialogues as a way to warn them to be less sharp in
their comment on corruption, special privilege, and other politically sen-
sitive problems. Some of the performers (and these people were, after all,
masters at wordplay) fought back with double entendre. Jiang Kun, a ris-
ing xiangsheng star at the time, announced that the interests of the Party
and of xiangsheng were yizhi ϔ㟈 ‘at one’. The phrase was officialese and
could not be gainsaid. He went on to say that xiangsheng “speaks what is in
the minds of the masses.”144 This, too, was politically correct verbiage and
impossible to oppose. But it harbored the second-level meaning that “the
masses” were angry at corruption, special privilege, and all those other
topics whose repression was precisely the reason officials had convened the
meeting in the first place. Officials then felt they had to answer Jiang, but
they could not contradict him without contradicting their own language.
Another example of combative use of double entendre (to choose but
one among a plethora) came in the late 1970s in Guangzhou with a play
on the word tiaozheng 䇗ᭈ ‘adjust’. Skilled people in urban work-units—
universities, hospitals, museums, libraries, and so on—had had no salary
raises, and had often suffered reductions, during the final decade or more
of the Mao years, and now the government was promising to tiaozheng
gongzi 䇗ᭈᎹ䌘 ‘adjust salaries’. A member of the Guangdong Political
144. Xiangsheng chuangzuo zuotanhui bangongshi, “Xiangsheng chuangzuo zuo-

tanhui jianbao” p. 1.
Consultative Conference punned on the official phrase to complain that

so far the raises had been tiny. We only got weitiao ᖂ䇗 ‘fine-tuning’, she
said. Another wag invented a pun that spread widely in the Guangzhou
area. The government gave us kongtiao ぎ䇗 ‘air conditioning’. The same
two Chinese characters, pronounced the same way, can mean “empty ad-
justment,” that is, raises of zero.
But double entendre is only a part of what language bifurcation makes
possible. The gap between what the Party says it is doing and what it actu-
ally does is a pervasive fact, with examples that are grounded in fundamen-
tal institutions. The Chinese constitution guarantees freedom of assem-
bly, yet people can be put in prison for joining unapproved groups. Political
cases are handled in “trials,” but only after authorities have decided what
the outcomes will be. An official can accuse a writer of “bourgeois indi-
vidualism,” when the writer’s actual offense was to expose corruption.
The large space between pretense and practice that examples like this es-
tablish leaves much room for people to use the language game as a tool for
resistance.
In early 1989, for example, shortly after his speeches on democracy
spread as “negative teaching materials” in reverse, Fang Lizhi applied for
permission to travel from Beijing to Taiwan. The Chinese government
had recently announced, as a part of a delicate game of words and face
with the Taiwan authorities, that it would now allow Chinese intellectuals
to visit Taiwan. Mainland authorities, of course, were not thinking of
people like Fang when they announced this policy, and Fang knew this,
and knew that an application from him would certainly be rejected, but
decided to play a round of the language game anyway. His goal was not to
get to Taiwan but to expose pretense. And why do that? Because pretenses
such as these were struts that supported the regime’s face and claims to
legitimacy. Fang’s gesture was small, but its purpose was not.
Later in the spring of 1989, student protesters at Tiananmen made
more explicit use of the gap between pretense and actuality in officialese.
They were protesting corruption, special privilege, and authoritarian rule,
all of which were “sensitive” topics whose mention could be risky. Repeat-
edly, they used official language as cover. They held up a banner that read
“Citizens of the People’s Republic of China have freedom of speech, publi-
cation, assembly, formation of groups, demonstration, and protest (Article
Politics 331
35 [of China’s constitution]).” They also drew words— carefully selected

for relevance—from China’s national anthem: “Arise, ye people who would
not be slaves. . . . The most perilous hour for the Chinese nation has
arrived,” and so on. The Communist anthem “The Internationale,” trans-
lated from the French, provided an even larger trove of stirring, double-
edged words:
Arise, wretched of the earth, convicts of hunger . . .

let us wipe the slate clean . . .
the world is about to change . . .
we are nothing, let us be all, this is the fi nal struggle . . .
there are no supreme saviours . . .
let us save ourselves . . . so that the spirit be pulled from its
prison . . .
hideous in their self-glorification, kings of the mine and rail—
have they ever done anything other than steal work?
Understood one way, such words were unspeakably seditious. But how
could the authorities order people to stop singing a classic Marxist anthem?
(What they could do—and did—was to play the language gambit in reverse
by claiming that the students felt a basic loyalty to the socialist motherland
because they were, after all, singing a Marxist classic.) At Beijing Univer-
sity, students invented perhaps the most sarcastic double entendre in all of
the song lyrics that emerged that spring. They revived the tune—widely
popu lar in the 1950s—“Without the Communist Party There Would Be
No New China” but now sang it, poker-faced, without comment on whether
“New China” was a good or bad thing.
After the June Fourth massacre, when officials across China held meet-
ings to promote the correctness of the “necessary military action,” people
found ways to turn even applause into sarcasm. After a political speech to
students at Peking University, an official was greeted with silence. Then a
pop! as one pair of hands came together, alone. Then a pause and another
random pop! here, and another there, and then a faster but irregular pat-
tern of isolated popping. What could the authorities do? Find the people
who had clapped and charge them with clapping? Ban applause after po-
litical speeches? At a required political meeting at a major publishing house
in Guangzhou, people were asked to discuss Deng Xiaoping’s report on

why the June Fourth crackdown was correct. Inwardly, this group of well-
informed intellectuals felt rage at the massacre. But what could they
say? After an awkward silence, someone thought of calling out hao! ད
‘good!’ and then repeated the word several times, ending with an extended
h-h-h-h-a-a-a-o-o-o-o-o-o! It was biting irony, of unmistakable intent.
In the early twenty-fi rst century, people working in China’s weiquan
㓈ᴗ ‘support rights’ movement have continued the tradition of exploit-
ing pretense in language. A small but influential group of “rights lawyers”
use a strategy of following the rules in the Chinese legal code and show-
ing how officials themselves violate these rules. It is well known on all
sides that political cases are decided by Party leaders, not by legal rules,
yet the pretense that the law is supreme remains vital to the Party’s claims
to legitimacy. What rights lawyers say, in essence, is “all right, if you pre-
tend that you follow the rules, then so will we, and we will hold you to
them.” The strategy can yield two kinds of benefit. Occasionally, it can
actually do some good for individual defendants. But, more important, it
challenges the entire system by exposing its hypocrisy— and does so in a
way that is theoretically unassailable because it depends only on the fol-
lowing of rules. (One reason why the regime so often avails itself of spuri-
ous charges— corruption, traffic obstruction, etc.—is that the actual of-
fenses of rights lawyers and other activists are that they follow rules, but
the written law does not include the “crime” of rule-following.) Vaclav
Havel has put the matter very well in writing about Czechoslovak ia in the
1970s, and it is perhaps relevant to quote Havel because Chinese rights
lawyers in the 2000s have taken direct inspiration from him:
A persistent and never-ending appeal to the laws—not just to the

laws concerning human rights but to all laws— does not mean at all
that those who do so have succumbed to the illusion that in our sys-
tem the law is anything other than what it is. They are well aware of
the role it plays. But precisely because they know how desperately the
system depends on it— on the “noble” version of the law, that is—
they also know how enormously significant such appeals are. Because
the system cannot do without the law, because it is hopelessly tied
down by the necessity of pretending the laws are observed, it is com-
pelled to react in some way to such appeals. Demanding that the laws
Politics 333
be upheld is thus an act of living within the truth that threatens the
whole mendacious structure at its point of maximum mendacity.145
In June 2009 the poet and literary critic Liu Xiaobo, who a year and a half
later was awarded the Nobel Peace Prize, was arrested and charged with
“inciting subversion of state power” because of his support for Charter 08.
In defending him, the famous rights lawyer Mo Shaoping made exem-
plary use of the strategy Havel describes. Liu had originally been de-
tained on December 8, 2008, on the authority of a document on which the
space for the item “suspected of the crime of . . .” was left blank. To leave
this blank was illegal, Mo pointed out. Then Liu was held for six months
under “residential surveillance.” This, too, Mo showed, was illegal be-
cause Liu was not kept at his own residence and was denied access to his
family and his lawyers. When the maximum period allowed by Chinese
law for “residential surveillance” had expired, Liu continued to be held
without charge. This, too, was illegal, Mo showed. When formal charges
were announced two weeks later, the police told Mo that he could not
defend Liu because he, Mo, had also signed Charter 08. This also was
outside the law.146 Did Mo believe that his pointing out of these various
legal infractions would lead to redress within the system? Probably not.
Was his inveterate appeal to law a mode of support for the regime? Hardly.
(One could view it this way only if the regime itself were fully and sincerely
supporting the law.) In defending Liu Xiaobo, Mo was playing a language
game whose proximate goal was to help Liu if possible but whose larger,
long-term enterprise was to push China further toward rule of law by ex-
posing the hypocrisy in the way things currently were.
Mo and other rights lawyers use their real names, but most resisters and
language-game players do not. Most are ordinary citizens—purveyors of
shunkouliu, oral jokes, graffiti, text messages, tweets, blogs, and so on—
who hide behind anonymity for the sake of freedom and safety. In the
first decade of the twenty-first century the Chinese government passed
145. Vaclav Havel, “The Power of the Powerless,” in Living in Truth (London: Faber
and Faber, 1989) p. 98.
146. A 2009 open letter by Mo Shaoping in defense of Liu Xiaobo appears in Perry
Link, Liu Xiaobo’s Empty Chair: Chronicling the Reform Movement Beijing Fears Most (New
York: New York Review E-books, 2011), pp. 52–56. See also letters from Zhang Zuhua
and signers of Charter 08, New York Review of Books 51, no. 13 (August 13, 2009), p. 76.
laws banning the use of pseudonyms on the Internet, and, although the
government can track down almost anyone if it really wants to, with hun-
dreds of millions of people using the Internet, it is not hard for most neti-
zens to speak anonymously. Messages often list no author or only a pseud-
onym. Like jokes in the West, material often passes from person to person
often without either the passer or receiver knowing who the original au-
thor was. But while authorship and readership remain fluid and largely
obscure, the principle of exploiting the gap between official pomp and
actual life remains constant. For example, in a dormitory restroom at Capi-
tal Normal University in early 2009, a sign that hung above the men’s
urinal read:
Shengming shaobuliao wenming yongyu ⫳ੑᇥϡњ᭛ᯢ⫼䇁

Renqun shaobuliao huansheng xiaoyu Ҏ㕸ᇥϡњ⃶ໄュ䇁
Speech in life should be polite

People should laugh in delight.
Beneath the sign an anonymous graffiti artist had added:
Guanfang shaobuliao wuliao biaoyu ᅬᮍᇥϡњ᮴㘞ᷛ䇁
Official slogans are always trite.147
It is important to note here that the content of the official sign is not the
main object of satire. Apparently some official, somewhere, had had the
idea—not a bad idea— of making use of the few seconds during which a
person urinates to remind people to be happy and polite. Who could argue
with that? It was not the sign’s idea but its source and its style— officialdom,
using its official voice, rendered pompous by the adoption of rhythm and
parallelism—that apparently inspired the sarcasm of the graffiti artist.
When economic recession threatened the world in early 2009 and it
looked as if only China’s labor-intensive growth engine could save inter-
national capitalism, an anonymous wordmeister in China reached back to
147. I am grateful to David Moser for the example. Email message to author, Febru-
ary 17, 2009.
Politics 335
Mao-era officialese to create a shunkouliu that, for its wizardly cleverness,

spread widely:
1949 nian: zhiyou shehuizhuyi cai neng 1949 ᑈ: া᳝⼒ӮЏНᠡ㛑ᬥ

jiu Zhongguo Ё೑
1979 nian: zhiyou zibenzhuyi cai neng 1979 ᑈ: া᳝䌘ᴀЏНᠡ㛑ᬥ
jiu Zhongguo Ё೑
1989 nian: zhiyou Zhongguo cai neng 1989 ᑈ: া᳝Ё೑ᠡ㛑ᬥ⼒Ӯ
jiu shehuizhuyi ЏН
2009 nian: zhiyou Zhongguo cai neng 2009 ᑈ: া᳝Ё೑ᠡ㛑ᬥ䌘ᴀ
jiu zibenzhuyi ЏН
1949: Only socialism can save China. (Mao’s assertion)

1979: Only capitalism can save China. (as Deng turned a corner
from Mao)
1989: Only China can save socialism. (as socialism collapsed across
eastern Eu rope)
2009: Only China can save capitalism. (as Western capitalism fails)
Almost continuously since the early 1950s, official discourse has claimed
that the Party has been campaigning against corruption, combating bu-
reaucratism, rooting out waste, and so on. Such claims are satirized in the
following shunkouliu about SARS, the infectious and sometimes fatal flu-
like disease that spread through China in 2003 precipitating extraordinary
health measures and forcing bureaucrats, on this issue, into unaccustomed
transparency. The piece is called “What the Party Can’t Cure, SARS
Can”:148
dachidahe dang zhibuliao, ໻ৗ໻ୱ‫⊏ܮ‬ϡњ, 䴲‫⊏݌‬њ;

feidian zhi le;
gongkuan lüyou dang zhibuliao, ݀ℒᮙ␌‫⊏ܮ‬ϡњ, 䴲‫⊏݌‬њ;
feidian zhi le;
148. Cited in Hong Zhang, “Making Light of the Dark Side: SARS Jokes and Hu-
mor in China” in Arthur Kleinman and James Watson, eds., SARS in China: Prelude to
Pandemic? (Stanford: Stanford University Press, 2005), p. 152. Zhang gives an excel-
lent translation, but here I have used my own.
wenshanhuihai dang zhibuliao, ᭛ቅӮ⍋‫⊏ܮ‬ϡњ, 䴲‫⊏݌‬њ;

feidian zhi le;
qishangmanxia dang zhibuliao, ℎϞⵦϟ‫⊏ܮ‬ϡњ, 䴲‫⊏݌‬њ;
feidian zhi le;
maiyin piaochang dang zhibuliao, प⎿Ⴊတ‫⊏ܮ‬ϡњ, 䴲‫⊏݌‬њǄ
feidian zhi le;
The Party couldn’t stop rampant banqueting, but SARS did;

The Party couldn’t stop junkets on the pubic till, but SARS did;
The Party couldn’t stop interminable meetings, but SARS did;
The Party couldn’t stop cover-ups, but SARS did;
The Party couldn’t stop traffic in sex, but SARS did.
A piece from early 2008 satirizes the habit in official language of packag-
ing political catchphrases in numbers (the Gang of Four, the Four Mod-
ernizations, the Four Basic Principles, etc.). It is called “The Four Clears
and the Four Unclears”:
Wei shenme kai hui buqingchu, kai hui zuo nar qingchu;
Shei song li bu qingchu, shei mei song li qingchu;
Shei gan de hao buqingchu, gai tiba shei qingchu
He shei shui bu qingchu, shui jiao gan shenme qingchu.
ЎҔМᓔӮϡ⏙Ἦ, ᓔӮതા⏙Ἦ;
䇕䗕⼐ϡ⏙Ἦ, 䇕≵䗕⼐⏙Ἦ;
䇕ᑆᕫདϡ⏙Ἦ, 䆹ᦤᢨ䇕⏙Ἦ;
੠䇕ⴵϡ⏙Ἦ, ⴵ㾝ᑆҔМ⏙ἮǄ
Why hold a meeting?—Unclear

But who sits in what spot?— Clear
Who brought what gifts?—Unclear
But who brought no gift?— Clear
Whose work has been good?—Unclear
But who should be promoted?— Clear
Who sleeps with whom?—Unclear
What happens in bed?— Clear
Politics 337
Other anonymous authors have not been this coy about “what happens in
bed.” The piece “New Year’s Wishes for 2007” lists ten sexually laden
wishes for men and ten more for women in the new year. Each wish plays
on the name of a top leader. For men, it is wished that jiahuo xiang Li chun
yiyang chang ᆊӭ䈵ᴢ᯹ϔḋ䭓 ‘the thing is as long as [propaganda chief]
Li Changchun’, that it jueqi xiang Wu guo yiyang bang ዯ䍋䈵ਈ೑ϔḋẦ
‘rises as magnificently as [National People’s Congress chair] Wu Bang-
guo’, plus eight more. For women, the wishes are that jiahuo xiang Hu tao
yiyang jin ᆊӭ䈵㚵⍯ϔḋ㋻ ‘the thing is as tight as [President] Hu Jintao’,
that pigu xiang Wu guan yiyang zheng ሕ㙵䈵ਈᅬϔḋℷ ‘buttocks are as
proper as [chief of discipline inspection] Wu Guanzheng’, and eight more.
I include this example not to be salacious but, in part, to illustrate the ex-
tremes to which people can go when protected by anonymity.
During times of extreme political pressure, such as the late Mao years,
even authorial anonymity might not be enough to protect one from a mis-
step in the language game. Merely repeating incorrect words, regardless of
who had originated them, could be a political crime. Zhang Xianliang, in
his fictionalized memoirs of labor camp experience in the late 1950s, tells
of meeting a man named Ma Weixiao and talking with him, while the two
are alone in a field, about the causes of the terrible famine they were expe-
riencing. Ma ventures the view that Mao Zedong planned the famine in-
tentionally. The government “has plenty of grain,” Ma says. “Yes, they
have it, but they aren’t bringing it out to feed the people. They want the
people hungry.” Zhang Xianliang’s protagonist asks how that could pos-
sibly be. What would be the point in starving the people? “It’s the best
way there is of reforming people,” Ma answers. Ma follows with a lengthy
explanation of why persuasion and education cannot get done what Mao
wants to get done. Freezing and fire will not work, either. Only hunger
works, because everyone has to eat.149 Hearing these words, Zhang knows
that he cannot repeat them. Normally, in the Mao era, a person could
earn political credit by reporting the counterrevolutionary words of an-
other. But not in this case, because “what Ma had said was enough to get
both listener and speaker executed.” Zhang knew this, and Ma knew it,
149. Zhang Xianliang, Fannao jiushi zhihui ⚺ᙐህᰃᱎ᜻, translated by Martha Av-
ery as Grass Soup (Boston: Godine, 1993), pp. 177–179.
too. He knew that, ironically, “the more counterrevolutionary one’s words

were, the safer one would be.”150
Informal codes have sometimes been useful in avoiding the danger of
saying things explicitly. In the summer of 1988, the widely watched tele-
vision series Heshang (River elegy) fed a national debate over whether
China and its Communist Party were still mired in a mentality that was
left over from the imperial era. It was very much a political debate but was
called—for the sake of indirection— a wenhua re ᭛࣪⛁ ‘cultural fever’. It
included the problem of how the Communist Party operates but did this
under the rubric of discussing “feudalism,” not Leninism. Mao Zedong
and Deng Xiaoping were not mentioned by name, but emperors and “des-
potism” were analyzed. All these were, in a sense, codes. A few months
later, when students demonstrated at Tiananmen, many used the two-
fingered “V for victory” sign, and then, after the June Fourth massacre,
when it suddenly became too dangerous to flash that V sign, the sign per-
sisted in code as people asked yao mai huichong yao ma? 㽕ф㲨㰿㥃৫? ‘do
you want to buy roundworm medicine?’ The understood answer was liang
pian! ϸ⠛! ‘two tablets’, using the fingers to illustrate the number of tab-
lets wanted: a V sign. The gesture became an improvised xiehouyu ℛৢ䇁
‘implied-end phrase’151 that allowed the V-sign to be conjured without ac-
tually being flashed.
Coded terms expanded greatly in the early 2000s, with the rise of the
Internet and the need to evade government filters. Internet police began
using software to intercept words such as minzhu ⇥Џ ‘democracy’, renquan
Ҏᴗ ‘human rights’, and liusi ݁ಯ ‘June Fourth’, and Chinese netizens in
response began inventing coded substitutes. For a time, June Fourth, for
example, became 䰚㙚—the traditional way accountants wrote liu and si
in tamper-proof style. The term lingba xianzhang 䳊ܿᅾゴ, “Charter
08,” was blocked shortly after it appeared on the Internet, so users began
referring to it in several other ways, one of which was the near-homonym
150. Ibid., pp. 180, 181.

151. Xiehouyu resemble riddles in English except that they are terser than riddles,
and what is guessed is a word, not a thing. They always rely on puns or other double
meaning. “A hen on an empty nest,” for example, implies bujian dan ϡ㾕㲟 ‘no egg in
sight’, a pun on bujiandan ϡㅔऩ ‘not simple— complicated’ or ‘not easy—impressive’.
The indirection of xiehouyu allows a speaker to say something that otherwise would be
embarrassing or dangerous.
Politics 339
linba xianzhang ⎟Ꮘও䭓 ‘lymph county-magistrate’. Coded substitutes

such as these are easy for the police to discover and to block. But it is just
as easy for netizens to reinvent new codes and stay one jump ahead.
Some codes in popular usage are brought about less by political repres-
sion from the government than by broader social attitudes. In the early
2000s, for example, people in gay and lesbian subcultures in China began
referring to one another as tongzhi ৠᖫ ‘comrade’.152 Here the oppressing
force was not (or not only) the Party-state, hence it becomes an interest-
ing fact that the code word chosen by the oppressed side was a classic po-
litical term. It did not have to be; the subculture could have chosen a word
like “gay” or “queer”. Apparently, an assumption had taken root that the
terms to which coded language refer are most naturally Communist offi-
cialese, whose vocabulary was already well suited for irony.
Group loyalties based on any number of factors—hometowns, lineages,
classes in school, and so on—have given rise to senses of “we” versus “they”
in Chinese society, both in the Communist era and well before. The Mao
years, though, brought an important new sense of tamen ҪӀ ‘they’ into
the Chinese language. The term became a nickname for the ruling au-
thority. When used without an antecedent, tamen was understood to mean
the Communist Party, the lingdao ‘leadership’, the people who control our
lives. It implied a distance. One was circumspect in dealing with tamen,
who could be arbitrary. (Party jargon may have exacerbated this sense,
without meaning to, by its frequent reference to itself as wo dang ៥‫‘ ܮ‬me
[our] Party’.)
During political campaigns, especially in the Mao years, activists who
were not originally part of the ruling authority, not part of tamen, some-
times chose to speak as if they were. This was usually done in pursuit of
political advantage, but the tactic could backfire as well, because it could
undermine one’s position among one’s peers. Ji Fengyuan, a veteran of
Maoist campaigns, explains the difference between the normal, defensive
use of the official language and this sort of political opportunism:
People were forced to protect themselves and seek advancement by

verbal displays of revolutionary conformity. Words were cheap and
152. Ji Fengyuan writes that the usage spread to mainland China from Hong Kong.
Linguistic Engineering, p. 317.
everyone knew it. So while everyone understood, and respected, a

decent compliance with the norms of linguistic virtue, those who ex-
celled often became not admired models but objects of skepticism.153
Skepticism of this kind could turn into alienation from the redder-than-
red. Liu Binyan, in his classic reportage “People or Monsters?” writes:
Language is a strange thing. When Commissar Yang pointed to Wang

Shouxin as having a “completely red family,” he had meant to praise
her. Yet, in the mouths of the common people, the same phrase—
“completely red family”—was said as a curse.154
During the Mao years, farmers in agricultural brigades and communes

sometimes referred to Party secretaries as tuhuangdi ೳⱛᏱ ‘local emper-
ors’.155 Political operators who were good at manipulating personal rela-
tions got the nickname wanneng jiao ϛ㛑㛊 ‘all-purpose glue’.156 There
are many other such epithets, and they exhibit great variety, but they
have in common a sense of distance, alienation, a divide between us and
them.
The same kind of division has appeared in the popular language of
other Leninist states. In Poland in the 1980s, for example, Anna Wierz-
bicka finds a “polarization between ‘them’ (the people in power) and ‘us’
(the bulk of the nation)” and that “the language of official propaganda
gives rise to its opposite: the unofficial, underground language of
antipropaganda.”157 Vaclav Havel’s brilliant exposé of the language of pre-
tense in “The Power of the Powerless” is based on Czechoslovak ia in the
1970s.158 Miklos Haraszti (The Velvet Prison) for Hungary and Vladimir
153. Ibid., p. 299.

154. Liu Binyan, “Renyao zhi jian” ҎཪП䯈, trans. James Feinerman, in Liu Bin-
yan, Two Kinds of Truth: Stories and Reportage from China, ed. Perry Link (Blooming-
ton: Indiana University Press, 2006), p. 60.
155. Kate Zhou, How the Farmers Changed China: Power of the People (Boulder, Colo.:
Westview Press, 1996), p. 29.
156. Xiangsheng chuangzuo zuotanhui bangongshi, “Xiangsheng chuangzuo zuo-
tanhui jianbao,” day 5, p. 2.
157. Wierzbicka, “Antitotalitarian Language in Poland,” pp. 1, 2.
158. Havel, Living in Truth, p. 45 and elsewhere.
Politics 341
Voinovich (The Fur Hat) for the Soviet Union highlight the difference
between writing honestly and writing for and with “them.”159
A noteworthy feature of all these cases is that the context of repression
gives to ordinary language the potential to be extraordinary. Without re-
pression, a plain statement might be just plain, but under repression, it can
seem uplifting, inspiring, or even profound. Victor Erlich, writing of Soviet
literature in the 1950s, has observed that “when bureaucratic euphemisms
displace the unbearable actuality and explain it away, the simple act of call-
ing a spade a spade, of naming the unspeakable, becomes an epiphany.
When fraudulent official semantics distorts the normal relations between
the sign and referent, responsible and accurate use of language is a blow for
personal dignity.”160 The additional power that writing gets when done in
defiance of repression has enhanced Chinese dissident voices since the 1950s.
The enthusiastic response to Liu Binyan’s work in both the 1950s and the
1980s is largely attributable to Liu’s willingness to write down truths about
corruption and abuse of power in plain, clear language, forming a sharp
contrast with the surrounding official language. In the late 1980s, a group
of students in Beijing experimented with the power of calling a spade a
spade at the level of single words. Is minhang ⇥㟾 ‘the people’s airline’ re-
ally the people’s airline? No, they reasoned, so let’s call it guanhang ᅬ㟾
‘the officials’ airline’. As an experiment, they tried out terms such as guan-
hang on others in society. They found that people at first were startled, but
then, after reflecting for a moment, “got it.” We should observe that there
would have been no joke to “get” if the distance between guan and min had
not already been broadly assumed and accepted. We now turn to some re-
flections on how significant that kind of broad acceptance might be.
Effects of the Language Game in the Mao and the Post-Mao Eras Compared
In 2009, at the sixtieth anniversary of the founding of the People’s Repub-

lic of China, many analysts drew a distinction between the regime’s “fi rst
159. Miklos Haraszti, The Velvet Prison: Artists under State Socialism, trans. Katalin
and Stephen Landesman (New York: Basic Books, 1987); Vladimir Voinovich, The Fur
Hat, trans. Susan Brownsberger (San Diego: Harcourt Brace Jovanovich, 1989).
160. Erlich, “Post-Stalin Trends,” Slavic Review 23, no. 3 (September 1964), p. 418.
thirty” (1949–79) and “second thirty” (1979–2009) years. And indeed,

there has been a considerable difference. From the first to the second of
these two periods, political control of daily life declined considerably, the
country opened much more to the world, and the economy began to oper-
ate in very different ways. Party policies have changed, too, sometimes so
radically as to become the very opposite of what they were before: capital-
ism was excoriated, then embraced; Confucius was denounced as a root of
“feudalism,” then exalted as a symbol of Chinese pride; memory of Japa-
nese war atrocities was at first discouraged, then purposefully stimulated.
On the other hand, some key elements in the Communist Party’s politi-
cal tradition have remained the same: one-party rule, a Leninist govern-
ing structure, domination of political power at the top by the same group
of families, and the mobilization of the tools of governance (including the
police, media, education, and penal systems) to support the preservation
of the political and economic interests of the elite.
Along this spectrum of change and continuity, official language and
the playing of the language game are intermediate examples. Some of the
basic structures are the same, but there has been considerable evolution,
too. To trace every facet of change would be to write an encyclopedia,
which is beyond our scope here. But we can at least take stock of how the
language game in the two periods has seemed to relate, in general, to the
ways people think and behave. Two Western writers, Jean-François Bil-
leter and David Moser, have written brief but astute analyses of the social-
psychological consequences of official language use in the Mao era and
the post-Mao era, respectively. They are worth our consideration.
Billeter grounds his analysis in the Mao regime’s claim that its rule is
based on virtue. This claim has strong antecedents in Confucian culture,
even if the contents of Confucian and Maoist “virtue” have been funda-
mentally different. Any regime that bases its legitimacy on a virtue-claim,
Billeter argues, is “naturally unstable given the fact that, unlike qualifica-
tions of birth, virtue is not objectively measurable or certifiable. In this
kind of regime, everyone had better be virtuous, or at least seem virtuous.”
Next, Billeter argues, the need for everyone to present appearances of vir-
tue (and to play the language game, as I have been considering it here)
naturally makes “all virtue come to appear suspicious.” People dress up
self-interest as virtue, or— even when they do not do this— are suspected
Politics 343
by others of doing it. What Billeter calls a “pathology of virtue” becomes

pervasive:
Suspicion, hypocrisy, and opportunism settle in. Everyone is in danger

of being justifiably or unjustifiably accused of hypocrisy or opportun-
ism. A defense is sought in refusing to take risks or in refusing to expose
oneself to any criticism whatsoever, in other words, in conformism.
This is all the more true since the surest way of displaying revolution-
ary virtue is to denounce the lack of revolutionary virtue in others.161
Another question that is beyond the scope of this book but is well worth
asking is what have the long-term psychological consequences been, for
the generation whose formative years were the Mao era, of growing up
with feigned virtue and suspicions of feigned virtue so natural a part of
life. We have seen, above, many examples of how the language game dur-
ing the Mao era was shaped by the need to negotiate a world in which the
pretenses and suspicions Billeter identifies were dominant. When the ques-
tion “Can I have a bigger apartment?” is expressed as “Do you think we can
concretize Party Central’s policy on intellectuals?” Billeter’s point is illus-
trated. When an answer of no is expressed as “the policy may have difficul-
ties,” it is illustrated again. If the whole matter is dismissed because you are
a “white expert,” we have yet another example.162 Billeter writes that the
“pathology of virtue” can be “catastrophic” for society.
Moser, who worked for many years in Beijing for CCTV, finds a differ-
ent sort of “pathology” in the public’s accommodation of official language.
Looking primarily at the years since 2000, he refers to “schizophrenia,” a
word that he means not in the clinical sense but in the popular sense of
“split perception.”163 He finds that the official language game operates
within certain spheres of life—the politically important spheres—but also
161. Jean-François Billeter, “The System of ‘Class Status,’ ” in Stuart R. Schram, ed.,
The Scope of State Power in China (Hong Kong: Chinese University Press, 1985),
pp. 147–148.
162. See the examples in the Introduction at note 17 and in this chapter in the para-
graphs following notes 90 and 142.
163. David Moser, “Media ‘Schizophrenia’ in China,” www.danwei.org/media _and
_ advertising/media _schizophrenia _in _china _b.php, accessed June 30, 2012.
floats, like a block of styrofoam, in an ocean of other kinds of language use

with which it is fundamentally incompatible.
The situation came about, in Moser’s analysis, largely because the Com-
munist Party could not control the world revolution in digital technology.
During the 1980s, when the Party was combating such things as “bour-
geois liberalization” and “spiritual pollution,” Party leaders could rely on
their ability to seal out unwanted influences. They could never do this abso-
lutely, but they could always make it work pretty well. Newspapers, radio,
and television could be censored, foreign broadcasts could be jammed, the
importation of foreign films could be monitored and rationed, and there
was no Internet to speak of. Then, Moser writes,
this all changed in the mid-90s, when digital technology hit China
like an atomic bomb. Virtually overnight, waves of pirated CDs,
software, computer games, and VCDs became available through
underground bootleg channels. The effect was not merely the open-
ing of a spigot; China was suddenly inundated by a tsunami of for-
eign “memes” and intellectual products. Outdoor stalls in the open-
air markets began to offer counterfeit versions of Windows 95, Jane
Fonda workout videos, music CDs from Mozart to Megadeath, and
movies from Bambi to Basic Instinct. . . . TV junk food like Get Smart
and Charlie’s Angels suddenly appeared on Chinese screens, sand-
wiched between Peking opera and news footage of Li Peng.
What could the Party do? The tsunami could not be turned back. It was
too big and, moreover, liquid in the sense that it seemed flexible at a mo-
lecular level. During the Mao era, and even in the 1980s, people watched
films and plays in large groups, and much newspaper reading was done in
open offices or at billboards in public parks. Such activity could be con-
trolled through censorship and mutual surveillance that induced self-
censorship. But now people could stay at home, make their individual
choices of what to listen to and look at, and, if they liked, keep their activ-
ity pretty much private. So long as there was strong demand and money to
be made, illicit traffic in digital content was nearly impossible to stop.
Moser provides a photo taken in Kunming of a bootleg DVD shop right
next door to a police station. In short, Moser writes, “the propaganda
machine’s worst nightmare had come true.”
Politics 345
On the other hand, the Party could not just walk away from its claim
that official language embodies an immutable political correctness or the
myth that such language represents the unitary voice of “the Chinese
people.” To retreat on these fronts would be to risk the Party’s grip on
power. So the official language continued to march along, lonely in its
own sphere and immaculate in its internal consistency even while it be-
came impossible, indeed unthinkable, that it be integrated into the multi-
plicity that surrounded it. The result, Moser finds, is that Chinese media
content has split clearly between “the news” (i.e, the official news in the
state media, plus a few other topics that need the imprimatur of being
officially correct) and “everything else.”164
The incommensurability of the two realms has caused practical prob-
lems for media workers. Leakage from unofficial language into official
spheres has been hard to avoid and has caused what Moser refers to as
problems of “schizophrenia.” He relates a story about Qin Minxin, a dep-
uty director of the international department of CCTV’s Entertainment
Program Center. Qin was considering whether to air the American sitcom
Friends on CCTV. For years, Friends had been infectiously popular among
young Chinese, who passed it around on bootleg DVDs. Should CCTV
now air it? Qin was not sure:
I had thought the play focused on friendship, but after a careful pre-
view I found each episode had something to do with sex . . . the atti-
tudes of the six close-knit young friends in the play cannot be gener-
ally accepted by Chinese audiences yet.
Could he cut the parts that refer to sex? That would not work, Qin thought,
because
most youth on the Chinese mainland have watched the show and feel
passionate about it. If we make too much trimming, I’m afraid they
will not agree. But it is also impossible that we accept it uncritically . . .
much content of Friends, although considered healthy in the United
States, is unacceptable to the Chinese.
164. Moser notes in passing how an opposite trend emerged in the U.S. media over the
same years; in the U.S., entertainment and news reporting has merged, and the dis-
tinction has become blurred.
Moser then analyzes the several ways Qin has to struggle as he tries to
bring together the pretend world of CCTV language and the actual world
that surrounds it. Chinese audiences “cannot accept” the sexual implica-
tions of Friends, but CCTV must not censor this content because it attracts
audiences. Contradictory? Yes—but also no, because “cannot accept,” in
the official language, is pretend language for “should not, in our opinion,
accept.” In the pretend world, “the people” do not do anything they should
not do. It therefore also becomes meaningful for CCTV officials to pon-
der whether to “expose” people to Friends despite the fact that they are al-
ready well exposed to it. (Indeed, the fact that viewers find the show attrac-
tive is the very reason why CCTV is considering its “introduction.”) By
the end of Moser’s analysis, one almost sympathizes with Qin. He needed
to juggle two worlds, a real one and a pretend one.
The ersatz flavor of the pretend world can be sensed even when explicit
comparison to the real world is absent. It is enough for the real world to exist
only vaguely in the background. One of Moser’s examples of the self-
revelation of pretense is the annual CCTV Chinese New Year variety show:
The lavish costumes, the unrelenting upbeat tone, all the glitz and
flashy production values seem designed to distract from the empty
core of the affair, an over-compensation for the impossibility of of-
fering anything that reflects real life.
Moser calls the presentation of real life an “impossibility,” not just an

omission, because of the closed nature, and ultimately the fragility, of the
official language system. If a performer were to step out of line and crack
a joke about current affairs, or satirize a leader, or present a “national mi-
nority” person who was not wearing colorful native garb and a toothy
smile, the aberration would not stand on its own. It would threaten the en-
tire world of pretense. The failure of protocol at point A would expose its
artifice at all other points. Moser writes: “Since performers do not— and
cannot in principle—relate to the audience in an honest fashion . . . no
wonder the shots of the audience so often reveal a sea of stony-faced spec-
tators applauding robotically on cue.”
Does the pretend world of official language in the post-Mao era pose a
danger to the whole of society in the way that language based in a “pa-
thology of virtue,” in Billeter’s analysis, did during the Mao era? Probably
Politics 347
not, at least not to the same extent. Official language in post-Mao times
can no longer dominate as it did under Mao. Significant dangers do remain,
however. I see primarily two, one potential and one actual.
The potential danger is an intensification in the use of the official lan-
guage to stimulate and exploit nationalism. After the June Fourth mas-
sacre of 1989, when the image of China’s rulers was at a low point, Jiang
Zemin (backed by Deng Xiaoping) made the strategic decision to use na-
tionalism to try to recoup the Party’s image. A number of measures—
including the Beijing Olympics, the Shanghai World’s Fair, the denuncia-
tion of “splittism” in Tibet, Xinjiang, and Taiwan, the injection of
chauvinist sentiment into news reports of confl icts with Japan, the United
States, and other countries—have met with considerable success in this
regard. A stress on nationalism serves the interest of the ruling authority
in two important ways: first, it distracts attention from problems that citi-
zens otherwise complain about— corruption, special privilege, pollution,
rights violations, a growing wealth gap, and so on— and second, it helps to
change the image of Party leaders. Instead of being the targets of popular
resentment because of all the problems, these leaders can present them-
selves as standard-bearers and heroes of the Chinese nation. The danger,
as of the second decade of the twenty-first century, is that nationalism
could be magnified even much more. It could draw deeper on Chinese
national pride and on the sense of aggrievement about the history of the
previous two centuries. The nation’s textbooks already stress this ag-
grievement, and the potential to magnify it further, and to marshal ener-
gies behind the idea that China should be “number one,” is very consider-
able. I am not ready to predict that an outsized chauvinism will appear in
China. But one is possible, and it would be bad news for both China and
the rest of the world if it should come.
In my view, the greatest actual social-psychological cost of the official
language game has been something rather different: it is the general ac-
ceptance by the Chinese citizenry that the demands of the official lan-
guage game are “normal” and should be accepted as a part of daily life.
The game’s demands prohibit certain topics from public discussion: the
disasters of the Mao era, the 1989 June Fourth massacre, misbehavior
among top leaders or their family members, the Falun Gong movement,
political questions relating to Taiwan, Tibet, and Xinjiang, the prospect
of an end to one-party rule, and a number of others. All these are impor-
tant topics on which a healthy society would conduct open debate. But
many Chinese citizens have grown so accustomed to avoiding these topics
that their absence from the public sphere seems unremarkable. People go
through daily life—making money, enjoying fashions, playing sports,
traveling, finding romantic partners, and doing other things that people
could not easily do during the Mao era—while simply avoiding the areas
in the world of ideas that could cause “trouble.” This is, in a sense, a ratio-
nal way to behave. Why should a person spoil a good thing? Dissidents
report that their colleagues, neighbors, and even families sometimes find
it odd, and even a bit stupid, that they venture into politically forbidden
zones and do things that, although idealistically aimed to help the whole
of society, in fact are likely to hurt themselves. Most of the public wants to
be “smarter” than that. It is smarter to stay out of trouble and take what
you can get. My main worry about this pattern is not just that it isolates
the “dissidents”—the people who are courageous enough to speak about
ideals. My worry is that, for the public at large, a myopic outlook on the
world comes to seem ordinary and normal.
In his essay on “schizophrenia,” David Moser relates an incident in
which he was invited to be a guest on a CCTV talk show on the topic of
the Internet. Before the taping the host sought to put him at ease. “Just
relax and say anything you want,” the host counseled, “and it should be
okay.” Moser reports feeling surprise at the man’s apparent nonchalance.
“Anything that comes to mind?” Moser asked. “Does that include how the
government blocks sensitive sites and news sources? Can we freely discuss
Internet pornography? Or how chat rooms are monitored and censored?”
The host backed off. “Yes, well, almost anything,” he said, as if reminded
of another way to look at the world. The value of this anecdote is not just
to give yet another example of the well-known fact that China’s Internet
is censored. The more significant point is that a CCTV host can become
so accustomed to self-censorship, so relaxed with it, that even a word like
“anything” can be severely stunted in his usage without his seeming to
have noticed. He accepts the rules of a language game so completely that,
unless reminded, he does not realize he is doing so.
Epilogue
To view the matter superficially, I chose the three themes in this book—
rhythm, metaphor, and politics— simply because they are facets of the
modern Chinese language that I have found interesting over the years. I
had taken a lot of notes on them.
But this explanation harbors a deeper question: why did these particular
aspects of the language, and not others, draw my attention? Why was I not
taking notes on the bai hua vernacular movement, on debates over roman-
ization, or on conventions of punctuation or paragraphing? The topics one
might choose to study are almost endless. Why these three? Do they share
anything in common?
They do, I believe. Two commonalities in particular stand out.
One is that all three are features of language whose use normally goes
unnoticed. A few people—professional linguists and others—are con-
sciously aware of them, but people in daily life seldom are. Most speakers
of Chinese just “absorb” them— and use them correctly in both speech
and writing—but are unaware of doing so. This inadvertency contrasts
sharply with the very conscious manner in which other aspects of language
are learned. Chinese children are highly aware of what they are doing
when they labor to master character writing or punctuation rules; their
parents and teachers, too, pay plenty of attention when checking their
work for errors. Advanced students of Chinese literature, in their con-
scious work, study such topics as genre, form, narrative point of view, in-
tertextual influences, and the ideology and historical contexts of literary

works. Only a few scholars, working in linguistics or cognitive science,
turn conscious attention to how prosody or conceptual metaphor work in
daily-life language.
For example, the overwhelming majority of speakers of Chinese are
unaware—and do not need to be aware—of how a 2–2–3 (qiyan ϗ㿔)
rhythm can serve to formalize or exalt a piece of ordinary language; or how
the structural metaphor guo 䖛 ‘cross over’ works in expressions about en-
tering or leaving consciousness ( yunguoqu ᰩ䖛এ ‘faint and go-over’,
xingguolai 䝦䖛ᴹ ‘awaken and come-over’, etc.); or how, in political lan-
guage, a phrase like jishaoshu ᵕᇥ᭄ ‘tiny minority’ adds a nonlexical
derogation that causes the phrase to mean something like “tiny (trouble-
making) minority.” Still— and this is the interesting part—speakers of Chi-
nese completely master these conventions and rely on them in daily life. In order
for daily-life communication to work, rules that people do not know about
are just as important as those of which they are consciously aware.
The other commonality among the three topics of rhythm, metaphor,
and politics is that despite the inadvertency of their use, each of these ele-
ments affects meaning, that is, makes a difference in what is communicated.
This is not something that can be said of all linguistic inadvertencies. For
example, English speakers who insert “filler” phrases like “you know” and
“I mean” in their sentences and Chinese speakers who use similar fillers
such as zheige, zheige, zheige 䖭Ͼ, 䖭Ͼ, 䖭Ͼ . . . (this, this, this . . . ) or
neige, neige, neige 䙷Ͼ, 䙷Ͼ, 䙷Ͼ . . . (that, that, that . . . ) are normally
adding nothing to the meanings of the overall sentences they are uttering.
At most, such words are but floor-holders, whose “meaning” might be
said to be “I would like to keep on talking even though I have not yet fig-
ured out what it is that I want to say next.” (The English word “like” in a
sentence such as “He asked me to dance and I’m—like—what?” also ap-
pears empty but probably is not entirely so. It does have a sliver of mean-
ing, as attested by the fact that, at least sometimes, it cannot easily be
omitted; but that sliver of meaning is extremely hard to articulate.)
By contrast, the “meanings” of rhythm, metaphor, and political con-
notation that I have examined in this book are usually more substantial
than this, and often can be at least partly articulated. The exaltation qiyan
provides, the spatial conception of consciousness yunguoqu implies, and
the denigration jishaoshu connotes are all fairly clear. Moreover, we can
Epilogue 351
count them as “meaning” (and not just ad hoc flavors) because they remain
reliably the same among the very large community of speakers of Chinese
who pass them back and forth. In using the qiyan pattern to make a phrase,
one can assume that one’s listeners will correctly apprehend the exaltation;
the negative connotation of jishaoshu, when used in political contexts, is
equally standard. This sort of “meaning,” to be sure, is vaguer than that of
a word like zhuantou ⷪ༈ ‘brick’. But even its quality of vagueness is some-
thing that travels effectively from speaker to speaker.
The combination of these two commonalities—inadvertency and
meaningfulness—is especially interesting because of the paradox it seems
to imply. We normally think of meaning as something that we mean, and
what we normally understand by “mean” is that we are aware of what we
are doing. If I step on your toe accidentally, I can apologize that I did not
mean to. In this book, though, we have studied aspects of meaning that we
do not notice. Meaning that we do not notice? The phrase almost seems self-
contradictory. But there can be no denying the phenomenon. It is there.
This book offers a range of examples of it, and my examples are only a
smattering of what is there to be found.
How should we describe this kind of “unnoticed meaning”? Metaphors
of plumbing, or of anatomy, come to mind. We are dealing with the under-
girding of language, with functions that are vital even if they are not obvi-
ous on the surface. Our awareness of these quiescent workings of words is
usually about as good as our awareness of our pancreas. Our life depends
on our pancreas, though we seldom think about it and (except for a few of
us) do not understand it or even try to. An analogy to bicycle-riding seems
useful as well. Most people who ride bicycles do not understand why it is
possible to ride a bicycle.
For such things—pancreases, bicycles, or the undergirdings of a lan-
guage—it is appropriate to ask the question, “Why study them?” Under-
standing for its own sake is, of course, always a defensible answer to this
question. But are there more practical benefits? In the case of the pan-
creas, avoidance of severe pain or death from pancreatitis is an obvious
practical payoff. What about the undergirdings of language use? Are
there any practical payoffs?
I believe that there are, and that some of the examples in this book can
show this to be so. For one, there is value in becoming consciously aware
of the ways “meanings” can be delivered inadvertently—bypassing, as it
were, the critical judgment that one normally would want to apply. For
example, when a message expressed in qiyan rhythm strikes us as being
authoritative, it can be useful to realize that the sense of authority comes
in part from the rhythm alone, not necessarily from any special stature of
the issuer of the words or from any special wisdom of the words them-
selves. This does not mean that we should turn rebellious when we see a
street sign in qiyan that says yi kan, er man, san tongguo ϔⳟ, Ѡ᜶, ϝ䗮䘢
‘first look, then go slowly, then cross’ or when a political leader commands
linghun shenchu gan geming! ♉儖⏅໘ᑆ䴽ੑ ‘make revolution in the
depths of the soul!’ We can still choose whether to respect authority or
not. The value of becoming aware of the rhythm is that we can also be
aware of the claim to authority the rhythm places on us. We can remind
ourselves that rhythms per se have no grounds to make such claims, and
that we can set them aside if we like. This leaves us more free to think for
ourselves.
Parallelism, alliteration, song, and other embellishments of words can
have similar effects. Chiasmus, as we have seen, is an especially clear ex-
ample of this kind of imposition on our intellect. Chiasmus seems to
claim special access to wisdom, and it is important that we be able to set
this claim aside if we like. When a politician says that Enron executives
were “either criminally stupid or stupidly criminal,” the art of his phrase,
and the power it generates, almost seem to say to us, “This is so obviously
right that you needn’t think about it anymore.” The politician may, of
course, have an excellent point about Enron. But we are better off if we
probe further and find exactly what that point is than if we allow ourselves
to be transfi xed by chiasmus and cut off our thinking.
Many of the examples I considered in Chapter 3 illustrate benefits that
can be had simply by understanding what political language is and how it
differs from daily-life language. The person who is officially labeled part
of a “tiny minority” ( jishaoshu) is obviously better off if he or she can un-
derstand that the phrase is not a description of the actual size of the com-
munity that shares his or her opinion but a pejorative term that a Party-
state uses in order to serve a purpose. In the late 1950s, learning to play
the official language game began to be a required part of Chinese daily
life, and in ensuing decades people grew skilled at using the official lan-
guage not only to defend themselves but to pursue interests of their own.
This kind of conscious use of official language brought the political un-
Epilogue 353
derpinnings of language to the surface more clearly than the others—

those relating to rhythm and metaphor—that this book has addressed. As
of the early twenty-first century, Chinese speakers remain much less
aware of the rhythmic and metaphorical undergirdings of their language
than they are of the political ones. This imbalance is understandable, of
course. It has sometimes been vitally important to understand what a
phrase like jishaoshu is doing, whereas understanding the hidden connota-
tions of a conceptual metaphor like yunguoqu, or of a qiyan rhythm, has
been much less important.
In cases where political language combines with either a rhythm or a
conceptual metaphor, it can be enlightening to observe how the under-
girdings sometimes clash. Metaphors and rhythms—because people have
examined them less than political language— are sometimes employed
even when they contradict conscious political messages. In metaphor use,
for example, the conscious ideology of the Communist movement says
workers are the “leading class” and officials “serve the people.” But con-
ceptual metaphors—in both official and unofficial language—have con-
sistently used shang Ϟ ‘up’ for people who have more power and xia ϟ
‘below’ for people who have less. When workers make a report to their
leaders, the report inevitably “goes up” (shangqu Ϟএ), even though in the
political ideology, the workers are already at the top and the officials the-
oretically should be serving them from “below.” In the case of rhythm, we
saw in Chapter 1 several examples of how Maoist culture used wuyan and
qiyan rhythms to exalt politically correct things, including Mao himself,
even while the same Maoist culture, at the conscious level, was denounc-
ing things like ancient rhythms under the rubric of “the four olds” (cus-
toms, culture, habits, ideas). There is value in noticing these inconsisten-
cies, and the value is not merely in noticing that the human mind can
handle conceptual inconsistencies and still get along in daily life. The
additional value, indeed the greater one, is in perceiving the double
standards they reveal. With rare exceptions, it has never really been true
that leaders in the Communist system have conceived of themselves as
servants of the working class, and understanding the inadvertencies of
their daily-life metaphors can help to make this point explicit. Similarly,
Maoist culture never really succeeded in rooting out “the four olds”—
even from itself, as its frequent use of wuyan and qiyan rhythms makes
clear.
In a broader sense, study of the inadvertencies of language considered

in this book can help us to understand how the human mind works. For
the cognitive scientists whose work I drew on in Chapter 2, this line of
inquiry has been the main goal. With rare exceptions, though, the sub-
field of cognitive science that works on conceptual metaphor has not done
too well at testing its hypotheses against non-European languages, and
there is something to be gained by trying to do so. By comparing Chinese
and English uses of fundamental conceptual metaphors such as “space for
time”— as Ning Yu does, for example— one can shed at least some light on
the elusive question of to what extent conceptual metaphor varies with
culture and to what extent it springs from structures that are likely uni-
versal in the human brain. To what extent the timeless “mind-body”
problem might be related to habits of metaphor use is but one example of
the possible benefits of studying conceptual metaphor comparatively.
At a more mundane level, awareness of how conceptual metaphors can
differ from one culture to another can help us avoid cultural misunder-
standing. In Chapter 2, I related an anecdote about how two eminent schol-
ars, one Chinese and one American, became very unhappy with each other
over a misunderstanding about what “yellow” meant. Their exchange would
have been entirely different—very friendly—if they had understood the
same connotations for the term; but they did not, and the result was pow-
erful anger. Such a pure example of metaphorical misunderstanding is
rare, but it illustrates a more general danger that affects many other and
more complicated cases. The metaphorical connotations of “dog,” for ex-
ample, are generally more negative in Chinese than in English, so that a
Chinese phrase like zougou 䍄⢫ ‘running dog’ loses some of its sting when
translated literally into English; on the other hand, the English word
“watchdog” can take on an unintended negative connotation if translated
too literally into Chinese. In any case, everyone is better off to be aware
of such things than not to be. It can be useful, as Dilin Liu has shown, to
reflect on the prominence of metaphor families in particular cultures.
American English makes much use of sports metaphors and car-driving
metaphors; Chinese has special predilections for metaphors of eating and
of stage performance, as well as—in the Mao and the post-Mao eras—
military metaphors. It is hard to be precise about how significant these
tendencies are, but it would likely be a mistake to say that they have no
significance at all.
Epilogue 355
More generally— although this, too, would be hard to show concretely—

I suspect that any person who notices the undergirdings of his or her
daily-life language, whether comparatively or not, is likely to reap benefits
just from thinking, and therefore speaking and writing, more clearly and
consistently. There are so many examples of this that one can cite them
almost at random. For conceptual metaphors, it is useful to understand
why the phrase houdai de qiantu ৢҷⱘࠡ䗨 ‘the future of future genera-
tions’ is not contradictory despite its surface appearance of being so. To
be clear about the two space-for-time metaphors that it combines is to be
able to use qian and hou metaphorically, in any number of contexts, with
more precision and confidence. Finally, to understand linguistic undergird-
ings can simply be enjoyable, even if no practical advantages are at stake.
By “enjoyable” I mean here something a bit more than the phrase “under-
standing for its own sake” that I used above. Understanding for its own
sake can be fun, but need not be. I fi nd it particularly fun to notice that a
3–3–7 rhythm appears in a Chinese nursery rhyme (ni pai yi, wo pai yi, yige
xiaohair kai feiji ‘you pat one, I pat one; one little child fl ies the plane’) and
also appears in an English rhyme (“this old man, he plays one, he plays
knick-knack on my thumb”), and to wonder why, and to discover a com-
plex history that begins at least fifteen hundred years ago and originates
(separately?) at two ends of the Silk Road as well as in Eu rope. Similarly, I
think, it is fun to note that the Maoist slogan gongye xue Daqing ᎹϮᄺ໻
ᑚ ‘industry should learn from [the] Daqing [oilfields]’ employs an ancient
Chinese rhythm. What would Stalin say? (What, indeed, would Mao
have said, had he noticed?)
There is also something fun—but even more than fun, something
deeply satisfying—to see how many up-and-down metaphors, space-for-
time metaphors, color metaphors, and other fundaments of the concep-
tual apparatuses of Chinese and English are the same, even though speak-
ers of the two languages often like to think that they and their cultures
are “very different” from the other. The commonality of the human ex-
perience seems underscored by such discoveries.
Acknowledgments
My interests in the topics of this book have grown on their own, without a clear
plan and in unforeseen ways, beginning primarily in the early 1990s and thanks
to conversations with a wide variety of people. I hesitate to begin listing their
names because I know that any list will be incomplete. I cringe to think of readers
who might scan the following list for their own names, not fi nd them, and then
judge me, rightly, to be either forgetful or insufficiently ungrateful. Still, to say
nothing at all would be even worse, so I will try.
I owe everything to Rulan Pian, who started me in Chinese at Harvard in
1963– 64, and to her genius-father, Yuen Ren Chao, whose books and articles have
not only taught me much but have set standards for me and others to emulate. My
mentor and dear friend Ta-tuan Ch’en at Princeton taught me an im mense
amount over decades of working together. Chapter 1 on rhythm owes much to
Feng Shengli. Sam Glucksberg provided crucial help in the field of conceptual
metaphor for Chapter 2. David Moser has given many examples and astute advice
on all three chapters. Others who have provided materials, offered insights, or
reviewed drafts include Nicholas Admussen, An Kun, Robert Bagley, Anthony
Barbieri-Low, Chen Ping, Joanne Chiang, Chih-ping Chou, Duanmu San, John
Frankenstein, Harry Frankfurt, Meow Hui Goh, Adele Goldberg, Hu Ping,
Monica Link, Liu Binyan, Victor Mair, Mao Ruxing, Mao Sheng, Daniel Osher-
son, Andrew Plaks, James Pusey, James Richardson, Patricia Russel, Michael
Schoenhals, Su Wei, Su Xiaokang, Tong Yi, Wang Haicheng, Wang Lixiong,
Wang Wei, Xiao Qiang, Ye Minlei, Yu Maochun, Yu Ying-shih, and Bell Yung.
None of these, of course, bears responsibility for my mistakes.
I am grateful for the support of the Chiang Ching-kuo Foundation for research
support during 1994– 95, when research for this book first began in a serious way;
to Chinese University Press for its permission to use, in Chapter 1, portions of
358 Acknowledgments
my essay “The Secret History of Classical Rhythms in Modern Chinese,” from

The Scholar’s Mind: Essays in Honor of F. W. Mote, ed. Perry Link (Hong Kong:
Chinese University Press, 2009); and to Tong Yi, Daniel Link, and Samuel Link
for their graceful indulgence of my use of time.
Index
Abstract nouns, 222–223, 244–245, 248, Authoritarian regimes: metaphors used by,
255–256 256–257; numbers used by, 266–267; and
Acting metaphor. See Stage per for mance official language, 258, 277
metaphors Authority: and abstract language, 256; and
Adjectives: in conceptual metaphors, 220; ambiguity, 288; and fourth tone, 94– 95;
nouns as, 261; in official language, 262; and qiyan patterns, 352
repetition of, 192–193
Adverbs in official language, 248 Baigujing xianxing ji (Ma Ji), 32
Advertisements: military metaphors in, Bartholomew, Terese, 100
253; parallelism used in, 105; rhythmic Basic Color Terms (Berlin & Kay), 148
patterns in, 34–35 Beijing Massacre (1989): metonyms for,
“Affection is warmth” metaphor, 185 210; and military metaphors, 19; and
Al-Kjshgarq, Mahmud, 169 political language game, 16, 246,
Alverson, Hoyt, 115, 226 270–271, 275, 294, 297–298, 320, 327;
Ambiguity: in official language, and stage per for mance metaphors,
247–248, 255–256, 287–288, 251
318–320; and self-censorship, 317; Being, 226–227
and stage per for mance metaphors, Berlin, Brent, 148–149
208 Billeter, Jean-François, 342–343
Ames, Roger, 203, 205, 227 Black, 149, 154
Analects (Confucius), 276 Bloch, Maurice, 90, 275–276, 278, 290
Anatomy metaphors, 351 Bo Yibo, 258
“Anger is heat” metaphors, 186–187 Bodde, Derk, 223
Animals, 269, 314 Book of Odes, 90
Anonymity of Internet, 334 Boroditsky, Lera, 130, 131, 132, 133, 146
Anti-Rightist Movement (1957), 33, 235, Borrowing of metaphors, 184–187, 232
250, 254, 259, 296–297, 322 Brightness metaphors, 190
Anti-Spiritual Pollution Campaign (1983), Brushes with Power (Kraus), 308
318
Apartheid, 291 Calligraphy, 204
Aspect markers, 262–264 Camel Xiangzi (Lao She), 24
Austin, J. L., 83 Can Xue, 294
360 index
Cao Changqing, 246, 248, 249, 254, 256, Consonants, 96–101

267, 295–296 Container metaphors, 213–214, 215,
Cao Xueqin, 219 216–218, 221–222
Cao Yu, 84 The Contemporary Theory of Metaphor:
The Captive Mind (Milocz), 13 A Perspective from Chinese (Ning Yu),
“Causation is emergence” metaphor, 214 138
Censoring mechanisms, 317 “Control is up” metaphor, 159–160
Central Chinese Television (CCTV), Cook, Sarah, 291
34–35, 345–346 Cooper, William, 177
Ceremonial trappings, 276 Cooperative principle, 119
Chan, Anita, 208 Covert sexism, 180–183
Chang, Jung, 292 Credibility, 243
Chao, Y. R., 2, 22–23, 41, 72, 83, 102, 207 Cultural memory, 84
Charges-of-convenience, 310 Cultural Revolution. See Great Proletarian
Charter 08, 327, 333, 338 Cultural Revolution
Chaucer, Geoffrey, 75 Cultural T-shirts, 27
Chen Dengke, 101
Chen Duxiu, 43 Daily-life language: and language game,
Chen Guangcheng, 310 278–341; metaphors in, 115–128; qiyan
Chen, Matthew, 62 patterns in, 25–26; rhythm patterns in,
Chen, Jenn-yeu, 131, 137, 144, 209 24–37; wuyan patterns in, 25–26
Chen Yi, 309 Dalai Lama, 252, 257, 281–282, 285, 307
Chen Yun, 303 Darkness metaphors, 190
Cheng Dachang, 167 Dead metaphors, 121, 251, 300
“The Chess Master” (Zhong Acheng), Deception, 91
293–294 Defi ning concept, 119
Chiang Kai-shek, 305 Deng Xiaoping: Mao’s criticisms of, 328;
Chiasmus, 106–109 metaphors used by, 8; and official
Children’s names, 100–101, 268 language, 242, 270, 274, 288–289, 303;
Chomsky, Noam, 11, 115, 129, 133, 134 repetition used by, 266
Chou, Jane, 221 Descartes, René, 11
Chuang, H. C., 243 Determinacy, 161
Clapper-tales, 30 Dickens, Charles, 219
Clash of Civilizations and the Remaking of Difficulty metaphors, 189–190
World Order (Huntington), 207 Ding Ling, 293, 303
“Closeness is strength” metaphor, Dingell, John, 108
193–195 Direction. See North and south
Coded language, 338 metaphors
Cognitive science: on color metaphors, Dittmer, Lowell, 243
150; and metaphors, 116, 117–118, 354; Doing Things with Words in Chinese Politics
on time metaphors, 145 (Schoenhals), 274
Color metaphors, 147–155, 355 Dominant external rhythms, 54–59
Committee on Scholarly Communication Dominant metaphors, 198–209
with China (CSCC), 2, 286, 322 Double meaning, 100, 328–329, 330
Community memory, 84 Down. See Up and down metaphors
Completeness, 93 Dream of the Red Chamber (Cao Xueqin),
Conceptualizing capacity, 224 205, 207, 219
Conceptual metaphors, 116. See also Driving of vehicles metaphors, 199, 209,
Metaphors 354
Confucius, 172, 173, 200, 276, 320, Drumsinging, 30
342 Du Fu, 12, 81
Consciousness metaphors, 9, 122–123, Duanmu San, 37, 51, 52, 53–54, 68, 73, 88
169–171 Dummy verbs, 17–18, 271
Conservative, 250 Dyads, privilege in, 174–183
index 361
“The East Is Red” (song), 12 Focal colors, 12, 149–150

Eating metaphors, 198–200, 209, 354 Fodor, Jerry, 129
Effortless Action: Wu- wei as Conceptual Formality: of official language, 234,
Metaphor and Spiritual Ideal in Early 242–243; and stage per for mance
China (Slingerland), 172 metaphors, 203; up and down metaphors
“Eight Honors and Eight Shames,” to convey, 161
34, 266 Foul odor metaphors, 188–189
Ekman, Paul, 186, 187, 188 “Four Basic Principles,” 247, 266, 270
English language: color metaphors in, “The Four Clears and the Four Unclears”
147–155; consciousness metaphors in, (anonymous), 336
169–171; container metaphors in, “Four Modernizations,” 266
216–218; dominant metaphors in, Four-syllable patterns: in daily-life
198–209; dyad privilege in, 174–183; language, 27; in Great Leap Forward
left-to-right orientation of, 144; length period, 44–45; in recessive rhythms, 61;
and volume as spatial metaphors in, speaker’s awareness of, 111; in xiang-
132; lexicon of, 244–245; meanings of sheng, 31
rhythmic patterns in, 86; pitch differ- Fourth tone, 94– 96, 101–102
ences in, 101; recessive rhythms in, 56; Frankenstein, John, 265
stress patterns of syllables in, 21–22; Freud, Sigmund, 123, 169
time metaphors in, 115, 133, 136–147; up Friedlander, Henry, 290
and down metaphors in, 145, 155–162, Friends (television sitcom), 345–346
255 Front and back metaphors, 195–197
Erlich, Victor, 322, 341 Fruit language, 246, 248
Euphemisms, 311–314 The Fur Hat (Voinovich), 341
Existence, 226–227 The Future Beckons (Zhao Zixiong), 242
External rhythms, 54–59
Gang of Four, 302, 311
False verb-object compounds, 72 Gao, 17–18, 271–273
Falun Gong, 306, 307 Gao Xingjian, 192
Family relations. See “Government is Gao Yubao, 71
family” metaphor Gay subculture, 339
Fan fenlie doctrine, 285 Ge Fei, 294
Fang Lizhi, 259, 280–281, 328, 330 Gendered terms, 180–183
Fear: and ambiguity in official language, Gender-neutral dyads, 179–180
318; and ontological metaphors, 216; Glucksberg, Sam, 123
and self-censorship, 317; and violent Goal orientation of official language,
language, 254 270–274
Feng Menglong, 63 Goebbels, Joseph, 296
Feng Shengli, 51, 53, 58–59, 60– 61, 67 “Government is family” metaphor,
Fiction: official language in, 302–303; 200–201, 209
wuyan and qiyan patterns in, 25–26 Graham, A. C., 205
Filler verbs, 17–18, 271 Grammar: and meaning, 54–55; moral
First tone, 101 weight of, 267; of official language, 234,
“Fit”: as kind of truth, 274–278; and 235, 237, 260–267; recessive rhythms
meanings of rhythms, 90; and parallel- altering, 73–74; rhythmic patterns
ism, 103 determined by, 37; and three-syllable
“Five black categories,” 66, 266 phrases, 51
“Five pay-attentions and four beautifuls,” Gravity, 130, 146
112 Great Leap Forward: euphemisms used in,
“Five Relations,” 200 311; four-syllable phrasing used in,
Five-syllable strings, 62– 63, 64– 66. See also 44–45; and official language, 235, 236,
Wuyan patterns 322; parallelism used in, 103; pattern
Flexibility, up and down metaphors to 3–3–7 used in, 80; rhythmic phrasing
convey, 161 used in, 45, 93
362 index
Great Proletarian Cultural Revolution: Infantilization, 265

aspect marker use in, 263–264; chiasmus Informality: codes for, 338; and official
used in, 106; difficulty metaphors used language, 323–324; of ordinary
in, 189; grammar use in, 261; and official language, 242–243; up and down
language, 249, 269–270; pattern 3–3–7 metaphors to convey, 161. See also
used in, 80; rhythmic phrasing used in, Daily-life language
16, 33, 45, 66; and stage perfor mance “Instrument is companion” metaphor,
metaphors, 204 210–211
Greek language: length and volume as Internal pressure metaphors, 187–188
spatial metaphors in, 132; time meta- Internet, 310–311, 334, 338–339
phors in, 133 Irony, 120–121
Green, 152 Item lists, 66– 67
“Green Dam Youth Escort” software,
310–311 Jenner, William, 279
Grice, Paul, 119 Ji Fengyuan: on military metaphors, 251;
Gu Hua, 325 on mindset of political language game,
Guo Moruo, 26 292; on official language, 243, 279, 297,
Guo Qiru, 31 302; on psychological effect of political
language, 300, 339–340
Ha Jin, 294 Jia Baoyu, 179, 205
Happiness, 157–158, 188 Jiang Kun, 329
Haraszti, Miklos, 13, 244, 340 Jiang Qing, 241, 305
Havel, Vaclav, 244, 332–333, 340 Jiang Zemin, 238, 266, 282–283, 284, 289
He Chi, 33 Jichu Hanyu (textbook), 261
He Qinglian, 310 Jituan, 305–306
Heat. See “Anger is heat” metaphors Johnson, Mark: on “affection is warmth”
Heng, Liang, 238 metaphor, 185; on “causation is
Heshang (television series), 152, 338 emergence” metaphor, 214; on
Hidden meanings, 100 “closeness is strength metaphor,” 193,
Higher-order conditioning, 300–301 195; on conceptual metaphors, 116–117,
Historical materialism, 239 119–122; on consciousness metaphors,
Hitler, Adolf, 256, 257 169; on cultural values in metaphors,
Homonyms, 100 198; on difficulty metaphors, 189; on
Horizontal movement metaphors, 125, 130, front and back metaphors, 196; on
138, 143, 144, 147, 170 “important is big” metaphor, 191; on
Hou Baolin, 31, 100 intonation, 102; on metaphors and
Hsia, T. A., 243, 251 thought, 8–11, 128–130, 133–134; on
Hsu, Mei-ling, 164 “mountain as person” metaphor, 197; on
Hu Fayun, 292 ontological metaphors, 15, 215–216, 245;
Hu Jintao, 34, 176–177, 201, 289, 313 on privilege in dyads, 177; on repetition
Hu Ping, 273–274 in metaphors, 192–193; on time
Hu Qiaomu, 242, 260 metaphors, 145; on up and down
Hu Shi, 43, 106 metaphors, 157–158, 159
Hu Yaobang, 240, 241–242, 254–255, 272 Jung, Carl, 169
Hu, ji, dan (Wang Guoxiang), 218 Jusczyk, Peter W., 77–78
Hua Guofeng, 275, 277, 284
Huntington, Samuel, 207 Kang Sheng, 311
Kant, Immanuel, 11, 115, 134
Implied conditionals, 70 Kay, Paul, 148–149, 150–151
“Important is big” metaphor, 191 Kennedy, John F., 106
Indonesian language: length and volume Keysar, Boaz, 124
as spatial metaphors in, 132; time Köhler, Wolfgang, 97
metaphors in, 133 “Kong Yiji” (Lu Xun), 4, 56
index 363
Kövecses, Zoltán, 158 Liu Binyan, 181, 235–236, 246–247, 278,

Kraus, Richard, 308 317, 340, 341
Liu, Dilin, 183, 198, 200, 201, 206, 209, 354
Lakoff, George: on “affection is warmth” Liu Shaoqi, 305, 320
metaphor, 185, 186; on “causation is Liu Xiaobo, 333
emergence” metaphor, 214; on Liu Xiaogan, 291
“closeness is strength” metaphor, 193, Logos, left-to-right orientation of, 143–144
195; on color metaphors, 150–151; on Lu Bingfu, 37
conceptualizing capacity, 224; on Lü Shuxiang, 243
conceptual metaphors, 116–117, Lu Xinhua, 254
119–122; on consciousness metaphors, Lu Xun, 4, 15, 56
169; on cultural values in metaphors, Luo Gan, 246
198; on difficulty metaphors, 189; on
front and back metaphors, 196; on Ma Ji, 32
“important is big” metaphor, 191; on Ma Weixiao, 337
intonation, 102; on metaphors and Ma Yuan, 294
thought, 8–11, 128–130, 133–134; on “The Magical Number Seven, Plus or
“mountain as person” metaphor, 197; on Minus Two” (Miller), 77
nominalization, 225; on ontological Male-female dyad, 179–180
metaphors, 15, 215–216, 245; physiologi- Mao: The Unknown Story (Jung Chang), 292
cal universality hypothesis of, 188; on Mao Zedong: awareness of rhythmic
privilege in dyads, 177; on repetition in patterns, 41–42, 57; calligraphy of, 308;
metaphors, 192–193; on time metaphors, and chiasmus, 106; criticisms of Deng
145; on up and down metaphors, Xiaoping, 328; difficulty metaphors used
157–158, 159 by, 189; and “five black categories,” 66;
Language game in politics, 278–341; Mao and gao, 18; and “government as family”
vs. post-Mao era effects of, 341–348; metaphor, 201; grammar use by, 260; and
ruled side of game, 321–341; rulers’ side language game in politics, 341–348; and
of game, 295–321 literary style, 302–303; metaphors used
Lao She, 24, 200, 272 by, 256–257; and official language, 14,
Laozi, 172 15, 235, 241–242, 243, 292, 304, 307; and
Lauer, R. H., 222–223 parallelism, 104–105; poetry of, 309;
Left-to-right orientation, 143–144, repetition used by, 265, 266; rhythmic
250–251 phrasing of, 4–5, 33, 47–48, 92, 353–354;
Length as spatial metaphor, 132 and self-censorship, 317; and stage
Lesbian subculture, 339 per for mance metaphors, 204, 208; up
Lessing, Doris, 256 and down metaphors used by, 155
Levi, Primo, 314 Maps, 163–166
Lexicon of official language, 244–260 Marketing metaphors, 198–199, 209
Li Bai, 12, 81 May Fourth movement (1919), 210
Li Denghui, 266 May 16 Declaration (1989), 294
Li Peng, 16, 88, 246, 251 McEwan, Ian, 56
Li Shangyin, 168 McGinn, Colin, 228–229, 230–231
Li Tuo, 13, 19, 293, 303 McGlone, Matthew, 123
Liang Shanbo, 179 Meanings: and chiasmus, 106–109; and
Liberal, 250 consonants, 96–101; contributory
Limericks, 55–56, 74, 87 features, 94–109; double meaning, 100;
Lin Biao, 69, 305, 311, 320 and grammar, 54–55; of metaphors, 350;
Lin Daiyu, 179 negative, 54; and parallelism, 103–106;
Linguistic Engineering ( Ji Fengyuan), 300 and pitch, 101–103; of rhythm, 5– 6,
“Linguistic Violence: The Power of 82– 94, 350; speakers’ awareness of,
Intimidation in Authoritarian Rule” 109–112; and three-syllable phrases, 51;
(Cao Changqing), 295–296 of tones, 94– 96; and vowels, 96–101
364 index
Meaning-transfers, 100 Mountain as person metaphor, 197

Medical metaphors, 254 Mozi, 205
“Me fi rst” theory, 177–178 Music, 55, 75, 89
Memory, rhythm as aid for, 84, 92– 93 “My Humble Opinion on the Reform of
Mencius, 172 Literature” (Hu Shi), 43
Metaphor, Culture, and Worldview (Dilin The Mysterious Flame: Conscious Minds in a
Liu), 183, 198 Material World (McGinn), 228
Metaphors, 113–233; “affection is warmth,” The Mysterious Stranger (Twain), 200
185; of color, 147–155; conceptual
differences rooted in, 209–215; of Nanjing Massacre (1937), 314–315
consciousness, 169–171; in daily-life Nathan, Andrew, 279, 282–283
language, 115–128; dead metaphors, 121, National Football League logos, 143–144
251, 300; divergences between Chinese Nationalism, 347
and English languages, 198–209; Nazis, 256, 257, 290–291, 312, 314
dominant, 198–209; meanings of, 350; Negative meanings, 54, 193, 267
mixed, 124, 136–137; north and south, “The Neural Lyre” (Turner), 77
162–169; in official language, 244–260; Neutral tone, 49, 60
philosophical problems generated by “New Year’s Wishes for 2007” (anony-
conceptual metaphors, 215–231; and mous), 337
privilege in dyads, 174–183; psychology Niehaihua (novel), 108
of, 125; and self in ancient thought, 1984 (Orwell), 298
171–174; shared by Chinese and English Ning Yu, 125, 138, 139, 141–142, 143, 145,
languages, 183–198; significance of 158, 185, 218, 354
conceptual metaphors in different Nominalization, 223–224
languages, 231–233; speakers’ aware- Nonhead stress rule, 52, 53, 68– 69
ness of, 349–350, 353; and thought, North and south metaphors, 162–169
128–136; of time, 136–147; up and Noun repetition, 192–193
down, 155–162 Nouns: abstract, 222–223, 244–245, 248; as
Metaphors We Live By (Lakoff & Johnson), adjectives, 261; and container metaphors,
8– 9, 102, 116, 128, 135 213–214; cultural preference for, 219,
Metaphor theory, 117–118 222
Metonyms, 209–210 Numbers: and metaphors, 230–231; official
Military maps, 165 language use of, 266–267, 336; patterns
Military metaphors, 19, 251–252, 253, 354 to memorize, 84, 92
Miller, George, 77
Milosevic, Slobodan, 257 Odor metaphors, 188–189
Milosz, Czeslaw, 13, 244 Official language, 243–278; “fit” as kind of
Mind-body problem, 150–151, 228–229, truth in, 274–278; goal orientation of, 17,
233, 354 270–274; grammar in, 260–267; lexicon
Misty poetry, 293 of, 244–260; metaphor in, 244–260;
Mixed metaphors, 124, 136–137 moral weight of, 267–270; rhythm in,
Mo Shaoping, 333 260–267. See also Politics
Mo Yan, 294 Oliver Twist (Dickens), 219
Modisane, Bloke, 291 “On Establishing a New Literature”
Molière, 212 (Hu Shi), 106
Moral weight: and language use, 206, “On Literary Revolution” (Chen Duxiu),
304–308; of official language, 267–270; 43
and stage per for mance metaphors, Onomatopoeia, 96, 98, 193
203; of up and down metaphors, “On the ‘Natural Foot’ in Chinese” (Feng
155–156 Shengli), 58
Morphemes, 37–39 Ontological metaphors, 9–10, 15, 215–216,
Morrison, Toni, 107 218, 245
Moser, David, 158, 178–179, 181, 182, 342, Opera/acting metaphor. See Stage
343–347, 348 per for mance metaphors
index 365
Orwell, George, 15, 234, 244, 245–246, “The Power of the Powerless” (Havel), 340
248–249, 256, 298, 313 Predicative complement, 207
Ouyang Xiu, 62 Privilege in dyads, 174–183
Propaganda: and official language, 298,
Paired categories, 174–183 301, 316; rhythms in, 32
Parallelism: in daily-life language, 3, 4; and Prosody, 21. See also Rhythms
meanings, 103–106; in official language, Psychology of metaphor, 125
34–35, 352; in popu lar sayings, 29; and Pu Songling, 303
rhythm, 103–106; speaker’s awareness of, Puns, 100, 239, 330
43, 110
Pathology of virtue, 343 Qiao Shi, 88
Peck, Graham, 113–114, 137, 141 Qin Minxin, 345–346
Pekinese Rhymes (Vitale), 80 Qiyan patterns: in advertising, 253; in
Peking University, 42 daily-life language, 25–26; defi ned, 2–3;
Peng Ruigao, 34 and item lists, 67; and meaning, 89;
“People or Monsters?” (Liu Binyan), 340 meaning in, 352; moral weight of, 268;
People’s Daily: parallelism used by, 105; pattern 3–3–7 combined with, 78–79; in
rhythmic patterns used in, 45–47, 48–49; recessive patterns, 73; in recessive
on tifa, 274 rhythms, 56; as rhythmic preference,
Philosophical problems generated by 61– 62, 75
conceptual metaphors, 10–11, 172, Quyi, 30–31
215–231, 233
The Philosophy of Lao Zhang (Lao She), 200 Rebiya Kadeer, 281–282
Phuntsog Wanggyal, 285 Recessive external rhythms, 54–59, 61;
Physiological universality hypothesis, of favor, 60– 67; structural effects of,
188 68– 74
Pianwen, 63 The Records of History (Sima Qian), 298
Pinker, Steven, 128–129, 133, 134 Red, 11, 149, 153, 269
Pitch and meanings, 101–103 Redundant syllables, 69
Plato, 227, 228 Repetition: in metaphors, 191–192; of
Playing dumb, 327–328 nouns, 192–193; in official language,
Plumbing metaphors, 351 264–265, 296; of verbs, 192–193
Plus-minus pairs, 179. See also Dyads, Resnikov, L. O., 300
privilege in Responsibility-shifting, 320
Poetry: dominant vs. recessive rhythms in, Rhythms, 21–112; and chiasmus, 106–109;
57; meanings of rhythm in, 91; misty in Chinese vs. other languages, 37–40;
poetry, 293; pattern 3–3–7 used in, 81; and consonants, 96–101; dominant
political symbolism of, 309; qiyan external, 54–59; external, 54–59; fads in,
patterns in, 25; rhythmic patterns in, 44–49; meanings of, 5– 6, 82– 94, 350; as
41–42, 55, 77, 85; wuyan patterns in, 25 memory aid, 84, 92– 93; moral weight of,
Pol Pot, 257, 291 267, 268; in official language, 16, 234,
Politics, 234–348; bifurcation of official 260–267; and parallelism, 103–106; and
and unofficial language use, 13–14, pitch, 101–103; prevalence of patterns in
235–243; characteristics of official daily-life Chinese, 24–37; recessive,
language, 243–278; and language game, 54–74; roots of, 49–54; speakers’
278–341; Mao vs. post-Mao era effects of awareness of, 40–44, 109–112, 349–350,
language game, 341–348; of nationalism, 353; and tones, 94– 96; universality of
347 preferred rhythms, 74– 82; and vowels,
“Politics and the English Language” 96–101
(Orwell), 15, 244 Ritualized language, 290
Positionality, 222 Role-playing, 208. See also Stage per for-
Power: of official language, 241; tifa as mance metaphors
form of, 275; up and down metaphors to Romantic relationship metaphors, 122
convey, 156–157 Roosevelt, Franklin, 92
366 index
Rosemont, Henry, 203, 205, 227 Stress patterns, 21–23. See also Rhythms
Ross, John, 177 Strong-arm lie technique, 298
Ruoxi, Chen, 243 Structural metaphors, 116. See also
Metaphors
Sadness, 157–158 Struggle metaphors, 251–252, 291
Salzman, Mark, 241, 280 Su Xiaokang, 152, 154, 282, 284
Sapir-Whorf hypothesis, 128, 202 Subconscious, 123, 169. See also Conscious-
SARS, 335 ness metaphors
Scar literature, 239–240, 250, 252–253, 293, Sunrise (Cao Yu), 84
302–303 Suppression of Counterrevolutionaries
Schell, Orville, 283 campaign, 45
Schoenhals, Michael, 13, 258, 269, Syllabic balance, 49, 71, 87
274–275, 296–297, 298 Syllabic elongation, 103
Scientific method, 223 Syllabic stress patterns, 21–22
Scott, Amanda, 137, 146, 222–223 “Symbols, Song, Dance, and Features of
Searle, John, 117, 185–186, 220 Articulation” (Bloch), 90
Seat-of-subjectivity self, 172
Second tone, 101 Tai, James, 221, 261, 262–263
Self, 126, 171–174 Tamaki, Ogawa, 82
Self-censorship, 317, 318, 344–347, 348 Tartuffe (Molière), 212
Self-cultivation, 126 Thaw literature, 239
Self-perception theory, 300 Thought and metaphors, 128–136
Semantic inversion, 3, 106 “Three loves,” 112
Semantics and Experience (Alverson), 115 “Three Represents,” 238–239, 266, 289
Seven-syllable strings, 62– 63, 64– 66. Three-syllable strings, 50
See also Qiyan patterns 3–3–7 pattern, 28–29, 78– 82
Shakespeare, William, 75, 86, 119 The Tiananmen Papers (Zhang Liang),
Shandong “fast tales,” 75 282–283
Shanghai Municipal Propaganda Bureau, Tibetan protests (2008), 252, 265, 281–282,
34 285, 306
“Shanghen” (Lu Xinhua), 254 Tibet Daily on protests, 252, 262, 306
Shapiro, Judith, 238 Tifa, 274–278
Shiro, Azumo, 314 Time lines, 131, 136–137, 143, 147
Shunkouliu, 30, 84, 335 Time metaphors, 9, 114–115, 136–147,
Sima Qian, 298 209–210, 232, 355
Six-syllable strings, 63, 64 “Tiny minority” phrasing, 16, 306–307, 352
Slingerland, Edward, 90, 126–127, 172–174, Tones: and irony, 121; meanings of, 94– 96;
177, 227 of official language, 234
Smith, Craig, 34 Topographic maps, 165
So it [email protected] (Hu Fayun), 292 Transitive verbs, 161, 262
South. See North and south metaphors Truth: fit as kind of, 274–278; and official
Spanish language: spatial metaphors in, language, 281; and stage per for mance
132; time metaphors in, 133 metaphors, 205, 206–207
Spatial metaphors, 131–133, 134 Truth and Reconciliation Commission
Speakers’ awareness: of four-syllable (TRC, South Africa), 156–157
phrases, 111; of meanings, 109–112; of Turner, Frederick, 77, 89– 90, 91
metaphors, 353; of parallelism, 43, 110; Tutu, Desmond, 156–157
of rhythm, 40–44, 109–112, 349–350, 353 Twain, Mark, 200
Sports metaphors, 199, 209, 354 Two Kinds of Time (Peck), 113
Stage per for mance metaphors, 199, Two-syllable adjectives, 54
201–206, 209, 251, 354 Two-syllable strings, 60
Stative verbs. See Adjectives
“Stinky is bad” metaphor, 188–189 Uighur protests (2009), 281–282, 306
Storytelling, 30 Unconsciousness metaphors, 122–123
index 367
“Understanding is seeing” metaphor, Wierzbicka, Anna, 291, 340

211–212 Wiesenthal, Simon, 313
Unofficial language. See Daily-life Wilde, Oscar, 91
language Wild Swans: Three Daughters of China ( Jung
Up and down metaphors, 155–162; Chang), 292
conceptual clashes for, 125; cultural Wittgenstein, Ludwig, 83, 228
preference for, 355; in official language, Worldview and metaphor, 183–198, 202,
255; for time, 130, 137, 144, 147 213
Updike, John, 56 Wu Zuxiang, 236–237, 328
Wuwei, 172–174
Vacuous syllables, 69 Wuyan patterns: in daily-life language,
Validity effect, 300 25–26; and item lists, 67; and meaning,
The Velvet Prison (Haraszti), 340 89; moral weight of, 268; in recessive
Verb-object constructions, 51 patterns, 73; as rhythmic preference,
Verbs: cultural preference for, 219; dummy 61– 62
verb, 17–18, 271; false verb-object
compounds, 72; in official language, 248; Xiangsheng, 30–31, 32, 253, 329
repetition of, 192–193; transitive, 161, Xing Lu, 243
262 Xu Zhenya, 4
Vertical movement metaphors. See Up and Xu Zhimo, 74
down metaphors Xunzi, 172, 173, 276, 277
Violent language, 253–254
Virtue, 155–156, 342–343 Yan Dongsheng, 154
Visual metaphors, 211–212 Yashin, Alexander, 325
Vitale, Guido, 80 Ye Jianying, 309
Vocabulary of official language, 234, 237 Yellow, 11, 154
Voinovich, Vladimir, 340–341 Yellow journalism, 154
Volume as spatial metaphor, 132 Yu Gong shanchuan dili tu (Cheng
Vowels, 96–101 Dachang), 167
Yulihun (Xu Zhenya), 4
Wakeman, Frederic, 154
Wang Anyi, 4 Zhang Heng, 168
Wang, David, 108 Zhang Liang, 282
Wang Guoxiang, 218 Zhang Xianliang, 337
Wang Huo, 26 Zhao Zixiong, 242
Wang Luxiang, 152 Zhao Ziyang, 246
Wang Shuo, 70 Zhen Shiyin, 205
Wang Wei, 218 Zheng Yi, 315
War metaphors, 19, 251–252, 253, 354 Zhong Acheng, 293–294
Warmth. See “Affection is warmth” Zhonghua Book Company, 168
metaphor Zhongshan University, 14, 207, 239, 287,
Watson, Burton, 170–171 308–309
Wei, James, 100 Zhou Enlai, 25, 73, 309, 320
Wei Jingsheng, 309–310 Zhou Liming, 296
What If I Really Were? (play), 315–317 Zhou Yang, 308–309
“What the Party Can’t Cure, SARS Can” Zhu Dexi, 243
(anonymous), 335–336 Zhu Yingtai, 179
White, 149, 153 Zhuangzi, 9, 170–171, 172
Whorf, Benjamin, 128, 130, 133, 135 Zhuo Lin, 303

An Anatomy of Chinese

Uploaded by

Copyright:

Available Formats

An Anatomy of Chinese

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

An Anatomy of Chinese

Uploaded by

Copyright:

Available Formats

What is the book about?

What is the book about?

What are some of the main topics discussed in the book?

What are some of the main topics discussed in the book?

An Anatomy of Chinese

Library of Congress Cataloging-in-Publication Data

Link, E. Perry (Eugene Perry), 1944–

Metaphor and Thought 128

deans announce programs and deliver homilies to encourage “interdisci-

[Usually] I shall prefer to use a familiar term, with a warning against

1. Yuen Ren Chao, A Grammar of Spoken Chinese (Berkeley: University of California

folksongs, proverbs, and storytelling in many later eras. (In Chinese it is

Gaogao xingxing chu cheng zou 催催݈݈ߎජ䍄

Handai wenxue yu sixiang xueshu yantaohui lunwenji ⓶ҷ᭛ᅌ㟛ᗱᛇᅌ㸧ⷨ㿢᳗䂪᭛䲚

7. I am indebted to Adele E. Goldberg for this insight. See her “Constructions: A

9. George Lakoff and Mark Johnson, Metaphors We Live By (Chicago: University of

10. Ibid., p. 15.

11. Ibid., p. 26.

to shape thought in contemporary Chinese society. The Swedish scholar

17. Neng bu neng luoshi yixia zhongyang de zhishifenzi zhengce? 㛑ϡ㛑㨑ᅲϔϟЁ༂ⱘ

lexically—would have sounded wrong in the official language. There

and structural metaphors: how do these aspects of language that we use

1. Example is from Mark Liberman, “The Intonation System of English” (Ph.D.

for “rhythm.” In this book, my focus is on Mandarin, but only because I

ultimately mystical about stress patterns. But for my purposes in this

The Prevalence of Rhythmic Patterns in Daily-Life Chinese

Rhythms in oratory and literature can be lengthy, complex, and varie-

something. In writing of the Nanjing massacre of 1939, Wang Huo subti-

Bushi Guo lao chui de xiong ϡᰃ䛁㗕਍ᕫ‫ܛ‬

If our great Mr. Guo had not puffed this place up

12. Www.signese.com, July 9, 2006. Viewed June 21, 2012.

that in the 1940s, basketball cheerleaders at Nankai University in Tianjin

Yi, er, san; san, er, yi ϔѠϝ, ϝѠϔ,

One, two, three; three, two, one

Ma wuzi, hong zhangzi 咏ሟᄤ, ㋙ᐇᄤ

A room of hemp, a curtain of red,

It appears as well in nursery rhymes:

Ni pai yi, wo pai yi Դᢡϔ, ៥ᢡϔ

You pat one, I pat one

Ni pai er, wo pai er ԴᢡѠ, ៥ᢡѠ

You pat two, I pat two

The historical roots of these rhythmic patterns lie in popular sayings,

Ri mo yanzhi hong ᮹≦㛁㛖㋙

When the sky is red at the end of the day

Popular sayings often include the “poetic” conventions of parallelism as

Jian ren shuo ren hua 㽟Ҏ䁾Ҏ䁅

Speaking human language to a person

Qingchun xian gei dang 䴦᯹⤂㌺咼

I worked my whole life for the Party,

Xinxinkuku sishi nian 䕯䕯㢺㢺ಯकᑈ

For forty-some years, ever more perspiration

16. The common mistranslation of xiangsheng as “crosstalk” appears to be based on

Zuo ge xiangshengr yanyuan, ‫Ⳍןخ‬㙆‫ܦ‬ⓨવ

Performing xiangsheng is no easy matter;

luoli luosuo yida dui ಝ䞠ಝ૚ϔ໻ේ

a big pile of loquaciousness

But returns to 4:4 in saying that “refi ned” Beijing dialect is

duan xiao jing han (er) ⷁᇣ㊒ᙡ (㗠)

terse, forceful, (and) powerful in logic

Elsewhere, 4:4 is prefaced by a heavy beat at the beginnings of two paral-