Sie sind auf Seite 1von 376


Series Editor
Michael Krausz, Bryn Mawr College
Advisory Board
Annette Baier (University of Pittsburgh), Cora Diamond (University of
Virginia), William Dray (University of Ottawa), Nancy Fraser (North-
western University), Clifford Geertz (Institute for Advanced Study,
Princeton), Peter Hacker (St. Johns College, Oxford), Rom Harr (Linacre
College, Oxford), Bernard Harrison (University of Utah), Martha
Nussbaum (University of Chicago), Leon Pompa (University of Birmingham),
Joseph Raz (Balliol College, Oxford), Amlie Oksenberg Rorty (Brandeis
University), Georg Henrik Von Wright (University of Helsinki)

Constructive Engagement
This book is printed on acid-free paper.
Library of Congress Cataloging-in-Publication Data
Davidsons philosophy and Chinese philosophy : constructive engagement / edited by
Bo Mou.
p. cm. (Philosophy of history and culture ISSN 0922-6001 ; 23)
Includes bibliographical references and index.
ISBN 90-04-15048-X (alk. paper)
1. Philosophy, Chinese. 2. Davidson, Donald, 1917- I. Mou, Bo, 1956- II. Series.
B126.D34 2006
ISSN 09226001
ISBN-13: 978-90-04-15048-5
ISBN-10: 90-04-15048-X
Copyright 2006 by Koninklijke Brill NV, Leiden, The Netherlands
Koninklijke Brill NV incorporates the imprints Brill Academic Publishers,
Martinus Nijhoff Publishers and VSP.
All rights reserved. No part of this publication may be reproduced, translated, stored in
a retrieval system, or transmitted in any form or by any means, electronic,
mechanical, photocopying, recording or otherwise, without prior written
permission from the publisher.
Authorization to photocopy items for internal or personal
use is granted by Brill provided that
the appropriate fees are paid directly to The Copyright
Clearance Center, 222 Rosewood Drive, Suite 910
Danvers MA 01923, USA.
Fees are subject to change.
In memory of Donald Davidson (19172003),
who inspired and participated in the project
Acknowledgments ........................................................................ xi
Note on Transcription ................................................................ xv
Contributors ................................................................................ xvii
How Constructive Engagement of Davidsons
Philosophy and Chinese Philosophy is Possible:
A Theme Introduction .......................................................... 1
Bo Mou
Chapter One Relativism and Its Schemes ............................ 37
Michael Krausz
Chapter Two Davidson and Chinese Conceptual
Scheme .................................................................................... 55
Koji Tanaka
Chapter Three Making Room for Comparative
Philosophy: Davidson, Brandom, and Conceptual
Distance .................................................................................. 73
Stephen C. Angle
Chapter Four Where Charity Begins .................................... 103
David B. Wong
Chapter Five Davidsons Charity in the Context of
Chinese Philosophy ................................................................ 117
Yiu-ming Fung
viii cox+rx+s
Chapter Six Davidsonian Rationality and Ethical
Disagreement between Cultures ............................................ 165
Samuel C. Wheeler
Chapter Seven A Davidsonian Approach to Normativity
and the Limits of Cross-Cultural Interpretation .................. 189
Yujian Zheng
Chapter Eight On Two Kinds of Meaning and
Interpretation .......................................................................... 207
A.P. Martinich
Chapter Nine Metaphorical Use versus Metaphorical
Essence: Examples from Chinese Philosophy ...................... 229
Kim-chong Chong
Chapter Ten Reading the Analects with Davidson: Mood,
Force, and Communicative Practice in Early China .......... 247
Yang Xiao
Chapter Eleven From Donald Davidsons Use of
Convention T to Meaning and Truth in Chinese
Language ................................................................................ 271
Chung-ying Cheng
cox+rx+s ix
Chapter Twelve Truth Pursuit and Dao Pursuit: From
Davidsons Approach to Classical Daoist Approach in
View of the Thesis of Truth as Strategic Normative
Goal ........................................................................................ 309
Bo Mou
Index ............................................................................................ 351
My deep appreciation goes to late Professor Donald Davidson whose
thought and style of doing philosophy, and whose valuable partici-
pation at the earlier stages of this anthology project, as explained in
my theme introduction below, have signicantly inspired and con-
tributed to the project.
I am very grateful to all the other contributing authors of this vol-
ume for their valuable contributions, all of which are previously
unpublished pieces written expressly for this book, and for their
patience, cooperation, and understanding throughout the process,
during which I have learnt a lot from them in various aspects. Their
persistent support of this project and of this editors eorts have
become especially valuable when Donald Davidson passed away at
one important stage of this project, as Davidsons participation in
this project in several ways is one of the main momentums for our
reective eorts in this constructive-engagement project. I am espe-
cially indebted to Michael Krausz not merely for his role as an active
contributing participant but also for his timely and eective help in
determining a decent academic publisher like Brill.
To eectively fulll the constructive-engagement purpose, this
anthology project is accompanied with its conference project to pro-
vide a critical discussion and engagement platform. In this way,
although the anthology project per se is an independent project instead
of the conference proceedings, the latter is rather one indispensable
stage for the sake of fullling the goal of this anthology and for the
sake of eectively implementing the constructive-engagement strat-
egy. During the whole process of preparing for the anthology pro-
ject including its closely related conference project as one crucial
stage of critical engagement platform, we have received a large
amount of support, help and assistance in various ways from vari-
ous parties. During the process of reviewing the submissions, I am
grateful to Wan-Chuan Fang, Yiu-ming Fung and Linhe Han for
their valuable review work and their precious time. I am grateful to
Xianglong Zhang, my colleague in the 20022005 board of the
International Society for Comparative Studies of Chinese and Western
Philosophy (ISCWP), for his persistent support since the conference
xii .ckxovrrrovrx+s
project became the rst one in the ISCWP constructive engage-
ment international conference series. I am thankful to the Institute
of Foreign Philosophy, Peking University, China, for its assuming
the conference host for the originally-scheduled August-2003 con-
ference, which had to be postponed due to the SARS outreach in
spring 2003; I am especially grateful to Linhe Han for his active
role as the conference-host representative in coordinating various
preparations. I am grateful to the Institute of Philosophy, Chinese
Academy of Social Sciences for its assuming the conference host for
the re-scheduled June-2004 conference; I am thankful to Pengcheng
Li, Deputy Director of the Institute of Philosophy, CASS, and Jing
Sun, Director of its Research Coordination Oce, for their sub-
stantial support; I am especially grateful to He Li for his active role
as the conference-host representative in coordinating various prepa-
rations. I am grateful to those speakers other than the contributors
to this volume, Bo Cheng, Wan-Chuan Fang, Yi Jiang and Chuang
Ye for their valuable and engaging talks at the conference. My sin-
cere thanks also go to Lian Cheng, He Li, Jian Li, and Xiwen Luo
for their helpful and eective professional service as the conference
session chairs, and to Jigang Shan, Jihong Lei and Xiaojian Zhang
for their eective logistics supports for the conference.
I am indebted to the American Philosophical Associations
Committee on International Cooperation (CIC), under the leader-
ship of its chair Alan M. Olson, for its valuable support and co-
sponsorship for the above mentioned international conference project
on Davidsons philosophy and Chinese philosophy during 20022004
when I served as a member of the CIC.
I am grateful to Roger Ames, Editor of the journal Philosophy East
and West, and He Li, Editor of the Chinese journal World Philosophy,
for their valuable support and help in setting precious space in their
journals for publishing the call for papers and/or news of some
ISCWP academic activities including this project. I am also thank-
ful to Christ Caputo at the American Philosophical Association for
providing the space at the APA website to post the call for papers
for the project.
I am grateful to my school, San Jose State University, and its
Department of Philosophy for their various substantial supports that
are related to this anthology project. A California State University
Research Grant for 20032004 has signicantly contributed to my
work on this anthology. I am thankful to Phillip Willamson, who
.ckxovrrrovrx+s xiii
was my graduate-student assistant in spring 2005, for his professional
assistance that he completed timely. My sabbatical leave in fall 2005
has enabled me to eciently nish the nal phase of the whole
I am grateful to our editors at Brill, Marcella Mulder at the early
stage and Boris van Gool and Birgitta Poelmans at the later stage,
for their variety of kindly and timely professional assistance.
Bo Mou
Albany, California
December 15, 2005
Because of its ocial status in China, its relative accuracy in tran-
scribing actual pronunciation in Chinese common speech and con-
sequent world-wide use, we employ the pinyin romanization system
in this volume for transliterating Chinese names or terms. However,
those Chinese names or terms are left in their original romaniza-
tions (typically in the Wade-Giles system) in the following cases: (i)
the titles of cited publications; (ii) the names whose romanizations
have become conventional (such as Confucius); and (iii) the names
of the writers who have had their authored English publications
under their regular non-pinyin romanized names (such as Fung Yu-
lan). The title of a cited contemporary Chinese book and essay is
given in its pinyin transcription with its translation or paraphrase
given in parentheses. The following rule of thumb has been used in
dealing with the order of the surname (i.e., family name) and given
name in romanized Chinese names: (i) for the name of a historical
gure in Chinese history, the surname appears rst, and the given
name second (such as Zhu Xi); and (ii) for contemporary gures,
we follow their own practice in this aspect when they publish in
English or other Western languages (typically, the given name appears
rst, and the surname second). In the pinyin versions of Chinese pub-
lication titles and those proper phrases that contain two or more
than two Chinese characters, hyphens may be used to indicate sepa-
rate characters.
Transcription Conversion Table
Wade-Giles Pinyin
ai ei
ch zh
ch ch
hs x
ien ian
xvi xo+r ox +n.xscnir+iox
-ih -i
j r
k g
k k
p b
p p
szu si
t d
t t
ts, tz z
ts, tz c
tzu zi
ung ong
yu you
Table (cont.)
Wade-Giles Pinyin
Axorr, S+rrnrx C. is Director of the Manseld Freeman Center for
East Asian Studies and Associate Professor of Philosophy at Wesleyan
University, USA. He received his B.A. from Yale in East Asian
Studies and his Ph.D. in Philosophy from the University of Michigan.
He is the author of Human Rights and Chinese Thought: A Cross-Cultural
Inquiry (Cambridge U.P., 2002) and the co-editor and co-translator
of The Chinese Human Rights Reader (M.E. Sharpe, 2001). Angle stud-
ies Chinese ethical and political thought from the Song dynasty though
the present, and is also interested in issues in the methodology of
comparative philosophy.
Cnrxo, Cntxo-vixo is Professor of Philosophy at University of Hawaii
at Manoa, USA. He received his Ph.D. degree from Harvard University
(1964). Cheng is the founder and honorary President of the
International Society for Chinese Philosophy and International Society
for the Yijing; he is the Editor-in-Chief of Journal of Chinese Philosophy.
He is the author of many articles and books on Chinese philosophy
and comparative philosophy, including Peirces and Lewiss Theories of
Induction (1969), Modernization and Universalization of Chinese Culture (1988,
in Chinese), New Dimensions of Confucian and New-Confucian Philosophy
(1991). He is co-editor of Contemporary Chinese Philosophy (2002).
Cnoxo, Kiv-cnoxo is Professor of Humanities at the Hong Kong
University of Science and Technology. He was formerly with the
National University of Singapore, where he served as Head of
Department for several years. His interests are in ethics, Chinese
philosophy, and comparative philosophy. His publications include
Moral Agoraphobia: The Challenge of Egoism (Peter Lang, 1996); The Moral
Circle and the Self: Chinese and Western Approaches (co-edited, Open Court,
2003); and Early Confucian Ethics (Open Court: forthcoming).
Ftxo, Yit-vixo is Chair Professor of the Division of Humanities at
the Hong Kong University of Science and Technology. He received
his Ph.D. degree in philosophy from the Chinese University of Hong
Kong in 1984. Fung is the author of several books, including The
Methodological Problems of Chinese Philosophy (1989); Chinese Philosophy in
the Ancient Period, 4 volumes (1992); Kung-Sun Lung Tzu: A Perspective
xviii cox+nint+ons
of Analytic Philosophy (1999); and The Myth of Transcendent Immanence: A
Perspective of Analytic Philosophy on Contemporary Neo-Confucianism (2003).
He has also published more than 80 research papers both in Chinese
and in English.
Kn.xsz, Micn.rr is the Milton C. Nahm Professor of the Department
of Philosophy at Bryn Mawr College, USA. Krausz is the author of
Limits of Rightness (2000), Rightness and Reasons: Interpretation in Cultural
Practices (1993), and Varieties of Relativism (with Rom Harr) (1995).
He is contributing editor of Relativism: Interpretation and Confrontation
(1989), The Interpretation of Music: Philosophical Essays (1993) and Is There
a Single Right Interpretation? (2002). Krausz is contributing co-editor of
The Concept of Creativity in Science and Art (1981), Relativism: Cognitive and
Moral (1984), Rationality, Relativism and the Human Sciences (1986) and
Interpretation, Relativism and the Metaphysics of Culture (1999). In 2003 a
festschrift on his work was published by Rodopi: Interpretation and Its
Objects: Studies in the Philosophy of Michael Krausz.
M.n+ixicn, Arovsits P. is Roy Allison Vaughan Centennial Professor
of Philosophy and Professor of History and Government at the
University of Texas at Austin, USA. He received his Ph.D. from the
University of California at San Diego. He is the author or editor of
many books and articles. His books include Philosophical Writing 3rd
edition (Blackwell, 2005), The Philosophy of Language 4th edition (Oxford
University Press, 2001), Hobbes: A Biography (Cambridge University
Press, 1999), The Two Gods of Leviathan (Cambridge: Cambridge
University Press, 1992) and Communication and Reference (Berlin and
New York: Walter de Gruyter, 1984).
Mot, Bo is Associate Professor of Philosophy at San Jose State
University, USA. After receiving B.S. in mathematics, he received
M.A. from Graduate School, Chinese Academy of Social Sciences,
and Ph.D. from University of Rochester, USA. Representative pub-
lications include A Metaphilosophical Analysis of the Core Idea of
Deationism, Metaphilosophy (2000), The Enumerative Character of
Tarskis Denition of Truth and Its General Character in a Tarskian
System, Synthese (2001), Two Roads to Wisdom?Chinese and Analytic
Philosophical Traditions (contributing editor, 2001), and A Re-examination
of the Structure and Content of Confuciuss Version of the Golden
Rule, Philosophy East and West (2004).
T.x.k., Koi is a Postdoctoral Fellow in the Philosophy Department
at Macquarie University, Australia. From 2006, he will be a Lecturer
cox+nint+ons xix
in the Philosophy Department at the University of Auckland, New
Zealand. He served as Reviews Editor for Studia Logica, an inter-
national journal for symbolic logic. He has published widely in lead-
ing journals and made contributions to logic, philosophy of logic,
Buddhist philosophy as well as Chinese philosophy.
Wnrrrrn III, S.vtrr C. is Professor of Philosophy at the University
of Connecticut, USA. He received a B.A. from Carleton College
and a Ph.D. in Philosophy from Princeton in 1970. He has pub-
lished articles on vagueness, logical form, philosophy of language,
ethics, political philosophy, Plato, metaphysics, and deconstruction.
He is the editor of Public Aairs Quarterly. His book, Deconstruction as
Analytic Philosophy was published by Stanford University Press in 2000.
Woxo, D.\ir B. is a Professor and Chair of the Department of
Philosophy, Duke University, USA. He received his B.A. from
Macalester College and his PhD from Princeton University. Repre-
sentative publications include Moral Relativity (1984), Confucian Ethics:
a Comparative Study of Self, Autonomy and Community (co-editor with
Kwong-loi Shun), Coping with Moral Conict and Ambiguity,
Ethics (1992), Relational and Autonomous Selves, Journal of Chinese
Philosophy (2004), Dwelling in Humanity or Free and Easy Wandering?
in Technology and Cultural Value, ed. P. Hershock et al (2003), Pluralistic
Relativism, Midwest Studies in Philosophy (1996).
Xi.o, Y.xo is Assistant Professor of Philosophy at Kenyon College,
USA. He received his Ph.D. from the New School for Social Research
in 1999, and was a Post-doctoral Fellow at UC Berkeley in 19992000,
and a Post-doctoral Fellow at the Fairbank Center at Harvard in
200203. His most recent publication is How Confucius Does Things
with Words: Two Paradigms of Hermeneutic Practice in the Analects
and Its Exegeses, forthcoming in Journal of Asian Studies.
Znrxo, Yti.x is Associate Professor in the Philosophy Department,
Lingnan University, Hong Kong. With BS degree in engineering
mechanics, he turned to philosophy of science at the MA level in
China, and nally got his PhD in philosophy from Bowling Green
State University, Ohio, USA. He has numerable paper publications
in the overlapping areas of dynamic rational choice theory, philos-
ophy of mind and action, and moral philosophy. His current research
interests include evolutionary and naturalist account of normativity
or emergence of intentionality.
Donald Davidson and the volume editor, Bo Mou, at Davidsons UC Berkeley oce
discussing this anthology project on July 17, 2003, about one month before his
passing away. (Photographed by Annie Ren)
By constructive engagement I mean philosophical inquiry into how, via reective
criticism and self-criticism, distinct modes of thinking, methodological approaches
or points of view in dierent philosophical traditions or within (the complex array
of dierent approaches of ) the same tradition, can learn from each other and make
a joint contribution to the common philosophical enterprise.
By Chinese philosophy here I primarily mean various movements of philo-
sophical thought in China from the Zhou dynasty (roughly eleventh century to 256
B.C.) through the early Qing dynasty (1644mid 19th century) and their contem-
porary studies and developments.
By Western analytic philosophy or Western philosophy in the analytic tradi-
tion I mean a Western mainstream philosophy from Socrates, Plato and Aristotle
via Descartes, British empiricism and Kant to the contemporary analytic movement.
Bo Mou
In this theme introduction to the project titled Davidsons Philosophy
and Chinese Philosophy: Constructive Engagement, I intend to do
three things. First, I introduce the background, nature and theme of
the project and explain the signicance and value of the construc-
tive engagement between Davidsons philosophy and Chinese phi-
losophy. Second, I examine some relevant meta-philosophical and
methodological issues that are closely related to the current project,
including those concerning the distinct orientations and purposes of
comparative studies and their due relations, how to look at the ade-
quacy of a prospective project in comparative studies, etc. Third, I
introduce the major points and distinguishing approaches of the con-
tributors essays and the rationale by which those essays are orga-
nized in this volume.
It is especially philosophically interesting and challenging to investi-
gate how a constructive engagement
between Chinese philosophy
and Western analytic philosophy
is possible. For one thing, Chinese
2 . +nrvr ix+nortc+iox
philosophy and Western philosophy in the analytic tradition are two
major philosophical traditions that have made many distinct and
signicant contributions. For another thing, the two philosophical
traditions have been considered by many to be remote, alien or even
opposed to each other; some in each tradition have taken philo-
sophical practice in the other tradition to have merely marginal value
or regard the two traditions as being essentially alien to each other.
Indeed, some mistaken or at least seriously misleading stereotypes
have resulted from one partys ignorance, or lack of in-depth inves-
tigation, of the other partys philosophy, while others have resulted
from failure to recognize the genuine nature of even ones own tra-
dition or from some theoretical conations. Today more and more
philosophers in both traditions have realized that Chinese philoso-
phy (or the philosophical dimension of Chinese thought) and Western
philosophy (including its analytic tradition) are not essentially alien
to one another: they have common concerns with a series of fun-
damental issues and have taken their characteristic approaches to
them; thus they can learn from each other and jointly contribute to
the common philosophical enterprise through constructive dialogue
and engagement. Some systematic meta-philosophical discussions,
especially in view of constructive engagement of Chinese philosophy
and Western philosophy in the analytic tradition, have been carried
out concerning the nature of philosophy in dierent philosophical
traditions and the issue of comparative philosophical methodology.
These have made constructive preparation at the level of meta-philo-
sophical theory and at the level of reective practice for carrying
out further in-depth investigations, like the current project, of exactly
how the Chinese philosophy and the Western philosophy in the ana-
lytic tradition can jointly contribute to the common philosophical
Note that, besides indicating a historical connection between Western philosophy
in such a tradition and some methodological approaches taken in this tradition,
such phrases as Western analytic philosophy and Western philosophy in the ana-
lytic tradition used by this writer are not intended to imply that those method-
ological approaches are, intrinsically or conceptually, exclusively connected with
Western philosophy.
One recent result from such eorts is the anthology volume, Two Roads to
Wisdom?Chinese and Analytic Philosophical Traditions, Open Court, 2001, edited by Bo
Mou and given a foreword by Donald Davidson. It is noted that the eorts in this
connection have been made in the past dozens of years in some ways; for the sake
of the purpose of this Introduction I do not plan to review them here.
. +nrvr ix+nortc+iox 3
In the aforementioned background, and in view of the need and
signicance of constructive dialogue and engagement of Chinese and
Western philosophy, one eective way to carry out such studies is
to focus on one philosophically signicant gure or one signicant
movement of philosophical thought in either Chinese or Western
tradition in constructive comparison with various relevant thoughts
and strands in the other tradition. It has been rendered especially
philosophically interesting, rewarding, and signicant to carry out a
case investigation of constructive engagement between the philoso-
phy of Donald Davidson in the Western analytic tradition and Chinese
philosophy for a number of theoretical considerations.
First, Donald Davidson (19172003) is known as one of the most
important and inuential philosophers in the twentieth century whose
works involve a series of fundamental issues in philosophy. There
is no more creative or systematic philosopher at work in America
today than Donald Davidson . . . Davidson has already constructed
one of the most remarkable pillars of sustained philosophical rea-
soning to be found in any era (Ian Hacking).
Both those philoso-
phers who endorse Davidsons views and those who oppose (some
of ) his views have unanimously agreed that Davidsons philosophy
is one of the most creative sources to stimulate their philosophical
reections in depth. Davidsons works involve a series of fundamental
issues and concerns in philosophy many of which various thinkers
in the Chinese philosophical tradition have also explicitly or implic-
itly addressed and somehow made their distinct contributions to;
those issues include (but are not limited to): (i) the relations between
language, thought, and reality; (ii) philosophical issue of truth; (iii)
meaning and reference; (iv) understanding and interpretation; (v)
knowledge and objectivity; (vi) actions and events; (vii) philosophy of
mind; (viii) problem of human rationality; (ix) irrationality and prac-
tical reasoning; and (x) the issue of metaphors. Their constructive
engagement on those issues would jointly contribute to our under-
standings and approaches to them.
Second, one distinct portion of Davidsons philosophy concerning
the conceptual scheme, rationality, etc. has its signicant implication
to the relation and engagement among distinct modes of thinking,
The citation is from the back cover of Donald Davidson (2001), Subjective,
Intersubjective, Objective, Oxford: Clarendon Press.
4 . +nrvr ix+nortc+iox
methodological approaches or points of view in dierent philosoph-
ical traditions including Chinese and Western philosophies. On the
other hand, there are some signicant meta-philosophical thoughts
in Chinese philosophy in regard to how to look at dierent approaches
and points of view. It would be especially philosophically interesting
and signicant to investigate how some signicant meta-philosophi-
cal ideas in Davidsons philosophy and in Chinese philosophy can
contribute to (our understanding of ) constructive engagement among
distinct modes of thinking, methodological approaches or points of
view in dierent philosophical traditions as well as within (the com-
plex array of dierent approaches of ) the same tradition.
Third, because Davidson is a well-respected and inuential philoso-
pher in contemporary Western philosophy, a positive research result
from this test case study would play a positive or even strong exem-
plary role for further constructive engagements of this kind both in
view of methodological approach and in regard to substantial treat-
ment of some fundamental issues and concerns in philosophy.
Fourth, this project would provide an eective and unique way
to look at how the philosophy of one of the major gures in the
Western analytic tradition has crossed cultural and national bound-
aries to contribute to the common philosophical enterprise.
There is another contributing consideration. Though it seems to
be less theoretical and more related to Davidsons style of doing phi-
losophy, this consideration has indeed contributed to the fashion in
which and means by which the foregoing constructive-engagement
agenda has been implemented in regard to Davidsons philosophy.
It is Davidsons very positive and encouraging attitude towards
reective criticism and challenge, insofar as my personal contact with
him can tell. Let me start with some small things that I have per-
sonally experienced. About eighteen years ago, around 1987, when
I still worked at the Institute of Philosophy, Chinese Academy of
Social Sciences, I selected, and translated into Chinese, thirteen rep-
resentative essays of Davidsons writings, mainly from his Inquiries into
Truth and Interpretation (rst edition, 1984), to give a more or less sys-
tematic introduction of Davidsons thought into the Chinese philo-
sophical circle as well as for the sake of my personal research plan
to have a more precise understanding of his thought.
When trans-
The result of this work is a collection, Truth, Meaning, Actions and Events: Selections
. +nrvr ix+nortc+iox 5
lating one of the essays from his preceding volume, I thought there
was an error in the text; but I felt hesitant about this. For the sake
of translation quality, I wrote Davidson consulting him about it.
Davidson gave a prompt response and agreed that it is a real error.
I would have forgotten this if it were not for Davidsons formal
acknowledgement of this in his Preface to the Second Edition of
the above volume (second edition, 2001). During 19992000 when
I did research at UC Berkeley, I discussed with Davidson about how
to look at the nature of the enumerative character of Tarskis truth
denition. Though we have dierent views on the issue, Davidson
always patiently explained his position and listened to my explana-
tion of my view without a dismissive attitude.
Such things in my
personal contact with Davidson have convinced me that he sincerely
welcomes reective criticism and other challenges for the sake of
constructive engagement. Indeed, Davidson considers this as one cen-
tral feature of the analytic method as he understands and practices
it. The analytic method in philosophy, Davidson wrote, is a method
that starts with a question or a doubt and tries to nd an answer
or to resolve the doubt. This sets in train attempts to nd reasons
for or against theses that suggest themselves as answers to the ques-
tions or resolutions of the doubts. The analytic method can engage
with ideas at any level and from whatever quarter or discipline or
tradition. It provokes argument and when practiced with an open
mind it engenders dialogue. At its best, dialogue creates mutual
understanding, fresh insights, sympathy with past thinkers, and, occa-
sionally, genuinely new ideas. But before there can be dialogue the
parties must meet. . . .
That is where we need to start for this con-
structive-engagement project: we meet with an open mind, for crit-
ical challenge and engagement, and to create mutual understanding,
fresh insights and genuinely new ideas. Davidsons style of doing phi-
losophy has inspired me to implement the foregoing constructive-
engagement strategy in an active elenchus style of dialogue with the
orientation of, or emphasis on, serious critical engagement instead
from the Philosophical Writings of Donald Davidson, edited and translated by Bo Mou,
Beijing, China: the Commercial Press, 1993 (in Chinese).
This disagreement is discussed in my article The Enumerative Character of
Tarskis Denition of Truth and Its General Character in a Tarskian System,
Synthese vol. 124, Issue 1 & 2, January 2001, pp. 91122.
See Davidsons Foreword to the anthology Two Roads to Wisdom?Chinese and
Analytic Philosophical Traditions, edited by Bo Mou, Open Court, 2001, p. v.
6 . +nrvr ix+nortc+iox
of a mere celebration, especially in Davidsons case. The suggested
project consists of, primarily, an anthology project and, secondarily,
an international conference project whose main purpose is to serve
as a platform for critical discussion and engaging challenge; the
anthology project is an independent project, instead of the confer-
ence proceedings, primarily for the sake of quality control, while the
conference project is indispensable for implementation of the con-
structive engagement strategy of the anthology project, though it also
serves some other purpose, being sensitive to situations and need
(say, to bring good international academic-exchange opportunities to
local scholars of the conference host region). In late 2001 I talked
with Davidson about the idea; he loved it, especially its emphasis
on the critical engagement of his thought instead of a mere cele-
bration. He indicated that he was more than happy not merely to
contribute one essay to the anthology project and deliver a talk at
the conference project but also to assume the role of commentator
for each of the speakers presentation papers and provide a reply to
each of the anthology contributors essays.
With all these background matters in view, let me highlight the
theme and objectives of this anthology project. The central theme
of this anthology project is to investigate (i) how Davidsons philos-
ophy and some thoughts and strands in Chinese philosophy could
jointly contribute to the common philosophical enterprise in some
philosophically interesting ways and (ii) how some signicant meta-
philosophical ideas in Davidsons philosophy and/or in Chinese phi-
losophy can contribute to (our understanding of ) constructive
engagement among distinct modes of thinking, methodological
approaches or points of view in dierent philosophical traditions as
well as within the same tradition.
The volume has the following three objectives. (1) Through the
Later on in 2002, as the International Society for Comparative Studies of
Chinese and Western Philosophy (ISCWP) was established with its emphasis of the
constructive-engagement, the conference project became one of its projects, i.e., the
1st ISCWP Constructive Engagement international conference, co-sponsored by
Institute of Foreign Philosophy, Peking University, which also assumed the confer-
ence host, and by the American Philosophical Associations Committee on International
Cooperation. The conference was originally scheduled for holding in the summer
of 2003. Because of the SARS outreach in Beijing and China in the spring of 2003,
the conference had to be postponed to the summer of 2004 with the Institute of
Philosophy, Chinese Academy of Social Sciences, as the conference host as well as
one of its co-sponsors.
. +nrvr ix+nortc+iox 7
preceding theme of the project, the volume is to investigate how
Davidsons philosophy and Chinese philosophy could jointly con-
tribute to the common philosophical enterprise and the constructive
engagement among dierent philosophical traditions in philosophi-
cally interesting ways. (2) Through this challenging test case of con-
structive engagement between Davidsons philosophy and Chinese
philosophy, the volume is to show how Chinese philosophy and
Western philosophy (including its analytic tradition) are not essen-
tially alien to one another: they have common concerns with some
fundamental issues and have taken their characteristic approaches to
some of those issues; they can learn from each other and jointly con-
tribute to the common philosophical enterprise in complementary
ways. (3) Through (1) and (2) above, this project is to show how the
constructive engagement in comparative studies is possible and how
such comparative methodology of constructive engagement concern-
ing distinct modes of thinking, methodological approaches or points
of view in dierent philosophical traditions or within the same tra-
dition is important or even indispensable in general philosophical
This anthology has its several distinguishing characteristics. (1) This
volume is the rst of its kind to investigate at an in-depth level how
a major gure in the Western (contemporary) mainstream philoso-
phy in analytic tradition and some thoughts and strands in Chinese
philosophy could jointly contribute to the common philosophical
enterprise in philosophically interesting ways. (2) This volume as
whole (and many an individual contributed essay in the volume) is
to investigate some fundamental issues and concerns in philosophy
from some distinct comparative approaches that would resort to con-
ceptual and explanatory resources from both the analytic tradition
and the Chinese tradition instead of merely from one tradition. (3)
Through this case study of constructive engagement, the volume has
shown how Chinese philosophy and Western mainstream tradition
in the analytic tradition are not essentially alien to one another: they
have many common concerns with a series of fundamental issues
and could jointly contribute to the common philosophical enterprise.
(4) Through (2) and (3) above, this volume has shown how the con-
structive engagement in comparative studies is possible and how such
comparative methodology of constructive engagement is important
or even indispensable in general philosophical inquiry. (5) This vol-
ume, through the foregoing (1) and (2), can play a positive or even
8 . +nrvr ix+nortc+iox
strong exemplary role for further constructive engagement of this
kind both in view of methodological approach and in regard to sub-
stantial treatment of some fundamental issues and concerns in phi-
losophy. (6) All the contributed essays in this volume are previously
unpublished pieces, written expressly for this volume, and unavail-
able anywhere else.
As indicated above, Davidson himself originally planned to con-
tribute to this anthology by oering one unpublished essay and pro-
viding his reply to each of the contributors essays. His participation
in this anthology project actually went much beyond this formal
commitment to his prospective writings in print. During the two
years since I set out to work on this project in 2001, Davidson and
I had quite a few discussions via emails and get-together meetings
on various things involved.
Indeed, his style of doing philosophy
and taking care of some other relevant things is truly encouraging
in various ways for eectively carrying out a project of this kind. In
my last meeting with Davidson at his oce on July 17, 2003, I
brought to him copies of several contributors essays; he told me that
he planned to use the coming month to work out at least a detailed
outline of his own contributing essay to the volume, while reading
and thinking of these contributors challenging essays, so that I can
use this outline as part of the proposal package to contact a prospec-
tive publisher. Unfortunately, his sudden passing away in the next
month, on August 30 of 2003, makes it impossible for us to share
and learn his thought from his planned contribution writings to this
volume and see how Davidson would further develop his ideas when
facing criticism and challenges. It is a sad loss to this anthology pro-
ject whose successful completion Davidson himself seriously looked
forward to, as everyone else who has participated in the project does.
However, with the rm support from all the contributors in this
volume, the anthology project together with its engagement platform
plan continued; the contributors have had their consensus: besides
its value on its own, the successful completion of the anthology pro-
ject in a critical-engagement manner which Davidson favors would
be a tting memorial to Donald Davidson. All the 12 essays in this
I still clearly remember the situation of one discussion (in 2003) with Davidson
on some relevant things concerning this volume at a Starbucks coee shop at Solano
Avenue in Berkeley, located roughly in the middle between his home on the Berkerly
hill and my home at the Albany bay shore, which he considered to have a fair
driving distance to both.
. +nrvr ix+nortc+iox 9
volume have undergone their authors many revisions that have
beneted from the successful and eective critical discussion and each
others engaging challenges at the conference in June 2004. This
anthology is a collective achievement, which comes from all the
authors valuable contributions and from the rm support from all
the contributors and other participants in various ways as indicated
in the Acknowledgements page.
Having resulted from the critical engagement that Davidson most
favors, and with the mutual understanding, fresh insights and new
ideas which are generated from such critical engagement and to
which Davidson truly looked forward, this volume is dedicated to,
and in memory of, Donald Davidson, for his important contribution
to the common philosophical enterprise that has crossed cultural and
national boundaries, and for his inspiration and participation in the
current project in his unique way.
To understand the nature and signicance of the current project on
the constructive engagement between Davidsons philosophy and
Chinese philosophy, the reader is expected to understand several
major orientations with their distinct methodological approaches in
comparative studies and their due relations.
Without pretending to exhaust all working orientations, I intend
to highlight three major orientations and their distinct methodolog-
ical approaches in comparative studies whose due examination,
in my opinion, would be most helpful for a constructive develop-
ment of comparative philosophy.
I plan to do this by discussing
The major part of the content of this section is a further revision of the major
account in my article, Three Orientations and Four Sins in Comparative Studies,
the APA Newsletter, Fall 2002, Vol. 02, No. 1, pp. 4245, and of some passages on
certain involved methodological issues in the rst section of my article A Re-
Examination of the Structure and Content of Confucius Version of the Golden
Rule, Philosophy East and West, Vol. 54, No. 2 (April 2004), pp. 218248.
By comparative philosophy I mean not merely comparative studies of dierent
philosophical traditions but any comparative investigation concerning distinct modes
of thinking, methodological approaches (perspectives, guiding principles or instru-
ments) or substantial points of view in dierent traditions or within (the complex
array of dierent approaches of ) the same tradition, though what is focused on in
this volume is comparative engagement between Chinese philosophy and analytic
philosophical tradition via the case of Davidsons philosophy.
10 . +nrvr ix+nortc+iox
the appropriateness of four sins that are oft-cited, explicitly or implic-
itly, in critically evaluating a comparative project. The reason that
I take this strategy is this: the appropriateness or legitimacy of the
four sins depends on the nature, purpose, and orientation of a com-
parative project that would decisively determine which kind of method-
ological approach should be taken and what kind of expectations
are appropriate; the strategy is an eective way to identify how cru-
cial aspects and purposes of those orientations and approaches are
distinct and so, in treating ones own comparative project or criti-
cally evaluating some others comparative project, to be more sen-
sitive to its distinct purpose and orientation and thus to what it is
appropriate to expect.
When comparative projects are critically evaluated, there seem to
be four sorts of complaint. The alleged sins are these: (1) over-
simplication; (2) over-use of external resources; (3) exaggerated dis-
tinction; and (4) blurring assimilation. They, or some of them, are
sometimes taken for granted in two ways: rst, it is thought that any
simplifying the object of study or using external resources to char-
acterize it is doomed to be excessive and thus deserve to be charged
with negative over-character; second, it is assumed that the four
complaints may be made indiscriminately in evaluating any com-
parative project without regard to the orientation and methodolog-
ical strategy of that study. A metaphilosophical examination of the
four sins will help to eectively identify the distinct character and
objectives of a variety of orientations and their approaches.
The rst orientation under examination aims to give a historical
and descriptive account. That is, the primary concern and purpose
of this type of comparative study is to accurately describe relevant
historical matters of facts and pursue what thinkers in comparison
actually thought, what resources were actually used (by them), and
what appear to be similar and dierent. The orientation of this type
of comparative study thus might be called historical orientation,
and its methodological approach aims at accurate description of his-
torical matters of fact. The historical orientation requires its practi-
tioners to cover a vast range of historical data to give such factual
description. It seems that this orientation and its methodological
approach are typically taken in Chinese studies or Sinology as the
primary approach to Chinese and comparative philosophy.
There is no wonder that the aforementioned four oft-cited sins
would be assumed relevant to those comparative projects with the
. +nrvr ix+nortc+iox 11
historical orientation. First, to accurately describe something, it is
taken for granted that one should not simplify what is actually com-
plicated; in other words, simplication is always oversimplication:
any simplication is guilty of being negatively excessive; and
simplication is thus identical with falsication. Second, as for over-
use of external resources, any conceptual or explanatory resources
which are used to interpret a thinkers idea under examination but
were not actually used by the thinker herself are rendered inade-
quate or excessive: use of external resources is always over-use of
external resources.
Third, in this approach, exaggerated distinct-
ness often results from over-simplication of one or both parties
under comparative examination in the direction of ignoring part(s)
in one tradition or account that would share something common in
another tradition or account; in this way, insofar as the sin of over-
simplication has been already legitimately charged, the charge of
exaggeration of the due distinction (if any) between the two would
be appropriate. Fourth, in this approach, blurring assimilation often
results from over-use of external resources to interpret one or both
parties under comparative examination, especially when the external
resources used to characterize one party come from the other party;
to this extent, insofar as the sin of over-use of external resources has
been already legitimately charged, the resulting assimilation of blur-
ring the distinction between the two would be also adequately charged.
There would be nothing wrong or inadequate with the historical
orientation and its methodological approach per se, when the orien-
tation/approach is adequately taken as one of a number of alter-
native orientations/approaches, instead of the exclusive one, and
when one can see its limitations in serving other distinct purposes
in comparative studies. In view of this, one question would be nat-
ural: Are there any orientations and approaches other than the his-
torical orientation that would be adequate, and, more importantly,
necessary in view of certain purposes in comparative studies? With
a positive answer to the question actually being presupposed in the
preceding discussion, the question can be phrased in another way:
By external resources I mean those resources that were not actually used by
the ancient thinker under discussion when the resources are identied from the his-
torical point of view or with the historical orientation. Nevertheless, as I explain
later, using the very term external in some situations would simply miss the point
in regard to the purpose of the third orientation to be discussed.
12 . +nrvr ix+nortc+iox
How are other legitimate orientations and methodological approaches
possible and necessary? In the following, I focus on two other ori-
entations and their respective methodological approaches.
The second orientation in comparative studies is concerned with
through elaborating a thinkers ideas under exami-
nation; the primary purpose of this orientation is to enhance our
understanding of a thinkers ideas via some eective conceptual and
explanatory resources, whether or not those resources were actually
used by the thinker herself. It is clear that a purely historical approach
does not t here: To elaborate and understand the thinker does not
amount to guring out exactly how the thinker actually thought;
instead, such interpretation and understanding might include the
interpreters elaboration of the implications of the thinkers point,
which might not have been considered by the thinker herself, or the
interpreters representation of the thinkers point in clearer and more
coherent terms or in a more philosophically interesting way, which
the thinker herself might have not actually adopted.
In both cases,
given a thinkers ideas (in one tradition or account) under interpre-
tation, some eective conceptual and explanatory resources well devel-
oped in another tradition or account are consciously used to enhance
our understanding of, and elaborate, the thinkers ideas; those resources
used are thus tacitly and implicitly, but constructively, in compari-
son and contrast to those original resources by means of which the
insight or vision was somehow delivered, insofar as such compari-
son of the two distinct sorts of resources is not expressly and directly
conducted. The term constructively here means such tacit com-
parative approach intrinsically involves how the interpreter of the
Here I use the term interpretation in a narrow or straightforward sense as
specied here (in terms of elaborating and understanding) rather than in a broad
or implicit sense in which all the three orientations discussed here could be some-
how identied as interpretation-concerned.
Then, can these implications be said to belong to the thinkers ideas in the
text (and thus fall into what the thinker truly means/meant or what the thinkers
ideas truly has/had)? In an important sense, the answer would be yes; for these
implications are truly implied by the ideas delivered by the thinker, although one
can surely say that these implications were not actually expressed by the thinker,
and one thus might say that they are not what the thinker actually (or truly?)
means/meant. (At this point, one can see that such expressions as what a certain
thinker truly means/meant or what she truly has/had tend to be ambiguous and
vague and thus deserve clarication, especially when one intends to make claims
about what a thinker truly means/meant or what her ideas truly has/had.)
. +nrvr ix+nortc+iox 13
thinkers ideas could learn from another tradition or account regard-
ing resources to enhance the interpreters understanding of the thinkers
ideas; therefore, some constructive philosophical engagement between
distinct resources in dierent traditions is tacitly involved in this ori-
entation and its corresponding methodological approach.
In this way, the so-called over-use of external resources is not nec-
essarily a sin but might really enhance our understanding of a thinkers
ideas or clarify some original unclear or confusing expression of her
ideas. Consequently, the endeavor per se of using external resources
in this orientation is not automatically inappropriate and thus is not
doomed to be a sin, as it would be in the historical orientation. Note
that, when those explanatory and conceptual resources are used, they
are not intended to assign the same degree of articulated systemati-
zation and of mastering some conceptual and explanatory resources
to an ancient thinker but to enhance our understanding of her ideas
delivered in the text. For this explanatory purpose, it is not merely
legitimate but benecial to employ more explicit or clearer concep-
tual resources to elaborate some otherwise implicit and hidden thing
(say, coherence and connectedness) in a thinkers ideas that was some-
times less clearly delivered or even ill-expressed for lack of those
contemporary explanatory and conceptual resources that are unavail-
able to the ancient but to us.
It is also noted that, when a thinkers
line of thought and her ideas lack in articulated systematicity in their
language expressions, that does not amount to saying that the thinkers
line of thought and her ideas per se go without (implicit and hidden)
coherence and connectedness deep in a thinkers ideas. Consequently,
we cannot base ourselves merely on this lack of articulated system-
aticity in language expression to judge that the thinkers text itself
is not a philosophical work when the text was indeed intended to
deliver her reective ideas. At this point, with the previous and cur-
rent methodological considerations, some further elaborations of the
thinkers line of thought and her surrounding reective ideas via ade-
quate conceptual and explanatory resources available to us is gen-
uinely needed, instead of being the mere issue of preference, for the
It is another matter when a thinker intentionally uses some seemingly para-
doxical remarks to make some points. However, such occasions imply neither that
the ideas delivered by these remarks per se are actually incoherent nor that the
points in question could not be delivered eectively in clearer terms without para-
doxical appearance.
14 . +nrvr ix+nortc+iox
sake of enhancing our understanding of the thinkers ideas includ-
ing their due implications.
As indicated in discussing the historical orientation, blurring assim-
ilation might result from over-use of external resources when inter-
preting one or both parties under comparative examination, especially
when the external resources used to characterize one party come
from the other party. But, for the purpose of interpretation, the
resulting assimilation is not necessarily a sin but might illuminate the
essential connection and common points between the assimilated
ideas at the fundamental level so as to enhance our understanding
of those ideas.
It is clear that a comparative project with the interpretation-con-
cerned orientation, instead of the historical orientation, is free, or
indeed tends, to focus on a certain aspect, layer or dimension of a
thinkers ideas based on the purpose of the project, the reective
interest of the person who carries out the project, etc. Indeed, instead
of a comprehensive coverage of all aspects or dimensions of the
object of study, focusing on one aspect or dimension is a kind of
simplication. Now the question is this: Is any simplication per se
doomed to be indiscriminately a sin of over-simplication? It should
be clear that, if the purpose of a comparative project is to focus on
interpreting or elaborating one aspect or dimension instead of pre-
tending to giving a comprehensive historical description, charging
the practitioner of this project with over-simplication or doing some-
thing excessive in simplifying the coverage into one aspect or dimen-
sion would be both unfair and miss the point.
Let us agree that a comparative project should be guided by some
comprehensive understanding. But a comparative project taking a
certain methodological perspective through focusing on one aspect
of the object of study is not incompatible with a comprehensive
understanding. At this point, what needs to be recognized is an
important distinction between a methodological perspective as the
current working perspective and the methodological guiding princi-
ple that an agent presupposes when taking the methodological per-
spective and that would be used by the agent to guide or regulate
how the current perspective would be applied and evaluated in view
of some other relevant perspective(s). Ones reective practice per se
of taking a certain methodological perspective amounts to neither
reectively rejecting some other relevant methodological perspective(s)
nor presupposing an inadequate methodological guiding principle
. +nrvr ix+nortc+iox 15
which would render irrelevant other relevant methodological per-
spectives (if any).
We have discussed three sins (i.e., over-simplication, over-use
of external resources, and blurring assimilation) that might be
charged against a comparative project with the interpretation-concerned
orientation. How about the other one, the sin of exaggerated dis-
tinction? This case is more complicated than what it may appear.
This sin, as discussed before, is connected with the sin of over-
simplication when the comparative project assumes the historical
orientation. But when a comparative project takes the interpretation-
concerned orientation and does simplify the object of study by focus-
ing on one aspect of the object of study, is it automatically guilty
of the sin of exaggerated distinction? The preceding distinction
between the methodological perspective and the methodological guid-
ing principle is helpful here again. What is at issue is whether the
interpreter has assumed an adequate methodological guiding princi-
ple to guide and regulate how to look at the relation between the
current methodological perspective used as a working perspective and
other relevant methodological perspective(s) that would point to other
aspects of the object of study. Consequently, when one evaluates a
comparative project, what really matters is for one to look at what
kind of methodological guiding principle is presupposed behind the
working perspective; only when this is examined can the charge of
exaggerated distinction be adequately evaluated.
Now let us move onto the third orientation under examination, i.e.,
constructive-engagement-concerned orientation aiming at joint con-
tribution to common philosophical issues. The primary purpose of
this orientation in comparative studies is to see how, through reective
criticism and self-criticism, both sides under comparative examina-
tion could jointly and constructively contribute to some commonly
concerned issues of philosophy,
rather than to focus on providing
For a detailed and systematic discussion of the distinction between the method-
ological perspective and the methodological guiding principle and its implications,
see Bo Mou (2001), An Analysis of the Structure of Philosophical Methodology
in View of Comparative Philosophy, in Bo Mou ed. Two Roads to Wisdom?Chinese
and Analytic Philosophical Traditions, Chicago: Open Court, pp. 337364.
It is arguably right that many issues that were traditionally identied as some
unique issues in dierent traditions have turned to be concerned primarily with
dierent aspects, layers or dimensions of some commonly concerned, more general
issues of philosophy, especially from a more broadly philosophical vantage point.
This is one point that I have endeavored to make and illustrate in my several writ-
ings mentioned above.
16 . +nrvr ix+nortc+iox
a historical or descriptive account of each or on interpreting some
ideas historically developed in a certain tradition or account. Typically,
in comparatively addressing a certain commonly concerned issue of
philosophy, some substantial ideas historically developed in distinct
philosophical traditions or accounts are explicitly and directly com-
pared with the aim of showing how they could jointly and comple-
mentarily contribute to the common concern in some philosophically
interesting ways. Insofar as constructive engagement in dealing with
various common concerns and issues of philosophy is most philo-
sophically interesting, this comparative orientation and its method-
ological strategy directly, explicitly and constructively conducts
philosophical engagement and is thus considered to be most philo-
sophically interesting. To highlight the characteristic features of a
comparative project with this as its primary orientation, let us exam-
ine the appropriateness of three charges, among the aforementioned
four, that have been sometime or even often brought against com-
parative projects with this orientation, that is, the sin of over-
simplication, the sin of over-use of external resources, and the sin
of blurring assimilation.
A typical procedure of conducting a philosophically constructive
engagement in such comparative projects could be both conceptu-
ally and practically divided into three phases: (i) the pre-engagement
phase in which certain ideas in dierent traditions or accounts that
are relevant to the common concern under examination and thus to
the purpose of the project are focused on and identied; (ii) the
engagement phase in which those ideas internally engage with each
other in view of that common concern and the purpose to be served;
and (iii) the post-engagement phase in which those distinct ideas from
dierent sources are now absorbed or assimilated into a new approach
to the common concern under examination. The three sins afore-
mentioned may be considered to be typically associated with dierent
phases. The sin of over-simplication regarding a certain idea
identied from a certain tradition may be typically associated with
reective eorts in the pre-engagement phase; the sin of over-use
of external resources regarding elaborating a certain idea from a cer-
tain tradition may be typically associated with reective eorts in
the engagement phase; and the sin of blurring assimilation may be
typically associated with reective eorts in the post-engagement
phase. Now let me briey evaluate the appropriateness of the three
sins respectively in the corresponding three phases; looking at the
. +nrvr ix+nortc+iox 17
sins in this way will help to highlight features of comparative pro-
jects primarily with the constructive-engagement-concerned orientation.
(1) In the pre-engagement phase, it might be not only legitimate
but also adequate or even necessary to have simplication and abstrac-
tion of some ideas in one tradition or account into such a perspec-
tive: this perspective per se is presented in most relevant terms to the
common concern addressed, and the purpose served in a construc-
tive-engagement-concerned comparative project, while without involv-
ing those irrelevant elements in the tradition or account from which
such a perspective comes, though those irrelevant elements in that
tradition might be relevant to guring out the point of those ideas.
The reasons are these. First, the primary concern of the project is
not with how such an idea is related to the other elements in the
source tradition or account but with how it is relevant to approach-
ing the commonly concerned philosophical issue. Second, while one
needs to understand the point of an idea in the context in which it
was raised, once one understands the point (either through employ-
ing data provided by projects with the rst two orientations or through
ones own background project with one of the rst two orientations),
there would be no present purpose served by discussing background.
Third, it is clear that such an approach per se does not imply deny-
ing the social and historical integrity of the idea in the source tra-
dition; the point is that the existence of such integrity cannot
automatically guarantee an indiscriminate priority or even relevance
of expressly addressing it in any comparative projects without regard
to their orientations and purposes.
(2) In the engagement phase, relevant perspectives from dierent
source traditions would constructively engage each other. From each
partys point of view, the other party is something external without;
but, from a more broadly philosophical vantage point and in view
of the commonly concerned issue, the distinct views may be com-
plementary within. In this context, the term external would miss
the point in regard to the purpose here: the pivotal point is not this
or that distinct perspective but the issue (and its comprehensive
approach) to whose various aspects those perspectives point; in view
of the issue, all those perspectives are internal in the sense that they
would be complementary and indispensable to a comprehensive
(3) In the post-engagement phase, some sort of assimilation typi-
cally results from the preceding constructive engagement; that is,
18 . +nrvr ix+nortc+iox
such assimilation would adjust, blur and absorb dierent perspec-
tives into one new approach as a whole; this would be what is really
expected in this sort of constructive engagement in comparative stud-
ies, instead of a sin.
It should be noted that, if a comparative project which explicitly
has one of the preceding orientations is considered as a project-
simplex in comparative studies, a comparative project in philosoph-
ical practice might be a complex that goes with a combination of
two or more orientations. For example, a comparative project con-
cerned with a historical gure often consists of such a combination.
Recognition of the characteristic features of the above three distinct
comparative orientations and their respective methodological approaches
would help us discriminatively treat dierent stages or parts of a
comparative project-complex.
Traditionally, to my knowledge, comparative projects with the
above third and second orientations (especially when resorting to
contemporary development and resources of philosophy) have yet to
receive due emphasis for some reasons. First, as far as comparative
projects regarding Chinese and Western philosophies are concerned,
a comparative project sometimes tends to be taken as a mere by-
product or extension of studies of the classical Chinese philosophy
which itself sometimes tends to be taken largely as merely historical
studies of the history of (the classical) Chinese philosophy. Second,
on the other hand, comparative approach as a methodological
approach has not yet been considered primarily as an eective
approach to doing philosophy per se. Third, the aforementioned four
sins (especially, those of oversimplication, over-use of external
resources and blurring-assimilation) have been more or less con-
sidered as some taken-for-granted sins and have thus discouraged
reective eorts in the direction of the third orientation (or even the
second orientation) which would often unavoidably but appropriately
commit those sins in many cases. Fourth, most importantly, Chinese
philosophy and Western philosophy (especially its mainstream tradi-
) are sometimes taken as being essentially alien to one another;
this kind of mentality would undermine or pre-empty any serious
I use the plural form of the term mainstream tradition to indicate that the
identity of a mainstream tradition is not exclusive but sensitive to regions and times
throughout the history of Western philosophy. In the twentieth century, the ana-
lytic tradition is a mainstream tradition in the English speaking countries while the
Continental tradition is a mainstream tradition in the European Continent.
. +nrvr ix+nortc+iox 19
reective eorts in comparative projects with the third orientation
and, in my opinion, negatively contribute to prejudice Western philoso-
phers as well as some scholars in studies of Chinese and compara-
tive philosophy may assume that Chinese philosophy is not philosophy
in the sense of the term philosophy that is intrinsically related to
a series of fundamental concerns and issues as addressed in Western
philosophy (especially its mainstream traditions).
Now, as more and more philosophers in the elds of Chinese and
comparative philosophy have a holistic understanding of Western
philosophy (both its past and its contemporary development, both
its appearance and its deep concerns, and both its distinct working
perspectives and its guiding principles at a deep level) and become
constructively engaged with Western philosophy on a series of fun-
damental common concerns and issues, it is more widely agreed
among philosophers who are familiar with both Chinese and Western
philosophies, as already indicated in the rst part, that they are not
essentially alien to one another: they have common concerns with
a series of fundamental issues in philosophy and have taken their
characteristic approaches to them. They thus could learn from each
other and jointly contribute to the common philosophical enterprise
through constructive dialogue and engagement. Consequently, there
is serious need to emphasize comparative projects of the third and
second orientations, though this emphasis certainly would not deny
the legitimacy and due value of the rst orientation as one eective
approach but stress its constructive compatibility with the other
The current comparative project Davidsons Philosophy and Chi-
nese Philosophy: Constructive Engagement, as highlighted by the
subtitle, is the one primarily with the aforementioned third orientation.
The main text of the anthology consists of twelve essays contributed
by experts in relevant areas of study, most of whom are established
scholars. With the foregoing constructive-engagement-concerned ori-
entation aiming at joint contribution to some commonly concerned
issues of philosophy, the main text of the book focuses on ve issues
or topics which are respectively the subjects of the following ve
parts into which the twelve contributed essay are organized:
20 . +nrvr ix+nortc+iox
Part One: Conceptual Schemes, Relativism, and Cross-cultural Under-
Part Two: Principle of Charity and Chinese Philosophy;
Part Three: Rationality, Normativity, and Inter-cultural Disagreement;
Part Four: Meaning and Interpretation;
Part Five: Truth Concern and Dao Concern.
Two notes are due. First, these topics are both reectively interest-
ing on their own and most conducive to constructive-engagement-
concerned comparative treatment in view of Davidsons philosophy
and Chinese philosophy. Nevertheless, in so saying, that does not
mean that the current ve topics have ever exhausted all actual or
potential aspects of such comparative engagement of Davidsons phi-
losophy and Chinese philosophy; the current selection of the topics
more or less reects the points of interest where a certain level of
quality research results have been generated in the current scholar-
ship in this connection. Second, because of the comprehensive nature
of some of the contributed essays, the way to organize the entries
is not exclusive. The reader who focuses on reading the entries on
one topic (say, those in Part One) can thus have her cross reference
to some other entries in another part (say, Part Two and Part Three).
In the following, let me sketch the organizational strategy and out-
line how the entries are related to the thematic topics of the parts
into which they are organized.
One of the most philosophically interesting and signicant points
made by Davison that is closely relevant to the current enterprise
of comparative Chinese-Western philosophy with the foregoing
constructive-engagement orientation is Davidsons point in his argu-
ment against the very notion of conceptual schemes, the dualism
between scheme and content, and thus conceptual relativism, which
was rst systematically put forward in his well-known essay On the
Very Idea of a Conceptual Scheme (1974). The point is reassured
and sharpened in his Foreword to a recent volume on the rela-
tion between Chinese and analytic philosophical traditions in view
of cross-cultural understanding and constructive engagement in
We seldom stop to think how much, as philosophers, we share with
other philosophers from other ages, other countries, other traditions.
We tend to discover our common problems and interests as we read,
teach, and travel. The discovery surprises us for, to begin with, minds
are best compared by nding as many points of similarity as every-
. +nrvr ix+nortc+iox 21
day patterns of action and reaction aord. But once this tting of pat-
tern to pattern is accomplished, the remaining dierences loom out of
proportion. This perhaps explains why a rst exposure to a new tra-
dition seems to reveal an unbridgeable gap. What experience shows,
though, is that, as in other areas, dierences are to be understood only
as seen against a background of underlying agreement. The underly-
ing agreement may be largely unspoken and unnoticed, but it is always
available. Sometimes we need help in appreciating how philosophy
builds on what we all know. No world views or conceptual schemes
are truly incommensurable.
To Davidson, cross-cultural mutual understanding in philosophy is
certainly possible; any disagreement between philosophers from dierent
traditions can be understood only against a background of underly-
ing agreement. Davidson does not intend to deny the diversity; rather
he considers such diversity as the life of the philosophy. Davidsons
point is by no means to indiscriminately seek consensus and con-
formity but celebrates variety and dierences as shown in distinctive
perspectives to look at various aspects of the object of study for the
sake of a comprehensive understand, as Davidson says:
. . . we should not seek conformity in philosophy. On the contrary, in
our intellectual work we should celebrate variety and do all we can
to insure its survival. We should not make the mistake of supposing
that under ideal circumstances our institutions and our philosophical
proclivities would or ought to become more alike. If anything, we
should expect the opposite, and we should welcome it. We welcome
understanding, and with it tolerance, but variety and dierence is at
the heart of philosophy.
In the citation above, when Davidson emphasizes that No . . . con-
ceptual schemes are truly incommensurable, this saying seems to
presuppose that there are conceptual schemes or it seems to be com-
patible with the notion of conceptual schemes. However, Davidsons
position is stronger than what the preceding statement appears to
suggest: for the sake of arguing against relativism, Davidson opposes
the very idea of a conceptual scheme, which is presupposed by rel-
ativism as standardly understood, in the sense that he renders this
notion incoherent; another sort of argument against relativism which
Davidson (2001), Foreword, in Bo Mou ed. (2001), Two Roads to Wisdom?
Chinese and Analytic Philosophical Traditions, Open Court, p. v.
Davidson, op.cit., pp. vvi.
22 . +nrvr ix+nortc+iox
Davidson oers is that which opposes the dualism between scheme
and content. However, is the notion of conceptual schemes really
incoherent? How is relativism related to the notion of conceptual
schemes and the dualism between scheme and content? What are
the due implications of Davidsons point here towards comparative
philosophy? Can Davidsons resources alone be sucient for devel-
oping a sound foundation for a creative comparative philosophy
which lies at the core of an emerging global philosophy? These ques-
tions will be examined in Part One of the volume, Conceptual
Schemes, Relativism, and Cross-Cultural Understanding, which con-
sists of three essays. In his essay Relativism and Its Schemes,
Michael Krausz agrees with Davidson in disallowing the scheme-
content dualism and thereby those relativisms that assume it. However,
it seems to the author, not all relativisms need assume the scheme-
content dualism; we need not understand schemes as giving form
to the data of sensation but as matrixes in terms of which expe-
rience and data themselves are to be understood. The author con-
tinues to argue that the notion of conceptual schemes per se is coherent:
persons of one scheme may understand persons of another, although
the two schemes are not fully translatable; that is, disagreeing with
Davidson, the author argues that non-translatability is not a sucient
condition for non-understandability (mutual understandability between
schemes does not entail translatability). It seems to Krausz, the prob-
lem with those standardly understood relativisms that are tied to con-
ceptual schemes concerns not their incoherence but their characteristic
adumbrance (typically it is unclear just what truth or rightness is
meant to be relative to).
The author, Koji Tanaka, of the second essay, Davidson and
Chinese Conceptual Scheme, focuses on elaborating Davidsons
unique insight into the incoherent nature of conceptual relativism
and exploring the consequence of this insight for those comparative
studies of Chinese and Western traditions of philosophy that pre-
suppose their having built up mutually non-inclusive stocks of con-
cepts and take this kind of conceptual relativism as the foundation
of comparative philosophy. Through his case analysis of the early
Confucian and Western approaches to morality, the author casts
doubt on the legitimacy of the foregoing comparative projects because
of their incoherent character. But the author also emphasizes that
the way of carrying out comparative studies that he criticizes is not
exhaustive and there may be other alternative ways. Indeed, as many
. +nrvr ix+nortc+iox 23
essays in this volume eectively show, a worthy comparative project
need not presuppose such a radical conceptual relativism as the foun-
dation of its work. The next essay in this part can be seen as one
constructive endeavor in that direction which tries to incorporate
Davidsons ideas and some other resources into a foundation for
reaching cross-cultural understanding and carrying out creative com-
parative philosophy.
As explained in the last section, most philosophically interesting
projects in comparative philosophy do things more than, or proceed
with the primary goal other than, descriptive comparison: based on
cross-cultural understanding, they seek integration, mutual challenge,
and joint contribution to the common philosophical enterprise, a
kind of synthetic project. In his essay Making Room for Comparative
Philosophy: Davidson, Brandom, and Conceptual Distance, Stephen
C. Angle explores to what extent Davidsons work can provide what
is needed for developing a sound foundation for such a kind of com-
parative studies in philosophy as synthetic projects in his terms. Given
how dierent philosophical traditions can be from one another,
though, such ambitions can seem hopeless; the author argues that
Davidsons work provides grounds for comparative philosophers to
overcome such worries. But, at the same time, the author maintains
that in certain respects, Davidson makes overcoming dierence seem
easier than it really is. The author then draws on Robert Brandoms
inferential and pragmatic account of semantics to develop a foun-
dation for comparative philosophy that, while still Davidsonian in
important ways, shows us both why the synthetic projects of com-
parative philosophy are possible, and also why they can be dicult:
conceptual dierences can be robust and important, even if radical
incommensurability is not in the ong. These synthetic projects lie
at the core of what might more properly be called an emerging
global philosophy, and they are well worth their trouble.
Part Two, Principle of Charity and Chinese Philosophy, focuses
on Davidsons principle of charity which plays a crucial role in
Davidons approach to the issue of how to understand and interpret
other agents including how to reach cross-cultural understanding.
The core idea of Davidsons principle of charity is this: we must
interpret others on the assumption that they are rational beings, nav-
igating the same world as we are. But how to understand the nature,
explanatory power, scope and limit of this principle? The two essays
in this part explore the issue in view of Chinese philosophy.
24 . +nrvr ix+nortc+iox
In his essay, Where Charity Begins, David B. Wong rst exam-
ines Davidsons own critical reection and changes of the statements
of the principle of charity, from maximization of agreement to opti-
mization of agreement, pointing out that this is the right move but
also that the requirement of optimization itself required interpreta-
tion; in so doing, the author criticizes overreaching attempts based
on the principle of charity to restrict the variety of belief and desire,
and then attempts to identify what real guidance the principle pro-
vides for the project of interpreting others. It seems to the author
that charity is less a denite principle of interpretation but rather
the assortment of the various ways we have of explaining others
talk and actions. We do interpret others by analogy to our own
beliefs and desires, but if the principle of charity bids us to inter-
pret others as being like us, we need to interpret who we are. The
author argues that the we is bound to be diverse in belief and
desire over human culture, and in particular in the range of values
that are central to particular cultures. It is argued that pragmatic
factors play in a role in the decision as to who counts as one of
us and that our very conception of what is rational results partly
from the mechanism of natural selection through genetic permuta-
tions, and that our ideal of rationality must open to the possibility
that a range of solutions to human problems may have equal claim
to be called rational. The authors strategy of argument consists in
working backwards, from interpretations of the moral tradition of
Confucianism from a contemporary American perspective; the question
addressed is how analogy and models of rationality operate within
these interpretations to increase our understanding of Confucianism.
The other essay in this part, Davidsons Charity in the Context
of Chinese Philosophy, is intended to test and defend the explana-
tory power of Davidsons principle of charity when facing some chal-
lenges in the context of Chinese philosophy. The author of the essay,
Yiu-ming Fung, discusses three alleged problems with Davidsons
principle of charity. The rst problem is A. C. Grahams doubt about
whether the principle can be survived in the context of comparison
between Chinese and English language and philosophy. The author
argues that Grahams criticism is fundamentally self-defeating; it can-
not be understood as a real challenge to Davidsons principle of
charity. The second problem is about the methodological character
of the principle of charity which many tend to negatively render
. +nrvr ix+nortc+iox 25
transcendental; the author argues that this is a misunderstanding
due to the critics failing to take seriously some of the concrete exam-
ples provided by Davidson and not realizing a crucial distinction
between the two kinds of transcendental argument in the literature.
Suzukis interpretation of Zen Buddhism is considered to present
another alleged challenge to Davidsons principle. The author argue
that Suzukis idea that giving up rational thinking and logic is a nec-
essary condition of attaining Zen cannot escape from its landing
problem, i.e., how to make sense the relation between two kinds of
truth or two levels of world; the authors strategy is to use this prob-
lem to force the Zen masters into a dilemma: either to explain the
relation with our rational language and logic or to fall into an abyss
of totally isolated from the real world.
The foregoing discussions of Davidsons rejection of the very idea
of conceptual schemes and relativism and his principle of charity in
view of Chinese philosophy all directly or indirectly point to Davidsons
idea of rationality. Furthermore, it is known that Davidsons thought,
though wide and complex, has a remarkably unitary character; what
unies Davidsons views on various issues is his fundamental under-
standing and characterization of the very idea of rationality and nor-
mativity. The two essays in Part Three, Ratinality, Normativity,
and Inter-Cultural Disagreement, explicitly address the issue and
make their insightful elaborations of Davidsons line especially in
view of Davidsons idea on the intrinsic connection between the nor-
mative and the factual (or the descriptive). They also explore how
a Davidsonian concept of normativity and rationality can contribute
to our understanding of disagreements between cultures. Both think
that such disagreements do appear between dierent cultures or
dierent paradigms; but such disagreements can make real sense only
against the background or overall validity of Davidsons approach
to rationality or normativity.
In his essay Davidsonian Rationality and Ethical Disagreement
between Cultures, Samuel C. Wheeler extends several Davidsonian
ideas in order to sketch a Davidsonian ethics. It is a reconstructed
Kantianism holding that the normativity implicit in rational inter-
pretation of another agent that maximizes agreement supplies the
basis for understanding the normativity of ethical concepts and that
the normative is not dierent in ontological or epistemic status from
the factual; it is thus a kind of ethical objectivism holding that ethical
26 . +nrvr ix+nortc+iox
sentences have truth value, even without a basis in pre-conceptual-
ized desire-stu. Nevertheless, Wheeler argues, this Davidsonian
account is dierent from Kants account. In contrast to obligation
as the primary concept in Kants account, ought is the primary
ethical concept; and conditional ought sentences are formally akin
to conditional probability sentences, and their categorical form has
an implicit relativization (to various senses of ought and to back-
grounds); this would render the connection between rationality and
ethics dierent from Kants: ethical arguments and thinking are prop-
erly understood on the model of induction following rules of thumb,
rather than deduction following Kantian universal moral principles;
the rationality as the foundation of ethics would thus require more
than formal consistency. Given that ethical actors are situated in cul-
tures, this Davidsonian ethics would render a system of partialities,
which shows cultural disagreements, objectively right for a culture
and also render some of ethical disagreements between distinct cul-
tures irresolvable, because there is no general algorithm for deter-
mining what is the better course of action.
In his essay A Davidsonian Approach to Normativity and the
Limits of Cross-cultural Interpretation, Yujian Zheng rst examines
Davidsons distinctive approach to normativity. The author argues
that it is Davidsons distinctive sense of normativity, rather than
something else, that unies Davidsons seemingly diverse arguments
on various issues and that the crucial feature of Davidsons approach
to normativity lies in its holistic character in integrating various inter-
dependent things and in rendering the connection between the descrip-
tive and the normative intrinsic. The author then explores the possible
implication of the Davidsonian view of normativity to the issue of
interpretative ruptures in cross-cultural understanding and intends to
argue for the following point. Although Davidsons thesis for the
impossibility of wholesale rupture (or incommensurability) between
conceptual schemes (paradigms) is sound, it does not suce to expel
skepticism or relativism at more local levels of interpretation: tem-
porary, local ruptures of interpretation are inevitable between lan-
guage games; but the local failure of Davidsons thesis can make real
sense only against the background or overall validity of his approach
to normativity.
One can say that what the preceding authors treat are actually
three distinct dimensions of the same issue and that the three themes
. +nrvr ix+nortc+iox 27
can be viewed as three related issues in Davidsons theory of mean-
ing and interpretation. And it is true that, historically speaking,
Davidson put forward his views on the three aforementioned themes
in the due course of developing his theory of meaning and inter-
pretation. It is natural to move on to this topic, and the discussion
of it is not only signicant on its own but will shed further light on
the issues addressed before. Davidsons important contribution to
natural language semantics is his proposal, rst made in Truth and
Meaning (1967), to employ a Tarski-style theory of truth to supply
what are expected for a theory of meaning. Here I would like to
give a relatively more detailed account of Davidsons approach to
meaning for the sake of the readers understanding of how this topic
is related to the preceding three themes as well as for the sake of
the readers understanding of how Davidsons theory of truth as a
theory of meaning works by appealing Tarskis Convention T, whose
understanding is needed for understanding the central points of some
of the essays in the following two parts (for example, Chung-ying
Chengs essay).
In his characteristic approach to meaning, Davidson rst identies
the conditions that an adequate theory of meaning must meet: (1)
it must enable us to give the meaning of every sentence in a nat-
ural language L under discussion: for each sentence of L, it must
produce a sentence of the form s means that p, where s is replaced
by a structural description of any sentence of L, and p is replaced
by a sentence that is supposed to give the meaning of the sentence
whose structural description is s; (2) it must show how any of sen-
tences in L is constructed out of nite words and rules regarding
the combination of words; (3) it must show that its presentation of
the meaning of a sentence in L should be based upon those con-
cepts that are not beyond the concepts that are used to express the
sentence; (4) it must be empirically veriable. With all those require-
ments in mind, Davidsons proposal starts with this: the apparently
non-extensional means (that) is replaced by is T if and only if :
(T) s is T if and only if p;
we require of a theory of meaning for a language L that without
appeal to any (further) semantical notions it place enough restric-
tions on the predicate is T to entails all sentences got from schema
T when s is replaced by a structural description of a sentence of
28 . +nrvr ix+nortc+iox
L and p by that sentence.
This requirement, as Davidson sees
it, is in essence Tarskis Convention T. Tarskis Convention T is
Tarskis criterion or test of the material adequacy for a formal truth-
denition candidate, which claims that a denition is an adequate
truth denition if it has as its logical consequences all the following
instances of the schema (T):
(T) s is True-in-L if and only if p
where s is replaced by a structurally descriptive name of any sen-
tence of the language L for which Truth is being dened, and p
is replaced by the translation of the sentence in the meta-language
ML of L. Davidson thus claims that a Tarski-type truth denition
supplies all we have asked so far of a theory of meaning. One can
say that Tarskis Convention T is one stone almost hitting four
birds, i.e., meeting all the aforementioned four conditions. The rea-
son I say almost above is this: there is an important dierence
between Davidson and Tarski. Tarskis Convention T appeals to the
notion of translation, while Davidson cannot because of the afore-
mentioned condition (3).
Davidson cannot aord to buy in any
notion of translation in Tarskis way; for he thinks a satisfactory the-
ory of meaning needs to meet conditions (3) and (4) in a strong sense
and in a straightforward way: an interpretive theory of meaning for
natural language must be empirically veriable on the basis of evi-
dence plausible available to the radical interpreter who has no
antecedent understanding of the target language. Davidsons solution
is his account of radical interpretation in terms of the principle of
charity which works by appealing to the interpreter own norms of
rationality. That is where Davidsons approach to meaning is intrin-
sically related to his principle of charity, his understanding rational-
ity and normitivity, and his strategy against the very idea of a
conceptual scheme and relativism.
Davidson (1967), Truth and Meaning, in Donald Davidson (2001), Inquiries
into Truth and Interpretation (second edition), Oxford: Clarendon Press, p. 23.
In Truth and Meaning (1967), when saying the preceding requirement is
in essence Tarskis Convention T, Davidson actually also presupposed the notion
of translation when bringing in a Tarskis strategy to serve for his purpose of con-
structing a theory of meaning for natural language. In early 70s, Davidson real-
ized the problem and suggested his account of radical interpretation to solve the
problem. Cf., Davidson (1973), Radical Interpretation, in Davidson, op. cit.
. +nrvr ix+nortc+iox 29
The three essays in Part Three, Meaning and Interpretation,
explore, directly or indirectly, various dimensions of Davidsons
approach to meaning and interpretation in view of Chinese linguis-
tic practice and Chinese philosophy in various ways. In several pub-
lications, Davidson has argued that attributions of meaning to utterances
have to be accompanied by attributions of belief to utterers. His
view is entailed by a more general consideration of two senses of
meaning, which are examined by A. P. Martinich in his essay On
Two Kinds of Meaning and Interpretation. The sense of meaning
as applied to communicationthe author calls it c-meaningis
standardly distinguished from meaning in the sense of importance
or signicancethe author calls it s-meaning. These two senses of
the term meaning correspond to two senses of the Chinese coun-
terpart term, yi-si, of the term. Martinich argues that the two kinds
of meaning, which appear to be semantically unconnected, are in
fact semantically connected in ways that the author thinks have not
been noticed in view of English and Chinese linguistic practices:
namely, a similarity that appears in considering what is required for
understanding or interpreting the two kinds of meaning. It seems to
the author, an investigation of the judgments that an interpreter
needs to make in order to understand the c-meaning of an utter-
ance leads to the conclusion that such judgments also constitute an
understanding of s-meaning; the absolute distinction between c-mean-
ing and s-meaning is untenable.
Any adequate theory of language meaning needs to account for
a wide variety of language uses, not merely various literal utterances
but also uses of gure of speech such as metaphor. Davidsons
approach to metaphor is an extension or application of his general
theory of meaning to treatment of metaphor; it is a semantic approach
which holds that metaphors function in virtue of their literal mean-
ings, instead of a special kind of meaning, to make us see things
about the world. Given that there are some distinctive characteris-
tic uses of Chinese language as an ideographic language and the
Western phonetic language like English, it would be a good test, and
is reectively interesting to see, whether Davidsons approach can
well account for metaphorical uses of Chinese language. In his essay
Metaphorical Use versus Metaphorical Essence: Examples from
Chinese Philosophy, in view of some interesting usage of metaphors
in Chinese linguistic practice, Kim-chong Chong presents his eval-
uation of the scope, limit and potentiality of Davidsons treatment
30 . +nrvr ix+nortc+iox
of metaphor. According to the author, on the one hand, Davidson
takes a narrow conception of metaphor when he contrasts it with
simile, and describes as characteristic of metaphor that it cannot be
wholly paraphrased without remainder. This conception is unneces-
sary, since it can be diagnosed as part of the motivation for think-
ing of metaphor as having an inherent cognitive content, and that
he wishes to argue against. However, on the other hand, Davidsons
conception of metaphor as belonging to the domain of use liberates
us from this narrow conception, including essentialist views of metaphor
such as that of Lako and Johnson. Through various examples, this
essay criticizes tendencies to posit an essence of metaphor in certain
discussions of metaphor in Chinese philosophy.
Though Davidsons approach to meaning is a semantic one, instead
of a pragmatic one, this by no means implies that Davidson renders
the literal meaning of a sentence simply or typically conventional.
Rather, as highlighted in A Nice Derangement of Epitaphs,
sons point is this: for an adequate interpretation to occur, it is not
simply to accept a conventional understanding of the speakers words;
one must also understand what the speaker means by the words on
that occasion; that is, interpretation is theoretically ad hoc. To this
extent, one might as well characterize Davidsons approach in terms
of his own pragmatic strand. Along the line of Davidson in this
regard, in his essay Reading the Analects with Davidson: Mood,
Force, and Communicative Practice in Early China, Yang Xiao
engages in an important debate in contemporary philosophy of lan-
guage between Dummett and Davidson by drawing out its implica-
tions in the context of early Chinese philosophy and language. The
debate concerns the relationship between the grammatical moods of
a sentence and the pragmatic forces of the utterances of the sen-
tence. Dummett insists that there is a strict correlation between mood
and force, and that illocutionary force is always conventional. Davidson
rejects the thesis, contending that neither force nor ulterior purpose
of an utterance is governed by linguistic conventions. The author
argues that, when studying the Analects, we should not limit ourselves
to the study of grammatical and conventional features of sentences.
Following Davidson, the author suggests that we should make the
pragmatic turn, focusing on peoples linguistic or communicative
The essay was published in R. Grandy and R. Warner eds. (1986), Philosophical
Grounds of Rationality, Oxford University Press.
. +nrvr ix+nortc+iox 31
practice, namely the utterances of sentences in concrete occasions
on which the sentences are put to work.
Now, with all the foregoing important themes in Davidsons phi-
losophy (his thought on the very idea of conceptual schemes and his
rejection of relativism, his principle of charity, his understanding of
rationality and normativity, and his approach to meaning, etc.), one
can see that one crucial concept emerges from behind: it is the con-
cept of truth that plays a central explanatory role in Davidsons phi-
losophy; to this extent, the preceding themes can be somehow
incorporated into what is called Davidsons theory of truth. The
central role of truth in Davidsons account is certainly not because
of Davidsons mere preference but because, it seems to Davidson,
the concept of truth as a matter of fact plays a central role as
explanatory basis to explain other important things such as our propo-
sitional thought, our understanding of the world and other minds.
It seems to Davidson, Without a grasp of the concept of truth, not
only language, but thought itself, is impossible.
Davidson treats
the concept of truth, along the line of our pre-theoretic understanding
of truth, as an axiom-like primitive that is already understood and
is partially captured by such instances of the form (T) as Snow is
white is true if and only if snow is white. Now, in view of Davidsons
philosophy and Chinese philosophy, some philosophically interesting,
signicant questions concerning the role of truth in philosophy can
be raised. Among others, rst, can Davidsons theory of truth be
validly used to examine, and enhance our understanding of, Chinese
philosophy, on the one hand, and be enriched or even constructively
developed by the recourses of Chinese philosophy, on the other hand?
Second, given that the truth concern is a long-term, fundamental or
dominant concern in Western philosophical tradition, as Davidsons
approach prominently, eectively and fruitfully illustrates, while the
dao concern has been considered a long-term, fundamental or dom-
inant concern in Chinese philosophical tradition, what is a due rela-
tion between the truth concern and the dao concern, generally speaking,
and the truth pursuit and the dao pursuit, specically speaking? Is
the dao concern dramatically dierent from the truth concern so that
Chinese philosophy and Western philosophy (say, Davidsons phi-
losophy) are alien to each other in this fundamental connection?
Surely, one can directly and eectively discuss the rst question while
Davidson (2005), Truth, Language, and History, Oxford: Clarendon Press, p. 16.
32 . +nrvr ix+nortc+iox
simply presupposing a positive answer to the second question; on
the other hand, a meta-philosophical examination of the second ques-
tion is clearly reectively worthy. The two essays in Part Five, Truth
Concern and Dao Concern, focus respectively on the two issues.
In his essay From Donald Davidsons Use of Convention T to
Meaning and Truth in Chinese Language, Chung-ying Cheng exam-
ines the foregoing rst issue from his onto-hermeneutical point of
view. In the rst part, Cheng rst characterizes Davidsons approach
to interpretation of truth in terms of the Tarskian Convention T
and then explains how the Davidsonian project can relate to under-
standing Chinese language and Chinese philosophy; in so doing, he
also raises critical questions concerning some restrictions of Davidsons
approach in view of Chinese language and how to re-construct the
Davidsonian theses in light of our understanding of Chinese lan-
guage and Chinese philosophy. In the second part, author rst sug-
gests ve principles to characterize the ontological and cosmological
visions embodied in both Chinese language and Chinese philosophy:
(1) the principle of open syntax and grammar; (2) the principle of
natural creativity; (3) the principle of internal reality externalized; (4)
the principle of historical reference (historicity) and life-world; (5) the
principle of comprehensive totality and ultimate origin. The author
then integrates the ve principles in a system which modies Davidsons
major theses in view of a philosophical reection on Chinese lan-
guage and for the sake of interpreting the dao-concerned Chinese
Addressing the aforementioned second issue in my essay Truth
Pursuit and Dao Pursuit: From Davidsons Approach to Classical
Daoist Approach in View of the Thesis of Truth as Strategic Normative
Goal, I focus on two related issues. First, I explore the due point
of the thesis of truth as (strategic) normative goal (the TNG thesis)
through analyzing Davidsons approach and drawing some morals
via three distinctions (between truth nature and truth criterion, between
truth pursuit as strategic norm and truths pursuit as tactic norm,
and between the semantic-ascent version and the paraphrase-explana-
tory-reduction version of the TNG thesis). Second, in view of the
TNG thesis and of Davidsons approach, I explore the relation
between the truth pursuit and the dao pursuit in regard to their roles
as normative goals respectively in the Western and Chinese (Daoist)
philosophies. I argue that the dao pursuit of classical Daoism is essen-
tially the truth pursuit in the way as capturing by the point of the
. +nrvr ix+nortc+iox 33
TNG thesis. I further explain how the classical Daoism as presented
in the Dao-De-Jing and the Zhuang-Zi, can make its substantial con-
tribution to our understanding of the truth concern in philosophi-
cally interesting ways.
The issue of how it is possible for Chinese philosophy and Western
philosophy in the analytic tradition, generally speaking, and Davidsons
philosophy and Chinese philosophy, specically speaking, can con-
structively engage each other is now not merely a theoretical issue
at a meta-philosophical level but also an issue of how to evaluate
the fruitful reective practice in this connection. As evidenced by the
contributors creative works in this volume, the current level of
reective inquiry into the issue has reached beyond a purely meta-
philosophical discussion of the possibility of such engagement to the
extent that many scholars have already eectively and fruitfully carried
out comparative studies of Chinese philosophy and Western philos-
ophy in the analytic tradition with the aforementioned constructive-
engagement purpose. One advantage of the situation is this. If such
reective practice with the constructive engagement as its primary
purpose does enhance and enrich the readers understanding of the
involved common themes, issues or concerns of philosophy as addressed
in the essays of this volume and does help the reader in her treat-
ment of them, that would not merely provide a prima facie positive
answer to the question of whether those perspectives from Chinese
philosophy and Western philosophy (specically here, from Davidsons
philosophy) under comparative engagement are really comparable,
i.e., whether the constructive engagement as characterized above is
possible; that would also constitute a best evidence and eective illus-
tration of, generally speaking, how such constructive engagement is
actually possible, instead of merely theoretically possible, and, specically
speaking, exactly how Chinese philosophy and Davidsons philoso-
phy in the Western analytic tradition can jointly contribute to the
common philosophical enterprise. If so, the purpose of this volume
would be fullled in one important connection. Now it is time for
the reader to tell.
I am grateful to Chad Hansen, Douglass Heenslee, Chenyang Li, Youzheng
Li and Xianglong Zhang for their helpful comments and criticism of an early ver-
sion of the main contents of Section 2. My thanks also go to A. P. Martinich for
his helpful identication of a number of awkward expressions and typos in Sections
1 and 3.
Michael Krausz
Donald Davidson oers two sorts of arguments against relativism as
standardly understood.
One sort (with which I agree) opposes the
dualism between scheme and content. The other sort (with which I
disagree) opposes the coherence of the very idea of a conceptual
scheme. I shall suggest that the notion of a conceptual scheme is
coherent and that persons of one scheme may understand persons of
another. Yet relativism need not assume the oending dualism between
scheme and content. A relativist need not, though characteristically
does, fall into the trap of assuming the oending dualism.
Standardly, relativism holds that cognitive or value claims involving
truth or rightness (or their cognates) are relative to the conceptual
schemes in which they appear. Such schemes may include cultures,
societies, civilizations, traditions, historical epochs, points of view,
perspectives, standpoints, world views, paradigms, forms of life, prac-
tices, languages, linguistic frameworks, networks of categories, modes
of discourse, systems of thought, disciplinary matrices, constellations
of absolute presuppositions, symbol systems, or the like. Accordingly,
the truth of a proposition or the rightness of a way is said to be
relative to one or another of such schemes.
* For their helpful comments and suggestions I extend thanks to Cheryl Chen,
Catherine Elgin, Jay Gareld, Bo Mou, Sam Wheeler, and David Wong.
Donald Davidson (1982), On the Very Idea of a Conceptual Scheme, in
Relativism: Cognitive and Moral, edited by Jack Meiland and Michael Krausz, Notre
Dame: Notre Dame University Press, pp. 6680.
When I use the term scheme I have in mind (as does Davidson) something
like Thomas Kuhns idea of a paradigmwithout, though, Kuhns hint of a cor-
respondence with an independent World or Nature. It is a matrix in terms of which
modes of inquiry are pursued. The constituents of schemes come as holistic pack-
ages and are not reducible to isolated beliefs or theories.
38 cn.r+rn oxr
Before considering Davidsons arguments, here are twelve prelimi-
nary points.
(1) Schemes and their cognates are adumbrant in the sense that
they have no clear boundary conditions; they are open; they are
indeterminate or indistinct as to their limits. Where, for example,
do the Inuit and Canadian cultures begin and end? Where do the
Mexican and American cultures begin and end? Where do the
Medieval, Renaissance and Baroque periods begin and end? Where
do Marxist and feminist perspectives begin and end? For one who
holds that truth or rightness is relative to these, it is dicult to spec-
ify clearly in relation to what such relativizing should operate. (I shall
return to this point.)
(2) In the above characterization of relativism, I used the word
standardly to signal that some denitions of relativism do not
invoke schemes as dening features. For example, Joseph Margolis
denes his non-standard robust relativism in terms of two essen-
tial doctrines: (i) that, in formal terms, truth values logically weaker
than bipolar value (true or false) may be admitted to govern other-
wise coherent forms of inquiry and constative acts, and (ii) that sub-
stantively, not merely for evidentiary or epistemic reasons, certain
sectors of the real world open to constative inquiry may be shown
to support only such weaker truth-values. That is all.
(3) Relativism is often motivated by the recognition of historical
or cultural diversity. Yet that recognition does not amount to rela-
tivism, since cultural or historical diversity is logically compatible
with either relativism or anti-relativism. An anti-relativist might react
to the diversity of beliefs or practices by employing the notion of
progress, according to which pertinent beliefs and practices which
dier from ones own are thought to be unenlightened, or backward,
or something of the sort. Such an anti-relativist might assume that
ones views are or should be the end product of a process toward
Joseph Margolis (1989), The Truth About Relativism, in Relativism: Interpretation
and Confrontation, edited by Michael Krausz, Notre Dame: Notre Dame University
Press, 1989, p. 232. For a discussion of Margoliss view, see Michael Krausz (1999),
Interpretation, Relativism, and Culture: Four Questions for Margolis, in Interpretation,
Relativism and the Metaphysics of Culture: Themes in the Philosophy of Joseph Margolis, edited
by Michael Krausz and Richard Shusterman, Amherst, NY: Humanity Press, pp.
nrr.+i\isv .xr i+s scnrvrs 39
a non-relative truth. Of course, other people whose beliefs and prac-
tices dier from ours might reach a similar conclusion about our
own beliefs and ways. What the anti-relativist demands is a non-rel-
ative criterion applicable to everyone. It would reveal truth and right-
ness as such.
But the relativist holds that there is no such criterion. Any would-
be neutral criterion would reect the biases or prejudices of ones
home scheme. Accordingly, to say that some belief or practice is
true or right relative to a scheme is to say more than that individ-
ual who embraces it happens to believe something, or happens to
live according to certain practices. Relativism of truth or rightness
is not just an ascription of a belief or an anthropological fact.
The relativist holds that at the boundaries of schemes where stan-
dards of evaluation give out, one has no way to adjudicate between
contending claims. One has no context-neutral objective way to
appeal to a putative overarching nature, human nature, absolute
principle or the like. Yet the relativist does claim to have resources
necessary to discriminate between distributive claims that fall within
a pertinent scheme. That is, the relativist has the resources to say,
for example, that it is true that this sentence is composed on a com-
puter, without recourse to extra-scheme considerations.
(4) Anti-relativism does not entail absolutism, the view that there
is a permanent or eternal foundation of meaningfulness, existence,
truth or rightness. One may be an anti-relativist but not an abso-
lutist. One may oppose relativism and remain agnostic or even deny
absolutism. Davidson, for example, rejects relativism, but that does
not commit him to absolutism.
(5) Relativism is not skepticism. The skeptic holds that knowledge
about matters of fact or value is impossible. Unlike skepticism, rel-
ativism arms that such knowledge is possible, yet relative to a per-
tinent scheme. The relativist sees the skeptic as setting up an impossible
goalabsolute truthand then damning the relative truth that one
might attain because absolute truth is impossible. It is as if an explorer,
on a fruitless quest for a mythical treasure, were to toss aside the
lesser treasures that he or she might acquire along the way. In
contrast to the skeptic, the relativist holds that truth or rightness is
attainable, even if it is not the absolute truth rst desired by the
(6) Sometimes relativism is confused with fallibilism. Fallibilism is
the view that, at any stage of ones inquiry, one may be wrong. It
40 cn.r+rn oxr
serves as a tonic for those who might hold that truth is absolute,
and that they in fact have attained it. But one could be a fallibilist
and still endorse an absolutist notion of truth, as Sir Karl Popper
does. He says: A . . . doctrine of absolute truth, in fact a fallibilist
doctrine . . . asserts that mistakes we make can be absolute mistakes,
in the sense that our theories can be absolutely false, that they can
fall short of the truth. Thus the notion of truth, and that of falling
short of the truth, can represent absolute standards for the falli-
That is, one could embrace the thought that truth or right-
ness are not relative to schemes, but that at any stage one could be
wrong about ones beliefs or ways. Relativism has no special claim
on fallibilism.
(7) An absolutist might worry that if there is no absolute truth to
be discovered, then there is no worthwhile goal of inquiry. If we
cannot aim for an absolute truth, for what can we aim? If there is
no absolute truth for everyone, how can we ever say beliefs or prac-
tices are true or right? The absolutist worries that by ruling out
absolute truth one rules out the possibility of progress in knowledge.
The relativist replies that there is no reason to assume a global view
of progress. Knowledge can be progressive, if only in a local way
according to standards linked to designated schemes.
(8) Relativism is characteristically dened as the view that, in the
absence of overarching standards of adjudication between pertinent
schemes, they are equally admissible. I have distinguished such a view
from multiplism whereby, in the absence of such standards, not all
admissible schemes or interpretations are equally preferable.
The multiplist
claim that several opposing interpretations may be admissible does
not preclude our giving good reasons for preferring one interpreta-
tion over others. Often there are no univocal commensurating stan-
dards between admissible interpretations. I call cases in which there
are no such standards inconclusive. Inconclusiveness does not entail
arbitrariness. Rather, it allows for critical comparability and reasoned
preferability of admissible interpretations. And it allows for reasoned
preferability of a narrative in which a given interpretation is nested.
Karl Popper (1976), The Myth of the Framework, in The Abdication of Philosophy:
Philosophy and the Public Good, First Edition, Open Court Publishing Company, p. 35.
See Michael Krausz (1993), Rightness and Reasons: Interpretation in Cultural Practices,
Ithaca and London: Cornell University Press; and Michael Krausz (2000), Limits of
Rightness, Lanham, MD: Rowman and Littleeld.
nrr.+i\isv .xr i+s scnrvrs 41
(9) Relativism is sometimes understood to entail that mutual under-
standing between those of dierent schemes is impossible. But rela-
tivists need make no such assumption. As Alasdair MacIntyre says,
we are not condemned to or imprisoned within our own particu-
lar standpoint
And Popper concurs when he says: Frameworks,
like languages, may be barriers; but a foreign framework, just like
a foreign language, is no absolute barrier. And just as breaking
through a language barrier is dicult but very much worth our
while . . . so it is with breaking through the barrier of a framework.
Popper arms that untranslatability between two languages can be
transcended when he says further:
In the comparative study of these languages we use, as a rule, our
own language as a metalanguage . . . in a critical way, as a set of rules
and usages which may be somewhat narrow since they are unable
completely to capture, or to describe, the kinds of entities which the
other languages assume to exist. But this description of the limitations
of English as an object language is carried out in English as a meta-
language. Thus we are forced, by this comparative study, to transcend
precisely those limitations which we are studying. And the interesting
point is that we succeed in this.
(10) Self-referential arguments against relativism are well known.
They concern the issue of how one should answer the question
whether the thesis of relativism itself is true or right. If we say it is
absolutely true, a contradiction results. If we say it is relatively true,
its reach is limited to the scheme in which it appears. The latter of
these alternatives is not vicious if we allow that the relativists pur-
pose in arguing need not be to convince the non-relativist. The rel-
ativist may use arguments to present his or her view, only to better
articulate the view.
(11) If Davidsons argument against conceptual schemes is sound,
it unseats those standard forms of relativism which hold that cogni-
tive or value claims are relative to schemes. And his argument asso-
ciates schemes with languages in this way:
We may accept the doctrine that associates having a language with
having a conceptual scheme. The relation may be supposed to be this:
Alasdair MacIntyre (1989), Relativism, Power, and Philosophy, in Relativism:
Interpretation and Confrontation, edited by Michael Krausz, p. 199.
Karl Popper, The Myth of the Framework, p. 46.
Karl Popper, The Myth of the Framework, p. 38, emphasis added.
42 cn.r+rn oxr
if conceptual schemes dier, so do languages. But speakers of dierent
languages may share a conceptual scheme provided there is a way of
translating one language into the other. Studying the criteria of trans-
lation is therefore a way of focusing on criteria of identity for con-
ceptual schemes.
(12) Let us consider two possible examples of pairs of schemes. I
say these are possible examples because, should Davidsons argu-
ment be sound, they would be disqualied as bona de examples.
Consider rst shame and guilt cultures. Ruth Benedict reports:
True shame cultures rely on external sanctions for good behavior, not,
as true guilt cultures do, on an internalized conviction of sin. Shame
is a reaction to other peoples criticism. A man is shamed either by
being openly ridiculed and rejected or by fantasizing to himself that
he has been made ridiculous. In either case it is a potential sanction.
But it requires an audience or at least a mans fantasy of an audi-
ence. Guilt does not. In a nation where honor means living up to
ones own picture of oneself, a man may suer from guilt though no
man knows of his misdeed and a mans feeling of guilt may actually
be relieved by confessing his sin.
Notice that in Benedicts characterization nothing precludes a per-
son of the Japanese shame culture from understanding a person of
the North American guilt culture. As a bi-cultural or bi-lingual
anthropologist acquainted with Japanese shame culture and American
guilt culture, Benedict succeeds in comparing and contrasting them.
A person of one culture can understand a person of another. Yet
there may be no non-relative standard according to which one cul-
ture is right and the other is wrong.
Here is a second possible example of a pair of schemes: Indo-
European languages and the Hopi language as understood by Benjamin
Lee Whorf. He reports:
We are constantly reading into nature ctional acting entities, simply
because our verbs must have substantives in front of them. We have
to say It ashed or A light ashed, setting up an actor, it or light,
to perform what we call an action, to ash. Yet the ashing and
the light are one and the same! . . . Hopi can and does have verbs
without subjects, a fact which may give that tongue potentialities, prob-
ably never to be developed, as a logical system of understanding some
Donald Davidson, On the Very Idea of a Conceptual Scheme, p. 67.
Ruth Benedict (1974), The Chrysanthemum and the Sword: Patterns of Japanese Culture,
New York and Scarborough: New American Library, p. 223.
nrr.+i\isv .xr i+s scnrvrs 43
aspects of the universe. Undoubtedly, modern science, strongly reecting
western Indo-European tongues, often does as we all do, sees actions
and forces where it sometimes might be better to see states . . . This
is the trouble with schemes like Basic English, in which an eviscerated
British English . . . is to be fobbed o on an unsuspecting world as the
substance of Pure reason itself.
As in Benedict, nothing in Whorf s characterization precludes an
Indo-European from understanding a Hopi, and vice versa. Whorf,
after all, is writing in English about Hopi. Nothing precludes mutual
understanding. Benedict and Whorf compare and contrast their pairs
of examples from the standpoint of a guilt culture in English. Each
makes sense of the scheme alternative to hers or his. This does not
mean, however, that the schemes are intertranslatable.
Let us now turn our attention to Davidsons argument against the
dualism between scheme and content. With Davidson, I reject the
dualism. He says: Conceptual schemes, we are told, are ways of
organizing experience; they are systems of categories that give form
to the data of sensation; they are points of view from which indi-
viduals, cultures, or periods survey the passing scene.
objects to the idea that there is data of sensation there (or some ana-
log thereof ) to be organized. Accordingly he rehearses the dualism
this way:
The idea is then that something is a language, and associated with a
conceptual scheme, whether we can translate it or not, if it stands in
a certain relation (predicting, organizing, facing or tting) to experi-
ence (nature, reality, sensory promptings). The problem is to say what
the relation is, and to be clear about the entities related.
Davidson says further:
The images and metaphors fall into two main groups: conceptual
schemes (languages) either organize something, or they t it (as in he
warps his scientic heritage to t his . . . sensory promptings. The rst
Benjamin Lee Whorf (1962), Language, Thought, and Realty, New York: Technology
Press of MIT and John Wiley & Sons, Inc., pp. 2434.
Donald Davidson, On the Very Idea of a Conceptual Scheme, p. 66.
Donald Davidson, On the Very Idea of a Conceptual Scheme, p. 74.
44 cn.r+rn oxr
group contains also systematize, divide up (the stream of experience); fur-
ther examples of the second group are predict, account for, face (the tri-
bunal of experience). As for the entities that get organized, or which
the scheme must t, I think again we may detect two main ideas:
either it is reality (the universe, the world, nature), or it is experience
(the passing show, surface irritations, sensory promptings, sense data,
the given).
And John McDowell formulates the oending dualism between scheme
and content in this way. Scheme-content dualism is incoherent,
because it combines the conviction that world views are rationally
answerable to experiencethe core thesis of empiricismwith a con-
ception of experience that makes it incapable of passing verdicts,
because it removes the deliverances of the senses from the domain
of the conceptual. According to the dualism, experience both must
and cannot serve as a tribunal.
Davidson links relativism with the oending dualism between
scheme and content. But relativism need not be dened in this way.
We need not understand schemes as organizing experience or as
giving form to the data of sensation. Rather, schemes may be
understood as matrixes in terms of which experience and data themselves
are to be understoodthat is, without assigning them a status pre-
existing to the schemes in question. Accordingly, no fact of the mat-
ter beyond language (as in the Indo-European and Hopi cases) would
be presumed, and no fact of the matter about moral behavior (as
in the shame and guilt cultural cases) would be presumed. Some rel-
ativists might link their relativism with the oending dualism. But
they need not.
As Davidson says, the dualism must be rejected. Yet if we do
reject the dualism we do not forfeit objectivity. Davidson rightly says:
In giving up dependence on the concept of an uninterpreted reality,
something outside all schemes and science, we do not relinquish the
notion of objective truthquite the contrary. . . . In giving up the dual-
ism of scheme and world, we do not give up the world, but reestab-
lish unmediated touch with the familiar objects whose antics make our
sentences and opinions true or false.
Donald Davidson, On the Very Idea of a Conceptual Scheme, p. 74.
John McDowell (1999), Scheme-Content Dualism and Empiricism, in The
Philosophy of Donald Davidson, edited by Lewis Edwin Hahn, Library of Living
Philosophers, vol. XXVII, Chicago and La Salle Illinois: Open Court, p. 96.
Donald Davidson, On the Very Idea of a Conceptual Scheme, p. 79.
nrr.+i\isv .xr i+s scnrvrs 45
Again, versions of relativism to which Davidsons critique of the
scheme-content duality are vulnerable are those which assume the
duality between what is given to the mind through sensation on the
one hand, and concepts which the mind uses to organize this given
on the other hand. Davidsons critique aims at the idea of a reality
apart from organizing concepts. He objects to the thought that there
is an external reality on to which the relativist (or the absolutist)
grafts schemes, for no coherent sense can be made of such an idea
of an external reality. The point, therefore, is not whether there is
one or more scheme to be grafted onto such a realityas the tradi-
tional stando between relativists and absolutists presents itselfbut
whether the very idea of grafting is wrong headed to start with.
Davidson says: Even those thinkers who are certain there is only
one conceptual scheme are in the sway of the scheme concept; even
monotheists have religion.
Davidson opposes those relativisms and
absolutisms which presume conceptual schemes that are taken to
mirror the world(s). He says further: It would be . . . wrong to
announce the glorious news that all mankindall speakers of lan-
guage, at leastshare a common scheme and ontology. For if we
cannot intelligibly say that schemes are dierent, neither can we
intelligibly say that they are one.
Davidsons argument against the scheme-content dualism applies
to relativism and absolutism if both presuppose that there is a nature
of things independent of schemes which answer to, or are represented
by pertinent schemes. And if the scheme-content dualism were
indeed rejected, then the question of relativism versus absolutism, as
standardly posed, would not arise in the rst place. Yet, since the
idea of a conceptual scheme need not assume the scheme-content dual-
ism, relativism without the dualism may still deploy the idea of a
conceptual scheme. Rejecting the scheme-content dualism does not
undercut all relativisms, only those relativisms that assume it. The
resulting kind of relativism would be benign, if heterodox from
Davidsons light.
Let us contrast the kind of relativism that assumes the oending
scheme-content dualism with the more benign kind of relativism that
does not. Consider again the cases of Benedict and Whorf. Recall
Donald Davidson, The Myth of the Subjective, p. 66.
Donald Davidson, On the Very Idea of a Conceptual Scheme, p. 79.
46 cn.r+rn oxr
that Benedict reports that true shame cultures rely on external sanc-
tions for good behavior, not, as true guilt cultures do, on an inter-
nalized conviction of sin. One who links relativism with the
scheme-content dualism would construe Benedicts example this way.
Good behavior denotes a fact of the matter, a phenomenon that is
there antecedently waiting to be captured or accounted for or faced.
Shame cultures do it in terms of external sanctions and guilt cul-
tures do it in terms of internalized convictions of sin. In both cases
good behavior is presumed to be pre-existently constituted before
interpretive activity. In contrast, one who does not link relativism
with the scheme-content dualism would drop the pretense of such
pre-existence and would instead do what Benedict actually does,
namely, speak of good behavior in terms given by the schemes of
shame and guilt cultures. Good behavior would be just what is so
constituted by pertinent practices in designated cultures. Thats all.
Similarly, consider Whorf s contrast between the Indo-European
who, when making sense of action phenomena, reads into nature
ctional acting entities and the Hopi who does not. One who links
relativism with the oending scheme-content dualism would hold that
the nature of action phenomena is pre-existing, constituted before
interpretive activity, that the Indo-European and the Hopi both are
seeking a match between their respective schemes and action phe-
nomena. In contrast, as in Benedicts case, the relativist who does
not link her relativism to the oending dualism, would hold that
action would be just what is so constituted by pertinent practices in
designated cultures. Thats all. My general point about relativism
without the scheme-content dualism is captured by what I elsewhere
call constructive realism or by what Patricia Hanna and Bernard
Harrison call relativist realism when, in their new and important
book, they say:
All talk of what an expression designates or refers to is in the end
merely a shorthand way of talking about the manner in which that
expression engages with, or is involved in, some practice or other . . . The
supposed entities in the case would then dissolve, not quite into thin
air, but into modes of engagement. The mode of engagement of an
expression with a practice, now, is clearly not part of the furniture of
the natural, extralinguistic world. On the contrary, it is quintessentially
a work of human invention, as much a fabrication of ingenuity in the
forging of convention as, say, the Petrarchan sonnet form or the rules
of golf . . . when we speak of the entities referred to or designated by
nrr.+i\isv .xr i+s scnrvrs 47
expressions, we speak, so far as we speak of anything at all, of fabri-
cations of the mind.
Now, an absolutist might concede that the cases that I have adduced
namely, the guilt versus shame cultures and the Hopi versus Indo-
European grammarsare conventional. But she might insist that the
scheme-content dualism still applies to such cases as middle-sized
objects or to molecules. Yet, rst, I agree with Davidson that such
an absolutist response is unsustainable because his sound arguments
for the rejection of the scheme-content dualism apply globally and
not in piecemeal fashion. Second, contra Davidson, my concern here
is to suggest that the idea of a conceptual scheme is coherent. And
that is demonstrated at least in the adduced examples. A more sus-
tained treatment of such entities as middle sized objects and mole-
cules would follow along lines suggested by Hanna and Harrison.
Now let us turn to Davidsons argument against the idea of a con-
ceptual scheme. He argues that the coherence of the idea of a con-
ceptual scheme requires the coherence of the idea of an alternative
conceptual scheme. Yet, he argues, the idea of an alternative con-
ceptual scheme is incoherent, for if an alternative conceptual scheme
is translatable into the rst conceptual scheme, it is not alternative.
And the failure of intertranslatability is a necessary condition for
dierence of conceptual schemes.
So, if a putative alternative
scheme is not thus translatable, nothing intelligent can be said about
it to distinguish it from the rst conceptual scheme. Accordingly,
Davidson says, . . . we could not be in a position to judge that oth-
ers had concepts or beliefs radically dierent from our own.
grounds for distinguishing one conceptual scheme from an alterna-
tive scheme do not obtain, the distinction between them collapses.
Patricia Hanna and Bernard Harrison (2003), Word and World: Practice and the
Foundations of Language, Cambridge: Cambridge University Press, chapter 3. See also
Bernard Harrison and Patricia Hanna (2003), Interpretation and Ontology: Two
Queries for Krausz, in Interpretation and Its Objects: Studies in the Philosophy of Michael
Krausz, edited by Andreea Ritivoi, Amsterdam: Rodopi Publishers.
Donald Davidson, On the Very Idea of a Conceptual Scheme, p. 72.
Donald Davidson, On the Very Idea of a Conceptual Scheme, p. 79.
48 cn.r+rn oxr
And with that collapse, the coherence of the very idea of a con-
ceptual scheme is unseated, as is the coherence of those forms of
relativism that presume the coherence of the idea of conceptual
schemes. Let us call this Davidsons alternativity argument.
Davidsons alternativity argument applies to partial and complete
translatability. It applies to portions of a language or a scheme that
are non-translatable as well as language taken as a whole. If por-
tions or the whole of a language are not translatable they could not
be recognized as alternative, and if they are translatable they are
not alternative.
According to Davidson, we characteristically make sense of people
by listening to what they say, and that means that they must speak
a language that we understand. If we cannot understand a persons
utterances, we could not ascribe beliefs to them. We could not say
whether they had a dierent conceptual scheme. And we could not
tell whether they are speaking a language in the rst place. Accordingly,
Davidson advances his principle of charity; namely, we can make
sense of what a person means only if we begin by assuming that
their beliefs are largely in agreement with our own. Anomalies
inevitably arise. We can deal with them in one of two ways. Either
1) we can say there is a failure of translation and therefore a dierence
in conceptual scheme, or 2) we can say that there is a dierence in
belief. But there is no hard and fast rule that forces us to deal with
anomalies according to 1) or 2). Therefore, if we come across a puta-
tive partial failure of translation, we can always say that it arises
from a dierence of belief rather than a dierence of conceptual
But we must be careful here. From the fact that there is no hard
and fast rule in this regard, it does not follow that we can always
say that the partial failure of translation is a matter of belief rather
than scheme. Further, there is a diculty in the vocabulary of par-
tial and complete translation, since it is hard to designate what
is complete and what is partial in relation to what is complete. The
limits of language are always emerging and open. It is partial all the
way through. What, after all, could count as a complete natural lan-
guage like Chinese or English?
Thanks are due to Cheryl Chen for her here paraphrased characterization of
Davidsons view on this matter, in private correspondence, August 8, 2003.
nrr.+i\isv .xr i+s scnrvrs 49
One might retort to Davidsons alternativity argument that all that
it shows is that, while we may have no criterion to verify the presence
of an alternative scheme, it may still exist. But Davidson might well
reply that it is not enough to indicate the possibility of an alterna-
tive scheme. We must also give some reasons or grounds to believe
that such an alternativity exists. Yet such reasons or grounds can be
given. To start with, there is enough overlap between those of guilt
and shame cultures for its inhabitants to compare and contrast their
cultures with one another. Within that conversational spacewhich
is improvisatory and emergentthey may show one another their
disparities concerning the moral character of what they do or the
grammatical character of what they say.
One might be tempted to say that one can distinguish conceptual
schemes by taking a neutral stance by divesting oneself of all schemes
and then comparing those that present themselves. But Davidson
disqualies this possibility, because conceptual schemes are embed-
ded in languages. And to divest oneself of all schemes would require
giving up the use of language. Language is necessary for thought.
So, if one gives up the use of language, one could not even com-
pare. Davidson says: Speaking a language is not a trait a man can
lose while retaining the power of thought. So there is no chance
that someone can take up a vantage point for comparing concep-
tual schemes by temporarily shedding his own.
According to Davidsons alternativity argumentwhich associates
translatability with understandabilitythe two pairs of examples
adduced earlier in this paper, because of their would-be inter-
translatability, cannot be bona de. Each of the pair is not an alter-
native to the other. In contrast, I suggest that understanding between
people of dierent schemes is possible despite the fact that there may
be no complete translatability between them. Davidson might reply
that the examples I oer are not bona de as alternative schemes,
since understanding is possible and therefore translation is possible.
Such a reply would disallow that understood schemes can be non-
translatable. It would disallow the testimony of bilinguals that they
understand dierent languages despite the fact that full translations
are not forthcoming. Put otherwise, Davidson holds that nontrans-
latability is a sucient condition for non-understandability. But I
Donald Davidson, On the Very Idea of a Conceptual Scheme, p. 68.
50 cn.r+rn oxr
suggest that a failure of translation does not imply a failure of under-
standability. Mutual understandability does not entail translatability,
although translatability might entail understandability.
Alasdair MacIntyre observes that translation is not always possi-
ble. Yet we can make sense of alternative cultures, that is, we can
understand persons of alternative cultures. Translatability is not nec-
essary for understandability. So the non-translatability of schemes
does not bar one from saying that one is alternative to the other.
Accordingly, even if shame cultures and guilt cultures were not trans-
latable, a person of a guilt culture like Benedict could still under-
stand persons of a shame culture, and, presumable, vice versa.
I return to the question of partial translatability. As David Wong
most helpfully suggests, one can gloss terms of a language even if
we have no strict equivalent terms in ones home language. He says:
When translators of the Chinese word ren in the Analects render it as
goodness, or benevolence, or humanity, or authoritative person-
hood we know these are just approximations. We look at what peo-
ple say about the role of the concept as one for comprehensive virtue,
and we may bring to our understanding of comprehensive virtue, per-
haps, our understanding of Greek ethics, but we know that this is also
just an approximation. We are told that Confucius at one point asso-
ciates ren with loving persons (Analects 12:22), but we should also be
careful about equating whatever notions of love we have with the one
Confucius was using in this context . . . Goodness or humanity or
benevolence or authoritative personhood serves more like a pointer.
The pointer Wong speaks of should not be taken to point to a
determinate equivalent term or phrase. The term in an object lan-
guage (Chinese) need not have a ready-made counterpart term in
the metalanguage (English). The gloss should not be taken to point
to a determinate equivalent term. The idea that there must be a
determinate equivalent term ignores a characteristic context in which
questions of translation arise. When one translates from Chinese to
English a word in a legal document pertaining to the transfer of
David Wong, in correspondence, July 30, 2003. For the glosses of ren as 1)
goodness, 2) benevolence, 3) humanity, or 4) authoritative personhood, see respec-
tively: 1) Arthur Waley (1989), The Analects of Confucius, New York: Vintage Books;
2) D.C. Lau (1979), Confucius: the Analects, London: Penguin; 3) Wing-tsit Chan (1969),
A Sourcebook in Chinese Philosophy Princeton: Princeton University Press; and 4) Roger
Ames and Henry Rosemont, Jr. (1999), The Analects of Confucius: A Philosophical
Translation, New York: Ballantine Books.
nrr.+i\isv .xr i+s scnrvrs 51
property, for example, one assumes acquaintance with various insti-
tutions of property rights. The translator might well seek terms neu-
tral to the contending claims so as to avoid a verbal sleight. But
such avoidance requires foreknowledge of the implicated institutions.
And questions of determinate equivalent terms arise yet again in the
translation between those institutions. The point is that praxial
considerations systematically enter into the business of translation.
One cannot translate independently of the practices and purposes to
which translation is put. The multiplicity of pertinent practices and
purposes undermines eorts that require lexical equivalencies.
Accordingly, a given term (like Dao) may be glossed in many ways
according to a multiplicity of schemes, but it may be lexically equiv-
alent to none.
Further, an interpreter may have more than one home scheme,
without full translatability between them. Indeed, as MacIntyre
One of the marks of a genuinely adequate knowledge of two quite
dierent languages by one and the same person is that persons abil-
ity to discriminate between those parts of each language which are
translatable into the other and those which are not. Some degree of
partial untranslatability marks the relationship of every language to
every other.
Languages not fully translatable between one another need be no
barrier to ones understanding them. MacIntyre emphasizes the point:
Notice that this recognition of untranslatability never entails an acknowledgment of
some necessary limit to understanding. Conversely, that we can understand
completely what is being said in some language other than our own
never entails that we can translate what we understand. And it is this
ability both to understand and to recognize the partial untranslatabil-
ity of what is understood . . . [that creates] the predicament of the bilin-
gual speaker.
By distinguishing between translatability and understandability
MacIntyre claims that one can understand two languages or appro-
priate portions thereof while not being able to translate between
them. Indeed, bilingual speakers need to be able to do this in order
Alasdair MacIntyre, Relativism, Power and Philosophy, pp. 1889.
Alasdair MacIntyre, Relativism, Power and Philosophy, pp. 189, emphasis
52 cn.r+rn oxr
to determineas he or she doeswhat is not translatable from one
language to another. Just as untranslatability does not entail a limit
on understanding, understanding does not entail translatability.
MacIntyres point undercuts Davidsons association of untranslata-
bility with non-understandability. So even if one were to hold xed
non-translatability as a criterion of conceptual schemes, it would not
follow that one would not be in a position to recognize an alterna-
tive conceptual scheme as a bona de case. Without translatability
we need not remain silent about an alternative conceptual scheme.
The putative alternative need not be absorbed into the home con-
ceptual scheme. One can still distinguish conceptual schemes. Hilary
Putnam agrees when he says: If one recognizes that the radical
interpreter himself may have more than one home conceptual
scheme, and that translation practice may be governed by more
than one set of constraints, then one sees that conceptual relativity
does not disappear when we inquire into the meanings of the var-
ious conceptual alternatives.
In short, mutual understandability
does not entail translatability, and translatability is not necessary for
Here now are my own misgivings about those relativisms that are
tied to conceptual schemes and their cognates. As I said in my pre-
liminary remarks, the problem with such relativisms concerns not
their incoherence but their adumbrance. The schemes relative to
which truth or rightness is supposed to operate often cannot be
clearly delineated. Where, after all, does a guilt culture and a shame
culture begin and end? Where, after all, does an Indo-European lan-
guage and a Hopi language begin and end? For that matter, where
do Chinese and English begin and end? The diculty concerns the
fact that, in saying that truth or rightness is relative to such schemes,
there are no ready application procedures. If I say that truth is rel-
ative to my scheme, which is mine? What if I have several schemes
Hilary Putnam (1989), Truth and Convention: On Davidsons Refutation of
Conceptual Relativism, in Relativism: Interpretation and Confrontation, edited by Michael
Krausz, pp. 1801.
nrr.+i\isv .xr i+s scnrvrs 53
that blend into one another? What if I am a Chinese American New
Yorker feminist? My misgiving concerns the applicability of the rela-
tivist rubric upon pertinent schemes. That does not mean that such
schemes are incoherent. It only means that there is a problem in
applying the notion of a scheme to do the job that relativists stan-
dardly demand of it. Schemes are characteristically adumbrant. And
they may overlap in complicated and unstable ways. The adum-
brance of schemes does not disqualify understanding between inhab-
itants of each. But typically it is unclear just what truth or rightness
is meant to be relative to. Of course, as degrees of adumbrance may
vary, the severity of the problem of applicability may vary.
In sum, I join Davidson in disallowing the scheme-content dual-
ism, and thereby those relativisms which assume it. However, not
all relativisms need assume the oending dualism. Further, I allow
the coherence of the idea of conceptual schemes and their mutual
understandability, thus dislodging the link between non-translatabil-
ity and non-understandability. Finally, in face of the characteristic
adumbrance of schemes, those relativists who do not assume the
oending scheme-content dualism still need to provide procedures
under which their relativism can apply. Or perhaps we should turn
to non-standard relativisms that dene themselves along lines quite
distinct from conceptual schemes altogether.
Koji Tanaka
1. Comparative Philosophy
Comparative philosophy is a branch of philosophy which examines
and contrasts dierent traditions of philosophy. For example, a com-
parative philosopher may examine African, Buddhist, Chinese, Indian,
Muslim, and Western traditions of philosophy in comparison with
one another. Comparisons may also be made between sub-traditions
within a tradition: one may compare Confucianism and Daoism
within the tradition of Chinese philosophy, for example. The par-
ticular concern for us in this paper is the comparative study of
Chinese and Western traditions of philosophy.
One may question the legitimacy of examining Chinese philoso-
phy and Western philosophy in general terms because it suggests a
unied way of talking about these philosophical traditions. Now, it
is certainly true that there are disagreements within a tradition. These
disagreements are sometimes the very force behind the development
of a philosophy. However, the fact that philosophers within a tra-
dition are all engaged in a meaningful philosophical debate suggests
that they share something in common. Otherwise, it would not be
clear how they are able to communicate with each other and con-
sequently contribute to the development of their tradition.
Now, in comparative philosophy, as Larson points out, general-
ization, or what he calls holistic characterizations, of a tradition
* This paper was presented at the international conference Philosophical
Engagement: Davidsons Philosophy and Chinese Philosophy held in Beijing, China,
in June 2004. Many thanks go to the members of the audience whose comments
forced me to clarify the main point of the paper. Also I would like to thank Barry
Taylor for clarifying the point of Davidsons argument. I have also greatly beneted
from the comments of two anonymous referees as well as those of Bo Mou. Many
thanks also go to Bronwyn Finnigan for proof-reading the paper.
56 cn.r+rn +vo
may be misleading.
However, without any generalization, it is not
clear how we can even start an examination of a foreign tradition.
As Kasulis argues, [w]ithout the generalization, however, people
cannot proceed in their quest for understanding anymore than they
can use most databases without rst dening the elds of entry.
In fact, the opposite seems to be more harmful. If there is no gen-
eralization that can guide an investigation, all we can do is super-
impose our framework of thought onto others. As any comparative
philosopher would agree, that is not the way that the activities of
comparative philosophy should be carried out.
For these reasons, some generalization seems necessary in com-
parative philosophy. However, for a generalization to be possible, it
must be assumed that there is something common that is shared by
the members of the tradition. If this is true then it seems reason-
able to suggest that comparative philosophy is a way of understanding
the philosophies of dierent traditions: a generalization of the com-
monalities which underlie a philosophical tradition is thought of as
a way of understanding the tradition itself. In fact, Parkes suggests
exactly this:
The major concern of comparative philosophy, however, is an under-
standing of the philosophies themselvesand thereby, to a greater or
lesser extent, of the world.
Yet how can we discover the common features that can be attrib-
uted to a tradition? And how can the comparisons of traditions be
made possible based on those commonalities? If comparative phi-
losophy involves generalization, such questions seem to be funda-
mental to comparative philosophy. Unless we can answer these
questions, it is not clear how one can comparatively philosophize
about dierent traditions.
Larson, Gerald James (1988), Introduction: The Age-Old Distinction Between
the Same and the Other, Interpreting Across Boundaries, G.J. Larson and E. Deutsch
(eds.), Princeton: Princeton University Press.
Kasulis, Thomas P. (2002), Intimacy or Integrity, Honolulu: University of Hawaii,
pp. 78.
Parkes, Graham (1987), Introduction, Heidegger and Asian Thought, Graham
Parkes (ed.), Honolulu: University of Hawaii Press, p. 3.
r.\irsox .xr cnixrsr coxcrr+t.r scnrvr 57
2. Davidson on Conceptual Scheme
Whether one considers it with pleasure or sorrow, the answers to
the above questions are hard to come by. The diculty can be enun-
ciated by referring to Davidsons discussion of conceptual relativism.
Davidson argues that the notion of conceptual relativism is incoherent.
The notion that he rejects is based on what he calls conceptual
Conceptual schemes, we are told, are ways of organizing experience;
they are systems of categories that give form to the data of sensation;
they are points of view from which individuals, cultures, or periods
survey the passing scene.
Conceptual relativism is then the doctrine that what is considered
to be real is relative to a conceptual scheme:
Reality itself is relative to a scheme: what counts as real in one sys-
tem may not be in another.
An example of this claim is made by Whorf who famously argued
that the Hopi conceptualize reality dierently from Westerners pri-
marily due to a dierent conception of time.
In arguing against conceptual relativism, Davidson associates a
conceptual scheme with a language.
We may accept the doctrine that associates having a language with
having a conceptual scheme. The relation may be supposed to be this:
where conceptual schemes dier, so do languages.
However, a speaker of one language does not necessarily organize
their experience dierently from a speaker of another language. This
is because a language may be translated into another by invoking
only one conceptual scheme. And this is a conceptual scheme that
they both share. For example, blanc is a French word for white.
One need not think that blanc and white express two dierent
concepts; rather, they both express the one concept, i.e., white or
whiteness. That is, the words blanc and white can be translated
Davidson, Donald (1984), On the Very Idea of a Conceptual Scheme, Inquiries
into Truth & Interpretation, Oxford: Clarendon Press, pp. 18398.
Davidson, 1984, p. 183.
Davidson, 1984, p. 183.
Davidson, 1984, p. 184.
58 cn.r+rn +vo
into each other. Hence, as Davidson writes, Studying the criteria
of translation is therefore a way of focusing on criteria of identity
for conceptual schemes.
Conceptual relativism thus means the failure of the intertrans-
latability of languages. This is because conceptual relativism denies
that languages share a conceptual scheme, which is the criterion for
translation. Showing the existence of languages which are not trans-
latable is tantamount to making a case for conceptual relativism. We
may then ask the question of whether there are any languages which
are untranslatable.
In addressing this question, Davidson considers two kinds of fail-
ures of translatability: complete and partial. A case of complete fail-
ure is when no signicant portion of a language can be translated
into another language. Davidson examines two metaphors that may
be considered to show the complete failure of languages. The rst
is that languages organize experience or the data of sensation; another
is that languages t experience or the data of sensation. According
to the metaphor of organization, two languages are (completely)
untranslatable if they organize the same experience in such a way
that no signicant portion of the languages can be intertranslated.
Here the two languages must be said to organize one experience.
However, Davidson argues, [t]he notion of organization applies only
to pluralities.
We cannot speak of organizing one thing, such as a
closet, without referring to the plurality of things in it. The prob-
lem with the case under consideration is that it involves only one
thing: one and the same experience. Hence, there is no coherent
sense in which two untranslatable languages can be said to organize
this single experience. Thus the metaphor of organization does not
make the idea of untranslatable languages coherent.
The metaphor of tting faces much the same diculty. To say
that a language ts experience or the data of sensation is to accept
the sentences of the language as (largely) true. Hence, the criterion
of untranslatable languages is that they are largely true but not
The problem is that the notion of truth is not inde-
pendent of translation. Two dierent languages may be (largely) true
with respect to one experience. Yet the recognition of this assumes
Davidson, 1984, p. 184.
Davidson, 1984, p. 192.
Davidson, 1984, p. 194.
r.\irsox .xr cnixrsr coxcrr+t.r scnrvr 59
translatability of the languages.
Without this assumption, we could
not demonstrate that these two languages truly express the same
experience. Hence the metaphor of tting does not establish untrans-
latable languages. Therefore, neither of the two metaphors estab-
lishes a case for (completely) untranslatable languages.
Now the case of partial failure may be thought to be more plau-
sible. It assumes that one language shares something in common
with other languages. The suggestion is that some common part
acts as a neutral ground from which the failure of intertranslatabil-
ity can be identied. Plausible candidates for such common parts
are beliefs and other mental attitudes that produce certain speech
acts. However, it is not clear how we can understand a persons
beliefs without understanding their language, which itself is assumed
to be untranslatable. Furthermore, we cannot understand a persons
language if we do not know their beliefs and attitudes which serve
as the neutral ground. Thus, it is not clear upon what basis we can
understand the failure of intertranslatability of a persons language.
We may respond to the case of partial failure by arguing that two
people can come to have a general agreement on beliefs. They may
simply stipulate what is to be believed for the sake of communica-
tion. However, such agreement could only be reached in the same
language. Such a situation would not invoke a language which had
partially (or completely) failed to be intertranslated. If the dierences
can be explained and described at all, that would be because of the
common language that was used to do so. Thus, again, an untrans-
latable language cannot be identied.
we could not be in a position to judge that others had concepts or
beliefs radically dierent from our own.
if we cannot intelligibly say that schemes are dierent, neither can we
intelligibly say that they are one.
Davidson explains this point by referring to Tarskis T-scheme or what he calls
Tarskis Convention T (Davidson, 1984, p. 194). I have omitted such details for
the benet of those unfamiliar with this T-scheme.
Davidson, 1984, p. 197.
Davidson, 1984, p. 198. We can frame the problem that Davidson points out
in terms of Hansens reading of Zhuang Zi. (Hansen, Chad (1992), A Daoist Theory
60 cn.r+rn +vo
3. . . . and Comparative Philosophy
Davidsons thesis seems to have a consequence for comparative phi-
losophy, in particular for a comparative study between Chinese and
Western traditions of philosophy.
It is often thought that these tra-
ditions dier conceptually. A comparative study between them is
considered to be based on comparisons between two dierent con-
ceptualizations. Consider, for instance, the notion of personal iden-
tity. As is pointed out by many, for example Rosemont, (contemporary)
Western philosophers largely view persons as atomic and autonomous
On the other hand, the Chinese (at least early Confucians),
as argued by Rosemont, think of persons as roles, particularly fam-
ily and social roles such as child, parent, teacher, neighbor and so
on. Hence, for the Chinese, to be a person is conceptualized to be
part of a network of relationships that are dened by roles.
Despite the fact that the Chinese concept of personal identity
seems radically dierent from the Western concept, Western philoso-
phers seem to be able to understand the Chinese concept of per-
sonal identity and vice versa as above. This may indicate that the
above two concepts are, in fact, not radically dierent. However, for
Western philosophers to learn this new Chinese concept of personal
of Chinese Thought, Oxford: Oxford University Press.) If Hansen is correct in his inter-
pretation, Zhuang Zi was concerned with the basis from which we accept and reject
social conventions. Zhuang Zi argues that there is no perspective which we can
accept to settle the issue of which convention to accept and which to reject because
the basis upon which we accept and reject the convention must also be in ques-
tion. All we can do is accept or reject a discrimination pattern based on a certain
perspective, and this perspective is one we happened to have accepted for what-
ever reason. Thus, there is no way to get to the common basis from which we can
settle the issue. However, given that this paper is written in English, addressing
issues to English-speaking philosophers, whether Zhuang Zis perspective can be
presented and understood at all is precisely what is at issue.
Davidsons primary target is in fact the very idea of a conceptual scheme. His
rejection of the idea is often thought to be based on a rejection of the content/scheme
distinction. Granted that he does reject the content/scheme distinction, Davidsons
thesis is more than the rejection which pregured in Quine. (W.V.O. Quine (1951)
Two Dogmas of Empiricism, Philosophical Review, Vol. 60, pp. 2043.) Davidson
rejects the very idea of a conceptual scheme by rejecting conceptual relativism. My
primary focus in this paper is Davidsons unique insight into the incoherent nature
of conceptual relativism, as it is conceptual relativism that has some bearing on
comparative philosophy.
Rosemont, Henry Jr. (1991) A Chinese MirrorMoral Reections on Political Economy
and Society, La Salle: Open Court.
r.\irsox .xr cnixrsr coxcrr+t.r scnrvr 61
identity, they must share something in common with the Chinese.
As Davidson explains in another place:
We tend to discover our common problems and interests as we read,
teach, and travel. The discovery surprises us for, to begin with, minds
are best compared by nding as many points of similarity as every-
day patterns of action and reaction aord. But once this tting of pat-
tern to pattern is accomplished, the remaining dierences loom out of
proportion. This perhaps explains why a rst exposure to a new tra-
dition seems to reveal an unbridgeable gap. What experience shows,
though, is that, as in other areas, dierences are to be understood only
as seen against a background of underlying agreement. The underly-
ing agreement may be largely unspoken and unnoticed, but it is always
available. Sometimes we need help in appreciating how philosophy
builds on what we all know. No world views or conceptual schemes
are truly incommensurable.
If Davidson is correct, what strikes us at rst glance as dierent may
in fact share a background of underlying agreement. This seems to
be the reason why Western philosophers can come to understand
the Chinese concept of personal identity and vice versa.
Perhaps it is the case that Western philosophers charitably inter-
pret the Chinese concept of personal identity based on what they
know about personal identity. Moreover, the fact that Chinese and
Western philosophers can communicate with each other may show
that they share a background of underlying agreement: the possi-
bility of communication between them is constituted by a background
of underlying agreement (to give it a Kantian ring). However, if the
background of underlying agreement is thought of in terms of the
sameness of the ways in which Chinese and Western philosophers
conceptualize personal identity, as is often done in a comparative
Davidson, Donald (2001) Foreword, Two Roads to Wisdom?Chinese and Analytic
Philosophical Tradition, B. Mou (ed.), Chicago and La Salle: Open Court, p. v.
It is dicult to talk about Davidsons philosophy without talking about lan-
guage. In this paper, I take for granted that (classical) Chinese diers from English
to the extent that the issue of translation arises. But comparative philosophers often
take the issue of translation as a rather trivial matter given the supercial dierence
between the two languages. They often shift the focus from languages to concepts
(or conceptualizations) by placing a primary importance on concepts that can be
materialized in languages. (This seems to be the way that comparative philosophers
think of the relationship between languages and concepts.) For this reason, I am
primarily concerned with concepts in this paper, while acknowledging that a prop-
erly Davidsonian analysis of comparative philosophy should take place in the con-
sideration of intertranslations of languages.
62 cn.r+rn +vo
study of Chinese and Western traditions of philosophy,
then it is
not clear that they could be in a position to judge that they are in
fact conceptualizing the same thing. If this is correct, then Western
philosophers (or Chinese philosophers) could not be in a position to
judge that the Chinese concept (or Western concept) is a dierent con-
cept of personal identity.
Let us narrow our focus on the problem. As we saw above, we
seem to be able to compare Chinese and Western conceptions of
personal identity, whether the dierence is radical or not. However,
for a comparative study of the conceptions of personal identity to
make sense, both conceptions must be recognized as conceptions of
personal identity. Yet, what will guarantee that they are both concep-
tions of personal identity? What makes the above Chinese concep-
tion a conception of personal identity? Given that it appeals to family
and social roles, why isnt that considered to be a conception of fam-
ily and social identity? How can Western philosophers know that
the Chinese philosophers have built their conception of personal
identity on what Western philosophers all know about personal iden-
tity? Western philosophers cannot refer to their own experience of
familial and other social interactions to answer this question since
they do not necessarily conceive of such interactions as matters of
personal identity. They cannot even refer to experiences of inter-
acting with other individuals, since the Chinese may not view these
interactions as interactions between persons. Furthermore, to say that
the Chinese conception truly expresses the experience which is viewed
as a matter of personal identity by Western philosophers is to assume
that the Chinese and Western conceptions are both conceptions of
personal identity and that, therefore, they can be translated into each
other as two dierent conceptions of personal identity. However, this
begs the question. Thus Western philosophers cannot make a case
for the idea that the Chinese conception expressed above organizes
or ts the experience they conceive of as being a matter of personal
identity. Analogous considerations show that Chinese philosophers
cannot demonstrate that the Western conception organizes or ts
the experience which they view in terms of personal identity. Therefore,
in the same way that Davidson argues against the idea of two ways
I do not claim that this is the only way in which comparative studies of Chinese
and Western traditions of philosophy are carried out. However, this seems to be
the assumption that is made in much literature on comparative philosophy.
r.\irsox .xr cnixrsr coxcrr+t.r scnrvr 63
of conceptualizing the same experience, we can cast doubt on the
possibility of comparing two dierent ways of conceptualizing per-
sonal identity. At the very least, neither Chinese nor Western philoso-
phers seem to be in a position to judge that they have two dierent
conceptions of personal identity.
This seems to suggest that a comparative study between Chinese
and Western traditions of philosophy cannot be based on the exam-
ination of dierent ways of conceptualization. How, then, is com-
parative philosophy possible? In the remainder of this paper I will
consider two approaches to comparative philosophy that seem to
have been adopted by many comparative philosophers.
I will exam-
ine them to see whether they provide any plausible answers to the
above question. In particular, I will examine whether or not those
two approaches fall pray to Davidsons argument.
4. Special Logic Resort
One popular suggestion for explaining the foundations of compar-
isons is (or was once upon a time) that each philosophical tradition
has a special logic of its own. According to this suggestion, as
Hansen puts it, What makes sense to Chinese does not make sense
to Westerners.
For example, in the case of personal identity, what
makes sense to the Chinese is that to be a person is to be in a net-
work of relationships dened by family and social roles. However,
this notion of personal identity does not make sense to Western
philosophers. For them, a person is an atomic and autonomous indi-
vidual who exists in some way prior to the society. In this way, it
is argued that comparisons of Chinese and Western philosophies can
I do not claim that there have been only two approaches to comparative phi-
losophy. Whether or not there are other approaches, I leave for another occasion.
Hansen, Chad (1983) Language and Logic in Ancient China, Ann Arbor: University
of Michigan Press, p. 20. There has been a vast amount of literature presenting
non-Western philosophies in this way, for example D.T. Suzuki in his presentations
of Zen Buddhism. See also Hansen, 1983, ch.1 in the context of Chinese philoso-
phy. One thing which is never clear in such literature is their use of logic. Logic,
for a contemporary Western logician, is about validity of arguments. It is about
what proposition or sentence, i.e., conclusion, follows from what propositions or sen-
tences, i.e., premises; it is not about what propositions or sentences make sense.
The approach considered in this section does not seem to conform to this standard
account of logic.
64 cn.r+rn +vo
be made based on the dierences in logics which dictate what
makes sense.
However, the special logic resort, as Hansen identies it, seems
to make comparisons untenable. If the Chinese have a logic which
is not shared by Westerners and if a certain logic dictates what makes
sense, Western philosophers, with their own logic, would be unable
to make sense of the philosophical discourses of the Chinese. Yet, if
the Chinese indeed had a special logic which prevented Western
philosophers from understanding their discourses, the Westerners
would be unable to recognize Chinese philosophy as meaningful at
all, let alone as being a philosophy. In such case, it is not clear how
Western philosophers can even have the idea that the Chinese have
a special logic in the rst place. Similarly, Chinese philosophers can-
not discover anything about the Western special logic either.
If we understand logic according to the special logic resort, it
seems implausible to suggest that a special Chinese logic underlies
Chinese philosophy and that it is the special logic that makes Chinese
philosophy dier from Western philosophy and vice versa. Thus, the
special logic resort falls prey to Davidsons discussion of the com-
plete failure of translatability. Since it cannot be identied as any-
thing meaningful, Chinese philosophy, let alone the Chinese
conception of personal identity, cannot be compared to Western
philosophy. Indeed, according to the special logic resort, there is
See Hansen (1983, p. 14) who also argues in the same way: If Western minds
are incommensurably dierent from Chinese minds then we could not discover any-
thing at all about Chinese thoughtincluding the actual details of the logic of
Chinese. Note also that the recent psychological studies such as Nisbett, Peng,
Choi and Norenzayan seem to face the same diculty. (Nisbett, Richard, Kaiping
Peng, Incheol Choi and Ara Norenzayan (2001) Culture and Systems of Thought:
Holistic Versus Analytic Cognition, Psychological Review, Vol. 108, pp. 291310.)
Nisbett et al. argue that cognition of East Asian people is dierent from Western
people. If this is true, it is not clear how anyone can cognize that the cognitive
processes of two peoples are dierent in the rst place. This diculty is also applic-
able to the philosophical discussion on ethno-epistemology by Nichols, Stich and
Weinberg, and Weinberg, Nichols and Stich. (Nichols, S., S. Stich and J.M. Weinberg
(2003) Metaskepticism: Meditations in Ethno-Epistemology, The Skeptics: Contemporary
Essays, Steven Luper (ed.), Aldershot: Ashgate, pp. 227247. Weinberg, J.M., S.
Nichols and S. Stich (2001) Normativity and Epistemic Intuitions, Philosophical
Topics, Vol. 29, pp. 429460.) They argue for the ethnicity of intuitions that under-
lie philosophical arguments by claiming that intuitions are culturally bounded. If
this is true, it is not clear how we can intuit that our intuitions are culturally
r.\irsox .xr cnixrsr coxcrr+t.r scnrvr 65
nothing that Chinese philosophy can be compared to. The same
must be said about Western philosophy.
5. Comparisons of Presuppositions
In recent years, a more sophisticated approach to comparative phi-
losophy has developed. For example, such an approach can be found
in the works of David Hall and Roger Ames. Hall and Ames,
primarily focusing on comparative philosophy rather than Chinese
philosophy per se, analyze Chinese philosophy by illuminating its
underlying presuppositions in contrast to those which underlie Western
philosophy. For Hall and Ames, presuppositions are usually unan-
nounced premises held by the members of an intellectual culture or
tradition that make communication possible by constituting a ground
from which philosophic discourse proceeds.
Based on these pre-
suppositions, it is claimed, the Chinese tradition of philosophy has
built up a stock of concepts, which form their beliefs, knowledge,
arguments and so on.
Hall and Ames
approach Chinese and Western presuppositions
with an attention to the historical development of the philosophical
system of each tradition. They provide a historical study describing
how the rst problematic, or analogical/correlative, thinking came
to dominate Chinese philosophizing and how the second problem-
atic, or causal, thinking came to dominate Western philosophizing.
Thus, English speakers associate the expression Heaven with a spir-
itually transcendent realm that is associated with God who is respon-
sible for the causal order of the universe. Chinese speakers, on the
other hand, associate tian (often translated as heaven) with an ances-
tral structure which analogically adumbrates their family structure.
Hall and Ames explain this dierence in terms of dierent ways of
thinking: causal and analogical. They seem to think of these two
ways of thinking as grounding the ways in which concepts are related
to each other. Because of the dominance of causal thinking in Western
Hall, David L. and Roger T. Ames (1995), Anticipating China, Albany: State
University of New York Press.
Hall, David L. and Roger T. Ames (1987), Thinking Through Confucius, Albany:
State University of New York Press, p. 11.
Hall and Ames, 1995.
66 cn.r+rn +vo
philosophy, Westerners associate Heaven with a spiritually transcend-
ent realm and God. The Chinese, on the other hand, associate tian
with an ancestral and family structure because of their analogical
thinking. Hence, Hall and Ames illuminate dierent sets of associ-
ated concepts and the nature of their association. It is in this way
that they contrast Chinese and Western traditions of philosophy.
However, analyzing the foundation of comparative philosophy in
this way seems to face the problem that Davidson points out. As
many comparative philosophers have noticed, the stock of concepts
within each tradition seems to be the product of the specic histor-
ical and cultural circumstances of the tradition. Each set of presup-
positions is a product of the historical and cultural development of
each tradition, as Hall and Ames themselves demonstrate.
succinctly describes the situation thus:
[Philosophy] necessarily draws on a cultural legacy for its terminol-
ogy, conceptualization of problems, and even relevance. Philosophy
develops not in total isolation but within a community of discourse.
However, that a tradition is indeed based upon certain historical and
cultural circumstances seems to create a diculty for accessing con-
cepts that are foreign to the tradition. Hence, it is not clear whether
or not we can judge that foreign concepts are the same or dierent
from our own concepts, as the methodology for and the concepts
used in our judgments will be historically and culturally sensitive.
Hence, we still face the problem that we are trying to solve.
6. . . . and the Possibility of Comparative Philosophy
Despite the above problem, Hall and Ames approach may be thought
of as providing the foundation of comparative philosophy. This may
Whether or not they would agree with my description of their study, we seem
to be able to present their study in this way without doing any injustice to them.
Whether or not this thought has been inherited from Heidegger and Gadamers
account of hermeneutics, I let the reader to be the judge.
Kasulis, 2002, p. 14. See also Rosemont who argues that our basic cognitive
frameworkwhich ranges from our unreective conception of what it is to be a
human being to our assumptions, beliefs, and presuppositions about the general fea-
tures of the physical universeis overwhelmingly determined for us by a set of
highly specic environmental circumstances ranging from social relations accompa-
nying stages of history and of culture, to the syntactical particularities of our native
r.\irsox .xr cnixrsr coxcrr+t.r scnrvr 67
be argued for in the following way. Consider, for example, the early
Confucians who are thought of having a concept cluster of ethics
that does not involve the rights-based concept cluster of morals
developed in modern western culture.
Based on their view of per-
sons as autonomous individuals, Western philosophers seem to base
their ethical concepts on a notion of rights that bears upon atomic
and free individuals. They consider murder, for instance, to be eth-
ically wrong because such act violates individual rights. The early
Confucians, on the other hand, base their ethical concepts on the
notion of piety, in particular lial piety, that adumbrates a social
Thus, the early Confucians would also think of murder
as ethically wrong, but not for the reason that it is a violation of
individual rights but because of an impiety that suggests a social
disharmony or discordance in the intricate network of relationships.
Hence, by explaining the bases from which the two traditions have
built their dierent ways of conceptualization, i.e., presuppositions,
we seem to be able to carry out a comparative study of ethics.
Indeed, to claim that the early Confucian ethics is not based on the
modern Western rights-based concept cluster seems to be to com-
pare the early Confucian and Western conceptions of ethics, so the
argument goes.
Hall and Ames
(and Kasulis)
seem to explain the above activ-
ity of comparative philosophy as follows. Consider one of the gestalt
pictures, such as the duck-rabbit picture. When the picture is pre-
sented for the rst time, it appears to some as a picture of a duck
and others as a picture of a rabbit, despite the fact that they are all
sighting the same thing. Depending on the visual orientation a per-
son has at the time, the picture may appear and be seen dierently.
Only with the same, or at least similar, visual orientation, is the pic-
ture seen as the same thing. In order to be able to explain what
a person sees, then, we need to forge an elucidation of the per-
sons orientation in such a way that others can adopt it and thereby
tongue. To be is to be the value of a pronoun formand an indexical at that.
(Rosemont, Henry Jr. (1988), Against Relativism, Interpreting Across Boundaries, G.J.
Larson and E. Deutsch (eds.), Princeton: Princeton University Press, p. 37.)
Rosemont, 1988, p. 63.
One may disagree with this description of early Confucians. However, the
point of the paper remains the same even if someone presents a dierent picture
of early Confucians.
Hall and Ames, 1995.
Kasulis, 2002.
68 cn.r+rn +vo
come to an understanding of what the other experiences. Similarly,
the early Confucians notion of social harmony, adumbrated by the
notion of piety, and the contemporary Western notion of rights may
be said to point to the same thing despite appearing to the Chinese
as piety and appearing as rights to Western philosophers. In order
to see it dierently, one would need an elucidation of the others
presuppositions, for example, their presuppositions about personal
identity. In this way, the early Confucian and modern Western
philosophers would be able to compare their conceptions of ethics.
What the above discussion suggests, so the argument goes, is that
the issue of ethics cannot be conceptualized in an absolute manner.
In the same way that what is seen in the duck-rabbit picture depends
on ones orientation, the demarcation of what counts as ethics seems
to be made only in virtue of the conceptualization performed via a
tradition of thought in which the philosophical thought has been cul-
tivated. One may then be led to think that dierent traditions of
philosophy can be compared by illuminating the dierent ways in
which concepts form a coherent whole. For Westerners, the concept
of ethics is associated with the concepts of rights and autonomous
individuals, whereas for the Chinese it is associated with those of
social harmony, piety, and family and social roles that dene rela-
tionships. However, by associating the expression ethics with the
concept of social harmony and so on, Western philosophers can
come to understand and compare the Chinese concept of ethics with
their own Western concept. Similarly, Chinese philosophers can asso-
ciate the expression ethics with the concepts of autonomous indi-
viduals and rights. In other words, by exploiting the exibility of the
boundary demarcating what counts as ethics, the Chinese and Western
conceptions of ethics can be identied and compared.
This seems to mean that the boundaries of comparisons are set
by the philosopher who is conducting a comparative study.
Potter seems to suggest exactly this when he writes: dierence in conceptual
system is assumed by the investigator, and dierences between persons, cultures, or
philosophers is thus made to t dierences in meanings of segments, pieces of behav-
ior, and specic categories. In truth, the choice of boundaries across which to com-
pare is the same choice as that among categories of investigation; as we saw, it is
the practical concerns of the investigator which ultimately dictate these boundaries,
they are not absolutely there to be discovered in the material investigated. (Potter,
Karl H. (1988) Metaphor as Key to Understanding the Thought of Other Speech
Communities, Interpreting Across Boundaries, G.J. Larson and E. Deutsch (eds.),
Princeton: Princeton University Press, p. 24.)
r.\irsox .xr cnixrsr coxcrr+t.r scnrvr 69
of the exibility and non-absoluteness of boundaries, ethics cannot
be conceptualized in an absolute manner. Hence, Western philoso-
phers may think that the expression ethics, for instance, can be
associated with piety and so on, even though it is customarily asso-
ciated with a radically dierent concept cluster. Though the choice
of the boundaries of comparisons may be thought to suggest that a
dierent tradition of philosophy is to be understood in an investi-
gators individual way, it is the exibility and non-absoluteness of
the boundaries that give rise to the possibility of understanding other
traditions and, thus, that of comparative philosophy. In this way,
Hall and Ames approach may be thought to provide the founda-
tion for comparative philosophy.
6. Despite all this . . .
Despite the above prima facie reasonable consideration, comparative
philosophers such as Hall and Ames have not completely overcome
the diculty that we enunciated earlier in the paper. Of course,
Western philosophers can examine the Chinese concept of piety and
roles. They can also compare it to their concept of rights and
autonomous individuals. They can even associate the expression
ethics with their concept of social harmony and piety by calling it
ethics, as we saw above. Similarly, Chinese philosophers can learn
the Western concept of rights and autonomous individuals and com-
pare it to their concept of piety and roles. They can even illumi-
nate practical consequences of adopting the modern Western notion
of rights-based ethics.
However, what could not be done is that Western philosophers
demonstrate that the Chinese concept of piety is an ethical concept
comparative with their own conception of ethics. Neither could Chinese
philosophers demonstrate the Western conception of rights to be an
ethical concept. In other words, the problem we enunciated above
with respect to personal identity has not been resolved. There does
not seem to be any guarantee that Chinese and Western philoso-
phers are talking about, or pointing to, the same thing. For Western
philosophers, according to Hall and Ames,
the discourse of ethics
At least, if my elucidation of their approach is accurate.
70 cn.r+rn +vo
involves the notion of free and autonomous individuals. However,
our problem is that we cannot show that the notion of free and
autonomous individuals is another way of conceptualizing what the
Chinese would view in terms of family and social roles. Similarly, it
is not clear that Western philosophers would conceive of the notion
of piety, based on family and social roles, as a matter of ethics, since
the Western concept of ethics is largely based on the notion of free
and autonomous individuals. How can Western philosophers come
to see the Chinese concept of piety as a concept of ethics while the
Western concept of ethics forms a conceptual cluster that does not
involve the notion of piety? What makes the association of ethics
with the concept of piety legitimate? Analogously, how can Chinese
philosophers come to understand the Western concept of rights as
an ethical concept given that their conception of ethics is based on
social harmony and so on?
As is shown by Davidson, an appeal to experience itself does not
answer these questions. To see this, consider, for instance, the situ-
ation of seeing someone who is murdered. This experience may be
organized or tted in two dierent ways; one in terms of the con-
cept of impiety and another in terms of rights violation. For Western
philosophers, for whom ethics is concerned with rights that bear
upon autonomous individuals, this experience would be viewed in
terms of ethics. For the Chinese, this experience would be a matter
of impiety and social disharmony. However, for Western philoso-
phers to infer from this that the Chinese would thus view the expe-
rience in terms of ethics is to assume that the notions of impiety
and social disharmony are ethical notions. Yet this is the very assump-
tion that we are trying to establish. As we saw above with respect
to personal identity, the provability of this assumption remains to be
seen. In fact, if Davidson is right, demonstration of the legitimacy
of associating ethics with what the Chinese would view in terms of
impiety and social disharmony seems to escape Western philosophers
judgment. Analogous considerations show that Chinese philosophers
could not judge that rights violation is a matter of ethics either.
One may argue that the reason why piety and rights are both
concerned with ethics is because of their associated terms found in
the texts. It seems true that contemporary Western philosophers use
the concept rights in the same discourse that they use ethics, and
Chinese philosophers do the same with piety and social harmony
(or perhaps ren (humanity)). However, to conclude from this that
r.\irsox .xr cnixrsr coxcrr+t.r scnrvr 71
Chinese philosophers are also concerned with ethics is to assume
that the notion of social harmony (or ren) can be translated as ethics
or that it has ethical contents, and hence begs the question.
Therefore, by rephrasing Davidsons conclusion, Western philoso-
phers could not be in a position to judge that the Chinese had con-
cepts of, or beliefs about, ethics radically dierent from their own,
nor could they intelligibly say that they are the same. Chinese philoso-
phers, with their own conception of ethics, could not be in an anal-
ogous position. The same can be said about personal identity and
other concepts.
This does not mean that comparative philosophy is a meaning-
less activity nor that comparative philosophy is impossible. Western
philosophers can compare their notion of autonomous individuals
with the Chinese notion of roles. Based on this comparison, they
can compare their rights-based notion of ethics with the Chinese
notions of piety or social harmony. If someone wishes to use the
term comparative philosophy in this sense, I will concede the term.
There may also be alternative ways of conducting the activities of
comparative philosophy than I have addressed in this paper.
Unfortunately, however, there are comparative philosophers who
seem to have thought that they were sincerely engaged in compar-
ative studies of ethics, metaphysics, epistemology and so on based
on the notion of dierent logics, dierent concepts or mutually non-
inclusive stocks of concepts given by dierent presuppositions through
which we organize our experience. This paper casts doubt on the
legitimacy of such activities and claims.
This discussion seems to be applicable to philosophy itself. In the same way
that there is no criterion to determine how ethics is conceptualized in two dierent
traditions, there seems no criterion to determine how philosophy is conceptualized
in two dierent traditions. What consequences this has for comparative philosophy
and my discussion in this paper, I leave for another occasion.
I wish I had had an opportunity to pose the question of how comparative
philosophy is possible to Davidson himself.
Larson complains that comparative philosophers have favored philosophical
boundaries of European thought since Descartes for identifying conceptual prob-
lems in general . . . Agenda items for comparative philosophizing have been selected
from ethics, epistemology, philosophy of language, and so forth (Larson, 1988, pp.
1112). Larsons complaint may be well justied. However, a more serious prob-
lem seems to be the fact that comparative philosophers have thought that they
could select conceptual items from ethics, epistemology, philosophy of language, and
so forth, and conduct comparative studies of those items based on the notion of
dierent ways of conceptualizing them.
Stephen C. Angle
The best comparative philosophy does more than compare. According
to three recent accounts, it seeks to integrate or challenge or
seek truth.
Each of these ideas depends on being able to com-
pare ideas, texts, and reasons across philosophical traditions, but each
goes beyond comparison to urge that we engage in creative philos-
ophy. The desire to synthesize aspects of dierent philosophical tra-
ditions that developed in relative isolation from one another (Yu
and Bunnin, 2001, p. 296), however, can seem hopeless. The many
dierences among the cultures and languages in which these dierent
philosophical traditions are imbedded have led to a set of worries
about the coherence of comparative philosophy: doctrines like rela-
tivism or incommensurability claim that reasoned comparison or
translation across cultural gulfs can be impossible. If comparative
philosophy would stick to comparisons, then perhaps these worries
could be side-stepped: any two things can be compared in one respect
or another. To integrate or challenge or seek truth, though, we must
tackle the worries head-on.
Over the last three decades, one of the principal resources on
which comparative philosophers could draw when they sought assur-
ance has been the work of Donald Davidson. In his Foreword to
Two Roads to Wisdom? Chinese and Analytic Philosophical Traditions, Davidson
A rst exposure to a new tradition seems to reveal an unbridgeable
gap. What experience shows, though, is that, as in other areas, dierences
are to be understood only as seen against a background of underly-
ing agreement. The underlying agreement may be largely unspoken
and unnoticed, but it is always available. Sometimes we need help in
See (Allinson, 2001), (Van Norden, 2001), and (Yu and Bunnin, 2001), respectively.
74 cn.r+rn +nnrr
appreciating how philosophy builds on what we all know. No world
views or conceptual schemes are truly incommensurable. (Davidson,
2001, p. v)
Davidson has argued convincingly that anything we can recognize
as a language must, in principle, be translatable into any other lan-
guage. As I will elaborate below, this argument has sometimes been
misunderstood, but it nonetheless stands as an important support for
the enterprise of comparative philosophy.
As signicant as Davidsons work has been, however, my thesis in
this essay is that comparative philosophers need still more than
Davidsons theory is able to provide. It is not enough to know that
translation is possible in principle: we need to be able to talk about
conceptual dierences with more subtlety, and to reason about what
is at stake in overcoming them. Keeping in mind that Davidsons
theory has at its heart theorems that translate sentences of an object
language into metalanguage, note the blas attitude he expresses
in this quote: Even when the metalanguage is dierent from the
object language, the theory exerts no pressure for improvement,
clarication, or analysis of individual words, except when, by acci-
dent of vocabulary, straightforward translation fails [Davidson 1984,
33]. The problem is that in an important sense, our vocabularies
are often no accidents.
Davidson is not wholly without resources to explain the range of
conceptual dierences and the dynamics of language change. I will
point to some promising ideas in his later work, as well as to some
elaborations of his basic theories proposed by others. But in the end,
the best solution to these matters is to be found in the work of
Robert Brandom. In many respects Brandoms views are Davidsonian;
turning to Brandom is not to abandon Davidsons core insights, but
to recast them in a framework that allows them fuller expression.
My goal will be to show why the creative philosophical projects of
contemporary comparative philosophy are possible, even though they
are often dicult.
My rst step is to quickly sketch Davidsons argument and then
to distinguish it from two similar but mistaken alternatives. The
essence of Davidsons argument is that a languages sentences can
always be translated into sentences with the same propositional con-
tent in any other suciently rich language, and furthermore that
there are no limits, at least in principle, on how languages can be
v.kixo noov ron covr.n.+i\r rnirosornv 75
enriched or revised. Let us start with the rst half: sentences can
always be translated. Davidson contends that anything we can iden-
tify as a sentence is amenable to translation, because the interpretation
of linguistic behavior cannot get o the ground unless we nd a way
to identify both what our interlocutor believes, and what she means
by her words. Since there are an innite number of false things one
might say in response to given situation, but a limited number of
(salient) true ones, Davidson proposes that only a principle of char-
ity can ground interpretation. That is, we start by provisionally inter-
preting our interlocutor as speaking truly. Interpretation proceeds on
this basis: while no particular sentence uttered by our interlocutor
must be true, we can only identify falsehoods against the background
of broad agreement.
This reasoning leads Davidson to conclude that anything we can
identify as language can always be translated because there is no
other means to identify something as a sentence other than by inter-
preting it. If all attempts at interpreting the noises or movements of
some creatures fail, then we (so far) have no grounds for attribut-
ing language to them. And this means that we can have no grounds
for attributing radical conceptual gulfs. To be sure, we can tem-
porarily fail to understand someone, but incommensurabilityunder-
stood as the doctrine that the concepts of two languages dier so
radically that they cannot be translatedis impossible.
Let me now briey turn to the second half of Davidsons argu-
ment, as I explained it above, namely that there are no limits, at
least in principle, on how languages can be enriched or revised. I
think this is clear, albeit almost always implicit, in Davidsons writ-
ing. He comes closest to making this point explicit when he writes
about the incredibly exible ways in which we can understand oth-
ers, and make ourselves understood, despite grammatical errors and
malapropisms. Davidson says that an interpreter has, at any moment
of a speech transaction, what I persist in calling a theory. (I call it
a theory . . . only because a description of the interpreters compe-
tence requires a recursive account.) However, the theory in ques-
tion is not unchanging: The theory we actually use to interpret an
utterance is geared to the occasion (Davidson, 1986, p. 441). We
adjust our interpretive theory on the spur-of-the-moment, taking into
account new things the speakers says. This leads Davidson to dis-
tinguish between prior theory and passing theory:
76 cn.r+rn +nnrr
For the hearer, the prior theory expresses how he is prepared in ad-
vance to interpret an utterance of the speaker, while the passing theory
is how he does interpret the speaker. For the speaker, the prior theory
is what he believes the interpreters prior theory to be, while his pass-
ing theory is the theory he intends the interpreter to use. (Ibid., p. 442)
What must be shared for communication to succeed, he concludes,
is our passing theories. Davidson insists, though, that passing theo-
ries do not correspond to an interpreters linguistic competence,
nor can they be said to be learned, or to be governed by conven-
tions. Of course things previously learned were essential to arriving
at the passing theory, but what was learned could not have been
the passing theory (Ibid., p. 443).
There are two sides to Davidsons conclusion. On the one hand,
we see that Davidson emphasizes a certain kind of linguistic change,
namely our everpresent practical ability to adjust our (linguistic) the-
ories to new (linguistic) evidence. On the other hand, in so doing
he rejects a very common way of understanding our linguistic com-
petence, namely the idea of a clearly dened, convention-governed,
shared structure which language users acquire and then apply to
cases. That is, as he puts it, he abandons the ordinary notion of a
language, [and thus erases] the boundary between knowing a lan-
guage and knowing our way around in the world generally (Ibid.,
pp. 4456). Both aspects of his argument are relevant to my con-
cerns here. As we proceed, I will argue that while much of what
Davidson says here is correct, he goes too far in abandoning a cen-
tral role for the norm-governed social practice that is language.
Thinking about language in terms of conventions is problematic, but
neglecting any role for norms makes linguistic change more easy,
and conceptual dierence more trivial, than they really are.
By way of further eshing out how Davidsons argument works
let me now turn to two alternative formulations. The rst, from
There are two dierent levels at which norms might be relevant to language:
at a general level, applicable to all language use; and at a specic level, applicable
only to the use of normative words like ought. Sam Wheeler has developed some
Davidsonian ideas to argue that the logic of conditional possibility ts how we use
ought better than the logic of obligation. See (Davidson, 1980) and (Wheeler,
1974). I am not convinced that these two options exhaust the alternatives, but will
not enter that debate here. Rather, I conne myself to the rst, more general level.
I will argue below that contra Wheeler, Brandoms talk of the ways in which we
commit ourselves via language use does make senseand this because the com-
mitment cannot be understood in probabilistic fashion, as he would have it.
v.kixo noov ron covr.n.+i\r rnirosornv 77
Martin Hollis and Steven Lukes, can appear to be the same as
Davidsons but is actually importantly dierent in ways that matter
a great deal to comparative philosophy. The second is Alasdair
MacIntyres version of, and response to, Davidsons argument. I will
show that while MacIntyre fails to refute Davidson, MacIntyre does
help to highlight a weakness that I will exploit later in this essay.
In a series of inuential articles published from the 1960s through
the 1980s, Hollis and Lukes argued for what they called a bridge-
In the very identication of beliefs and a fortiori of belief systems we
must presuppose comonly shared standards of truth and of inference,
and . . . we must further presuppose a commonly shared core of beliefs
whose content or meaning is xed by application of these standards.
(Lukes, 1982, p. 262)
Lukes adds that practical everyday beliefs are prime candidates
for the bridgehead. Truths of this kind cannot in general be assumed
not to be shared, since that would be strictly unintelligible (Ibid.,
p. 265). Lukes explicitly invokes Davidson to support his position,
though he does realize that in arguing for a specic bridgehead,
he is going beyond Davidsons more general principle of charity. His
reason is that Davidsons argument is inadequate as a method of
translation. . . . The principle of charity gives us no guidance as to
where agreements are to be assumed before disagreements can show
up (Ibid., p. 263).
To a signicant degree, Lukes is right. Davidson does not pro-
vide us with a method of translation, but only an understanding of
the nature of linguistic behavior which guarantees that translation of
another language will always be possible. In suggesting that Davidsons
argument has provided assurance to comparative philosophers, I have
not gone so far as to say that Davidson provides us with all the
tools we need. Lukes argues that Richard Grandys Principle of
Humanity is a better guide to translation than Davidsons Principle
of Charity.
Lukes also suggests that guidance from sociology and
anthropology will be needed to settle questions about what counts
(Lukes, 1982, p. 264f ). See (Grandy, 1973, p. 443), wherein Grandy explains
that the Principle of Humanity tells us that in interpretation, the imputed pattern
of relations among beliefs, desires, and the world [should] be as similar to our own
as possible.
78 cn.r+rn +nnrr
as good reasons for action or belief in a given context.
It is likely
that he is right on both of these counts, but what is important for
our purposes is to see that he is wrong in seeing Davidson and
Grandy (or Davidson and the anthropologists) as competitors.
Davidsons argument aims to establish possibility, and is based on
very general features of languages; Grandy and the rest are best
understood as aiming to help us actually translate in a particular
context. Even when engaged in this practical activity, though, we
must keep in mind Davidsons dictum that interpretation is always
provisional. I can grant many of Lukes points, therefore, but must
stop short of endorsing the idea of an inexible bridgehead. Hollis
supposes that an interpreter might pin down the native counter-
parts of English sentences like Yes, this is a brown cow in order
to establish a bridgehead, and insists that the bridgehead is not an
hypothesis that can be refuted or conrmed by evidence (Hollis,
1970, p. 215). Davidson has shown us, though, that there is no way
other than via interpretation to pin down the beliefs or meanings
of the natives, so even seemingly secure sentences about brown cows
could turn out to be misinterpretations, if we uncovered enough evi-
dence. Hollis seems to think that impossible, but counterexamples
are easy to manufacture.
Distinguishing Davidson from Lukes and Hollis has helped us to
see what Davidson can (and cannot) do for comparative philosophy.
If we turn to MacIntyres eort to rebut Davidson, we will uncover
a dierent kind of limitation to Davidsons approach. MacIntyre
believes that Davidsons argument has two premises: (1) all we have
to do to assure understanding of another cultures point of view is
to translate their language, and (2) nothing that we can identify as
a language could resist translation. MacIntyre contends that these
premises rest upon:
. . . a way of translating texts from alien and dierent cultures, and of
responding to them, which is central to the cosmopolitan cultures of
those modern internationalized languages-in-use, such as contemporary
Trans-Atlantic and Trans-Pacic English, one of whose central fea-
(Lukes, 1982, p. 274n39). Lukes notes that he and Hollis part company at this
point; Hollis insists that standards of rationality must be determined a priori, and
also is more explicit about the inexibility of the bridgehead. See below, and
(Hollis, 1982).
For example, see the discussion below about fair-weather vs. foul-weather animals.
v.kixo noov ron covr.n.+i\r rnirosornv 79
tures is that utterance in them presupposes only the most minimal of
shared beliefs. These are languages, so far as is possible, for anyone
at all to use, for those who are equally at home everywhere and there-
fore nowhere. (MacIntyre, 1991, p. 114)
MacIntyres idea, which he develops more fully elsewhere, is that
certain modern languages that are spoken around the world have
been, in a sense, neutered: in order to be usable by people from
widely-dierent cultural backgrounds, these languages have lost some
important characteristics that all local languages once had.
The chief
features that he says internationalized languages have lost are: rst,
naming systems that presuppose certain beliefs on the part of the
languages speakers, and second, a tight relationship between canon-
ical texts expressing strong, substantive criteria of truth and ratio-
nality and acceptable utterances (MacIntyre, 1988, pp. 377, 384).
What we are left with are languages in which the relationship of
a name to what is named will have to be speciable . . . indepen-
dently of any particular scheme of identication embodying the beliefs
of some particular community (MacIntyre, 1989, p. 193), and in
which formerly canonical texts now serve only as sources for liter-
ary allusions, not as standards of truth (Ibid., p. 194).
How does MacIntyre think that the putative transition from local
to internationalized languages might refute Davidsons argument?
Since the new languages have no tight connections to particular sets
of beliefs, MacIntyre reasons, the sorts of obstacles that would stand
in the way of translating from Aristotles Greek to Confuciuss Chinese,
for example, are gone.
The plausibility of Davidsons contention
that we can translate anything that we can identify as a language
into our own language, that is, rests on taking our own language to
be one of these neutered modern languages. But once Aristotle and
Confucius have been translated into English, they have lost their
essential ties to particular beliefs that helped to dene their stand-
points. Once rendered into English, MacIntyre concludes, they are
no longer genuine Confucianism nor genuine Aristotelianism, but
merely new menu items for the modern individualism of aestheti-
cized personal choice (MacIntyre, 1991, p. 115). Nothing has been
done, in other words, to dissolve the incommensurability between
the actual languages of Confucianism and Aristotelianism.
See (MacIntyre, 1988, ch. 19) and (MacIntyre, 1989, especially section 4).
See (MacIntyre, 1991).
80 cn.r+rn +nnrr
I am dissatised with MacIntyres response to Davidson for two
reasons. First, MacIntyre has overstated the extent to which mod-
ern languages have been neutered. It seems true that the use of
names may have changed somewhat along the lines he describes,
but the blanket claim that we no longer have canonical texts is surely
false, especially in ethics. Many speakers of internationalized English,
by no means limited to so-called fundamentalists, still look to the
Bible, for instance, as expressing strong, substantive criteria of right
and wrong. Many others look to more recent classics, like the
Constitution. Admittedly, appeal to the Bible or the Constitution
does not always settle disputes. For one thing, contemporary America
is composed of many overlapping communities with commitments
to dierent sets of canonical texts. It is also true that canonical texts
must be interpreted, and that they therefore supply standards of right
and wrong only together with the styles of reasoning and traditions
of interpretation that have grown up around them. But in this we
are no dierent from our predecessors of any age: texts are never
Second, MacIntyres whole discussion of internationalized languages
seems to me to misunderstand the nature of Davidsons argument.
The essence of MacIntyres response to Davidson, after all, is that
although it appears that Confucian terms can be translated into
English, this is in fact only a watered-down kind of quasi-translation.
This response would only be harmful to Davidson, though, if Davidsons
argument had used our apparent ability to translate Confucian Chinese
into English as evidence for his conclusion. As we have seen, though,
Davidsons argument does not depend on any specic instances of
successful translation, but instead relies on very general features of
languages which apply equally to modern and to pre-modern lan-
guages. Davidson could easily accept MacIntyres claim that current
translations from Chinese into English are mere quasi-translations,
For those familiar with Saul Kripkes argument that the meanings of proper
names cannot contain descriptive content, MacIntyre makes the following response:
What this argument shows is, not that the names of persons do not or cannot
have informational content, but that either they lack such content or it is true of
them that their use presupposes commitment to a belief, such that were this belief
discovered to be false, the name would not continue to be used in the same way
(MacIntyre, 1988, p. 377). For Kripke, see (Kripke, 1980, Lecture One, pp. 612
and passim.).
On styles of reasoning, see (Hacking, 1982); that texts do not interpret them-
selves is of course one of Wittgensteins most important lessons.
v.kixo noov ron covr.n.+i\r rnirosornv 81
in fact, because Davidson recognizes that it may be necessary to
enrich or revise our language before true translation will be possible.
Still, even if MacIntyre has not given us a rebuttal to Davidsons
argument, he has helped to point out a danger. I just said that
Davidsons argument does not depend on any specic instances of suc-
cessful translation. This is true, but Davidson can sometimes sound
like he thinks instances of ready translation across alleged concep-
tual gulfs abound. About Thomas Kuhn, one of the main propo-
nents of the idea that dierent scientic languages (or paradigms)
can be incommensurable, Davidson writes: Kuhn is brilliant at say-
ing what things were like before the revolution usingwhat else?
our post-revolutionary idiom (Davidson, 1984, p. 184). But Kuhn
has emphasized that his work is no simple translation into existing
vocabulary, but rather the learning and then teaching of a new lan-
guage (Kuhn, 1983). In a similar way, MacIntyre helps us to see
that theres a crucial dierence between a breezy rendering of
Confucian texts into trans-Atlantic English and a genuine transla-
tion. Even without endorsing either Kuhns or MacIntyres precise
diagnosis of the diculties that conceptual gulfs can cause, I think
we can see that Davidson may have been too blas. I now turn to
eorts to build onto Davidsons framework a sensitivity to these issues.
In 1989, Bjrn Ramberg published a splendid book on Davidsons
philosophy of language in which he tried to show that Davidsonians
can satisfactorily account for the phenomena that MacIntyre, Kuhn,
and others call incommensurability. Indeed, Ramberg argues that
Davidsons framework is better-suited than other models of seman-
tics to taking seriously talk of meaning change and conceptual gulfs,
because it does not make essential use of the idea of reference. He
realizes, though, that his goal might appear quixotic:
Assimilating the concept of incommensurability to a Davidsonian seman-
tics . . . would appear to be highly problematic. For the incommensu-
rability thesis is not only a denial of the view that the continuity of
reference is a necessary presumption of successful communication. It
is also, perhaps primarily, intended to positively identify a certain kind
of semantic obstruction between would-be communicatorsa seman-
tic obstruction which . . . is commonly taken by critics of the idea to
be intranslatability. And on the radical-interpretation model of seman-
tics, it is hard to conceive of any such obstruction. It is so hard, in
fact, that the radical-interpretation model is generally presumed to yield
a powerful argument against the very idea of incommensurability.
(Ramberg, 1989, p. 119)
82 cn.r+rn +nnrr
The powerful argument to which Ramberg refers is precisely the
Davidsonian argument I have been discussing from the beginning,
and endorsing as an important foundation of comparative philosophy.
The key to Rambergs eort to make Davidsonian semantics t
comfortably with the idea of incommensurability is a distinction he
draws between abstract linguistic meaning, which is modeled by the
synchronic truth-theories that radical interpretation is said to pro-
duce, and the actual production of meaning, which takes place
via language and is (according to Ramberg) constituted by lin-
guistic conventions.
Davidsons rejection of convention notwith-
standing, Ramberg says that our conventions bind us in various
wayswith varying viscosity, he saysand:
We rely on conventions to understand and make ourselves understood.
Incommensurability, as a communication breakdown, can be under-
stood as a breakdown of linguistic conventions, caused by changes in
use that are too abrupt to be absorbed smoothly, or changes that a
particular set of conventions are too rigid to accommodate. Semantically,
then, incommensurability is a disruption in the ongoing interpretation-
through-application of our linguistic conventions. (Ibid., p. 130)
With adequate time and changes to our conventions, Ramberg is
saying, anything is translatable: that lies at the center of Davidsons
argument. But viscous conventions can cling to us, keeping us from
seeing how dierent our conventions would have to be to correctly
interpret some foreign language, such that we repeatedly misunder-
stand the foreigners.
Ramberg concludes that incommensurability
is a diachronic relation, not a synchronic one; it is not a relation
between structures, but a symptom of structural change (Ibid., p. 131).
I agree entirely with Rambergs core contention, namely that the
phenomenon that has been labelled incommensurability is (1) no
threat to translation or comparison in principle, because (2) it arises
out of the diachronic relations between (communities of ) language
users, but still (3) has important practical signicance. Given the
always/in principle possibility of proper interpretation, I believe it
(Ramberg, 1989, p. 130), and see also (Ibid., ch. 8).
Ramberg writes that If Azanda magicians and Western physicists were to dis-
cuss causality, incommensurability would arise. But is would arise because transla-
tion would too often be wrong, which is to say that the interpocutors would frequently
believe they were using the same language when actually they were not (Ramberg,
1989, p. 131).
v.kixo noov ron covr.n.+i\r rnirosornv 83
makes more sense to say that languages are incommensurate than
incommensurable, but this is a minor point. Rambergs Davidsonian
version of the doctrine of incommensurability promises to simulta-
neously ground the possibility of comparative philosophy while cau-
tioning us about that enterprises diculties, and explaining those
diculties in terms of the viscosity of our linguistic conventions. It
is this last part, the explanation in terms of conventions, that I nd
problematic. My next goal in this essay is to argue not just that
Ramberg cannot successfully appeal to conventions to serve his needs,
but also that this is for reasons which come back to haunt Davidsons
picture itself. In the end, therefore, I will not oer here a revised
Davidsonian account, but an account which pushes beyond Davidson
in some crucial respects.
I have already introduced something of Davidsons opposition to
basing our understanding of linguistic competence on a language,
in the sense of a clearly dened, convention-governed, shared struc-
ture which language users acquire and then apply to cases. He repeat-
edly claims that while every-day interpretation is greatly facilitated
by the convergence of speech behavior that we usually mark by say-
ing that people speak the same language, we do well to ignore this
practical issue in constructing theories of meaning, of truth, and of lin-
guistic communication (Davidson, 1990, p. 311). He believes that:
. . . what interpreter and speaker share, to the extent that communi-
cation succeeds, is not learned and so is not a language governed by
rules or conventions known to speaker and interpreter in advance; but
what the speaker and interpreter know in advance is not (necessarily)
shared, and so is not a language governed by shared rules or con-
ventions. What is shared is, as before, the passing theory; what is given
in advance is the prior theory, or anything on which it may in turn
be based. (Davidson, 1986, p. 445)
Davidson adds that of course things previously learned were essen-
tial to arriving at the passing theory, but what was learned could
not have been the passing theory (Davidson, 1986, p. 443). Ramberg
helps us to see how radical Davidsons view is when he says that
Davidson is arguing, in eect, that even if we used nothing but mala-
propisms, communication would still be possible (Ramberg, 1989,
In other writings, I have tried to follow a strategy more like Rambergs, albeit
without reliance on conventions, but now nd the current strategy more compelling.
See (Angle, 1994) and (Angle, 2002).
84 cn.r+rn +nnrr
p. 101). Linguistic meaning is modeled by radical interpretation,
which does not (Davidson and Ramberg believe) essentially depend
on the notion of a language.
While Ramberg follows Davidson in all these matters, he does
believe that Davidsons rejection of language and conventions is too
complete, which keeps Davidson from seeing the proper signicance
of incommensurability. Ramberg writes that:
In dismissing the body of conventions that constitute a language as a
contingent feature of linguistic communication, as nothing more than
a complex, and immensely valuable, practical aid, Davidson obscures the
dialectical relation between meaning and what we might call the production of
meaning. Meaning, understanding the meaning of an utterance, is what
is modeled in radical interpretation. The production of meaning, on
the other hand, is modeled by what we call a language. (Ibid., p. 110,
emphasis added)
Radical interpretation involves the construction of a truth-theory for
a speaker, but Ramberg says that in a normal speech situation,
we do not advert to truth theories. Here, conventional strategies,
not the construction of truth-theories, determine what truth-conditions
we attach to utterances. In other words, in so far as we are speak-
ers of a language, the truth-conditions of the sentences of that lan-
guage are conventionally taken for granted (Ibid., p. 111). In at least
most cases of the production of meaningthat is, actual speaking
and interpretingRamberg thinks we are constrained by the con-
ventions of our language. Since these conventions change slowly,
while radical interpretations adjusts instantly to new evidence, a gap
can appear between the radical interpretation of meaning and the
conventional production of meaning. What Ramberg means by the
dialectical relation between meaning and meaning production,
therefore, is presumably that radical interpretation only sporadically
impacts our production (and consumption) of meaning, such that a
diusion of meaning [or] a blurring of linguistic understanding
(Ibid., p. 112) becomes possible. This blurring is incommensurabil-
ity. By restoring conventions to the broader picture of linguistic
understanding, Ramberg has made room for conceptual gulfs and
persistent failures of understanding, even while preserving the idea
that absolute intranslatability is impossible.
Or so it might seem. In fact, Ramberg is trying to have his cake
and eat it too, by relying on a fuzzy notion of convention. If we
sharpen the focus, we will see that Rambergs so-called conventions
v.kixo noov ron covr.n.+i\r rnirosornv 85
cannot do what he wants them to, which in turn will help us see
that the relation between meaning and the production of meaning
must actually run deeper than Rambergor Davidsonrealize.
Conventions, according to Ramberg, are heuristic devices, strate-
gic shortcuts, diachronic generalizations, and things that are con-
ventionally taken for granted (Ibid., pp. 1101). That is, what
Ramberg calls conventions are ways that we tend to talk, uncon-
scious habits that we persist in because they are useful. A rst point
that bears making is that this is not what most philosophers mean
by convention. According to one standard account, for instance, a
dierence between conventions and other regularities in our behav-
ior is that we follow conventions in part because we believe that
others will also conform, and we prefer that they do so.
On this
view, driving on the right in the U.S. is not just an unconscious
habit, convenient shortcut, or empirical generalization of peoples
behavior over time, though it may also be those things: It is a con-
vention established by the beliefs, desires, and intentions of U.S. dri-
vers. So Rambergs understanding of convention is unusual. In one
way, this is just as well, because as I discuss below, Robert Brandom
has shown that the more standard view of convention is inconsis-
tent with Davidsons central tenet that belief and meaning emerge
together. Still, the idiosyncrasy of Rambergs conventions suggests
that we should ask whether they can be conned to the limited role
he envisions for them.
Rambergs conventions, to repeat, are convenient (but dispensable)
shortcuts that facilitate the production of meaning. Our practical
dependence on these shortcuts leads us to the blurring of linguistic
understanding called incommensurability. But are conventions really
dispensable? An analogy that Ramberg uses to esh out his idea is
extremely revealing. He says that the individual speaker stands in
the same relation to the conventions of his/her language as Aristotles
phronimos does to the virtues. That is, If it is a convention to use
snow is white to say that snow is white, this convention is ham-
mered out only in a series of assertions that snow is white, in just
the way the meaning of a moral precept is hammered out for Aristotle
in the actions of the phronimos (Ibid., pp. 11112). The problem with
See (Lewis, 1969). Ramberg says that his own account is consistent with Lewis
(Ramberg, 1989, p. 113n1), but I see no evidence of this.
86 cn.r+rn +nnrr
this analogy is that according to Aristotle, one cannot understand
virtue apart from the actions of the virtuous person: they are con-
stitutive of virtue, not dispensable shortcuts. Ramberg makes it sound
like our repeated assertions, like the consistent actions of the phro-
nimos, simply imbue us with habits that become hard to shake. Of
course habit does play an important role for Aristotle, but virtuous
actions are not just any habits: they dene the standard of virtue,
and as such are normative: We should all emulate the actions of the
phronimos. Language, too, has this normative quality, as when we cor-
rect ourselves when we realize we have misused an expression.
Ramberg seems to think that self-correction can be understood on
his pared-down model of convention (Ibid., p. 110), but I cannot see
how. If I have come to habitually utter umm when pausing to
think, but then for some reason fail to do so, should I feel the need
to correct myself ?
I have begun to suggest here that language must be understood
in a more thorough-going normative fashion than Ramberg can allow
for, whether or not that means that we have to employ a more full-
bodied notion of convention. (In the end, I will agree with Brandom
that we must do without conventions, and follow his alternative.) A
broad notion of language is essential to understanding linguistic be-
havior, and not merely convenient. I believe that Rambergs and
Davidsons view that even if we used nothing but malapropisms, com-
munication would still be possible is wrong. As a supplement to the
considerations just oered about conventions, here is a more gen-
eral argument that Ramberg and Davidson (at his most radical) are
missing something crucial.
A central feature of Davidsons view of linguistic behavior is that
we are to be understood as possessing (largely implicitly) theories. I
mentioned the prior and passing theories briey above, and
quoted Davidsons statement that: I call it a theory . . . only because
a description of the interpreters competence requires a recursive
account. In much of his work he refers to these theories as truth
theories or T-theories, because he believes that truth conditions
that is, the conditions under which a given sentence is trueplay
critical roles in the theories. Davidson argues that a description of
the speakers competence requires a theory for two main reasons.
First, without a recursive theory able to generate innitely many new
theorems, we would be at a loss to explain our manifest ability to
generate and understand new sentences. Second, unless our under-
v.kixo noov ron covr.n.+i\r rnirosornv 87
standing of a given sentences is systematically inter-related to other
sentences (as well as sub-sentential units), interpretation will not be
Let me explain. In the imagined case where an interpreter uses
radical interpretation to construct a T-theory from the ground up,
we have no problem seeing that the theorems she attributes to her
interlocutor will be systematic. Clearly, we cannot demand that, in
order to have adequate evidence that a theorem is systematic, the
interpreter test all possible other theorems and axioms that might
bear on the truth of the theorem in question. On the other hand,
more is needed than our interpreter hearing her subject utter It
rains every Tuesday, and, observing that it is raining quite hard,
(A) The subjects T-theory states that It rains every Tuesday is true-
in-L at this time for her i it rains hard today.
to the speaker. This is clearly inadequate, since It rains every
Tuesday does not mean that it rains hard today. Of course, Davidson
never suggests that an interpreter should stop with (A). The provi-
sional attribution of (A) is no more than an appropriate step in the
long process of puzzling out a speakers T-theory, and it would be
followed by attempts to determine whether the speaker would utter
It rains every Tuesday on a clear day, on a day when it rains
lightly, and perhaps on a Tuesday without rain. These teststhe
very essence of radical interpretationwould lead our interpreter to
reject (A) since that interpretation, while adequate to explain her ini-
tial evidence, manifestly fails to be systematic, since it is not part of
a consistent theory.
Exposure to a long series of the speakers utterances is therefore
required for radical interpretation, as is the assumption that the
meanings of the speakers words do not change greatly over the
course of investigation. Without this assumption, we could have no
guarantee that a theorem could be conrmed or disconrmed by
the evidence: If the second time our interpreter heard It rains every
Tuesday she could not assume that it meant the same as the rst
time she heard it, she could not take the fact that it was not rain-
ing that day as tending to disconrm her initial hypothesis. Davidson
himself even suggests that it will be necessary to assume rough equiv-
alence of meaning across speakers in order to avoid unacceptable
level of indeterminacy:
88 cn.r+rn +nnrr
A theory for interpreting the utterances of a single speaker, based on
nothing but his attitudes towards sentences, would, we may be sure,
have many equally eligible rivals, for dierences in interpretation could
be oset by dierences in the beliefs attributed. Given a community
of speakers with apparently the same linguistic repertoire, however, the
theorist will strive for a single theory of interpretation: this will greatly
narrow his practical choice of preliminary theories for each individual
speaker. (In prolonged dialogue, one starts perforce with a socially
applicable theory, and renes it as evidence peculiar to the other
speaker accumulates.) (Davidson, 1984, p. 153)
The idealized process of constructing a T-theory for a speaker, we
can conclude, requires taking a language as the primary object of
ones theorizing. In one essay, Davidson acknowledges that, in some
of his early writings, he had neglected the possibility that someone
might know a set of theorems without knowing them to follow from
a T-theory (and thus without knowing them to be systematic) because
he imagined the theory to be known by someone who had con-
structed it from the evidence, and such a person could not fail to
realize that his theory satised the constraints (Davidson, 1984,
p. 173). Davidson does not go on to tell us how someone in a less
imaginary situation could be justied in attributing systematic theo-
rems to a speaker, but this is just the issue that we must now face.
We have seen that what justies the radical interpreter in mod-
eling her subject as possessing a T-theory isamong other things
both exposure to a long series of the speakers utterances, and
assuming that the meanings of the speakers words (and perhaps
even: the words of speakers in the relevant linguistic community) do
not change greatly over the course of investigation. Our problem is
to identify conditions that would justify an everyday interpreter in
modeling her subject in the same fashion. The answer seems obvi-
ous: both conditions are met, roughly, in the course of speaking a
language. We construct T-theories, that is, by relying on evidence
not from individual utterances, but from a language.
If the language has been returned to the center of our theoriz-
ing, we have come a long way from the notion that even if we
used nothing but malapropisms, communication would still be possi-
ble. Communication would not be possible under such circum-
stances. It is worth noting, though, that a language-based approach
can also handle the individual idiosyncrasies of T-theories which
seem to have driven Davidson to speak of prior versus passing the-
ories. Recall that Davidson did allow that things previously learned
v.kixo noov ron covr.n.+i\r rnirosornv 89
were essential to arriving at the passing theory. I think that a proper
appreciation of the force of this admissionso-called passing theo-
ries are essentially parasitic on theories that take as their subject a
whole language, rather than an occasion of utteranceshould con-
vince us to abandon talk of passing theories altogether. We begin
any linguistic interaction with a theory for the language-in-use of
our prospective interlocutor. As we converse, we may have to make
some adjustments to that theory. Some will be unconscious, auto-
matic, and passing, as when we take malapropisms into account.
Some will be conscious, lasting, and perhaps even carried out with
the interlocutors help, as when we need to ask What do you mean
by X? So long as X is another word for a concept that we already
possess, or stands for a simple concept that we can readily add to
our conceptual scheme, such adjustments will pose no diculties to
our communication.
Let me review. Comparative philosophers want two things from
a philosopher of language. On the one hand, we look for an assur-
ance thatsometimes only after hard workcommunication, com-
parison, and challenge are possible across languages, cultures, and
traditions. On the other hand, we do not want these things to seem
misleadingly easy, as if there were no diculties in arriving at cor-
rect translations or legitimate comparisons. Thusfar I have argued
that Davidson succeeds on the rst score but falls down on the sec-
ond. I looked at Rambergs promising-seeming eort to combine
Davidsons argument against untranslatability with the possibility of
incommensurability, but showed that both Ramberg and Davidson
himself still run into trouble by not taking seriously enough the role
of language in linguistic meaning. It is time to look at an alterna-
tive that strays farther from Davidsons theory than did Ramberg,
though still without giving up on the core orientation that provides
us with the rst of the assurances just mentioned.
The alternative I have in mind is Robert Brandoms inferential-
ist and pragmatist account of semantics. In the next several para-
graphs, I aim to show that although Brandom has little to say about
radical interpretation, he shares an approach with Davidson that
grounds the ultimate possibility of translation of all languages. He
carries this out quite dierently from Davidson, taking commitment
to be central instead of Davidsons stress on truth, but these dierences
do not overshadow their shared starting point in what Brandom calls
a relational theory of language and thinking. What is crucial for
90 cn.r+rn +nnrr
my purposes is that Brandom has found a way to do this that
nonetheless has a fundamental place for social norms. While Brandom
argues against talking of linguistic conventions, he shows that the
norms implicit in our practices play central roles in making possible
linguistic interactions. This role for norms will allow me to make
good on the second need of comparative philosophy, thus complet-
ing my task.
Recall that Davidson has argued that neither beliefs nor mean-
ings can be established independently of the other. This is illustrated
by the plight of a radical interpreter who knows neither what his
interlocutors words mean, nor what she believes. Davidsons solu-
tion is that we must provisionally assign true beliefs to the speaker:
only against a background of agreement can we come to identify
things on which we disagreethings that we think are true and she,
or they, think are false. In fact, Davidson goes father than this, argu-
ing that only in the context of interpretation can we make sense of
the notion of somethings being objectively true, as opposed to our
just thinking things are a certain way. Only when we can say some-
thing like She thinks that rabbits are bigger than hares, but I know
dierently do we have access to the dierence between thinking
that things are a certain way, and their actually being that way.
(This is independent of whether we are right about the way they
are, of course.) By putting interpretation at the center of his theory,
therefore, Davidson is able to accomplish two things: (1) show that
we must generally interpret people as speaking truly, and (2) show
how we come to appreciate the dierence between speaking truly
and speaking falsely.
Brandom follows Davidsons lead in arguing that the concepts of
objective truth and error necessarily emerge in the context of inter-
Brandom also puts the interpretive interaction between
speaker and hearer at the center of his theory. Like Davidson, he
believes that intentional states (cf. Davidsons beliefs) and speech
acts (cf. Davidsons meanings) are fundamentally of coeval con-
ceptual status, neither being explicable except in an account that
includes the other (Brandom, 1994, p. 152). It will not be lost on
readers, though, that meanings and speech acts are quite dierent
from one another; speech acts (like asserting) are typically thought
(Davidson, 1984, p. 169), cited in (Brandom, 1994, p. 152).
v.kixo noov ron covr.n.+i\r rnirosornv 91
to express meanings. Davidson says: start with someones behavior
that you think might be linguistic. The only way to arrive at the
persons beliefs and meanings simultaneously is to provisionally x
one, so you apply the Principle of Charity and assume that her
beliefs are true. Then try to build a theory of her language. If (in
principle) you cannot, then you conclude that she is not speaking a
languageher behavior was not linguistic, after all.
In contrast, Brandom says: start with that same behavior, which
you think might be linguistic.
To interpret someone as speaking is
to treat her as having taken part in a certain kind of practice, the
paradigm for which is assertion.
To perform an assertion is to take
on a certain kind of commitment: one becomes socially answerable
for ones performance.
If I say There is a red ball on that table,
and there is no ball on the table, or only a blue one, thensubject
to correction by further evidenceit seems that I am not playing
the assertion game. In the typical case, though, my interpreter will
understand me as having taken on various commitments through my
assertion, and will understand my utterance as an expression of these
commitments. Brandom argues that we should view the (semantic)
contents of my utterance in terms of the inferential relations I have
licensed by expressing these commitments: my saying that there is
a red ball on the table licenses the inference that there is a ball on
the table, that there is not nothing on the table, and so on. These
This paragraph is based on (Brandom, 1994, ch. 3); see in particular p. 142.
Brandom tends to use interpretation in the narrow sense, following Wittgenstein,
of explicit hypothesis formation. He notes that Davidson has been criticized for
thinking of ordinary intralinguistic understanding as this sort of interpretation, though
Brandom himself remains silent on whether this criticism is apt. Brandom does
allow that his deontic scorekeeping is a kind of interpreting, but it is implicit,
practical interpretation (Brandom, 1994, pp. 5089).
Sam Wheeler has argued (in a personal communication) that since (1) com-
mitment carries with it the logic of obligation (according to which the addition of
a new premise to a valid argument cannot invalidate the argument), and yet (2)
what we ought to believe based on our existing beliefs is better understood accord-
ing to the logic of conditional probability, so (3) Brandom is wrong to talk of com-
mitment. Wheeler writes: The rules of successful language use are not like rules
of games, but more like rules of thumb, that is, generalized conditional probabili-
ties. I think he is wrong about this, and we can see this if we think about one of
Wheelers examples. The person who believes that chien means dog, but does
not think that a chien is a mammal (because he does not believe that dogs are
mammals) is not just rare, but wrong. If we were to encounter such a person, we
would expect that when we presented him with the appropriate reason or evidence,
he would acknowledge a mistakethat is, the violation of something the use of
dog committed him to.
92 cn.r+rn +nnrr
various inferences make up the meaning of the utterance. It seems
clear enough, then, that for Brandom, attribution of intentional states
(taking on the commitments which undergird inferences, which
Brandom calls doxastic commitments) and attribution of the par-
ticular sort of performance called a speech act (assertion or whichever)
must go hand-in-hand. Brandom says that when we do this, we are
acting as deontic scorekeepers, by which he means that we keep
track of the commitments and entitlements that people we are inter-
preting as speakers take on. If you utter There is a red ball on the
table, I score you as committed to a variety of inferences. If you
then say Its the only thing on the table, I update my deontic
scorebook, since there are now new commitments you have taken
on. If you then push the ball o the table, I update the scorebook
again: since our shared situation has changed, your deontic status
has changed without your needing to say, for instance, Now theres
nothing on the table.
Instead of Davidsons approach, according to which we attribute
sentences held true, then, we start with the practice of assertion.
Brandom is quite explicit about inverting a central principle of
Davidsons, even while preserving the insight it embodies:
The attitude of taking-true is just that of acknowledging an assertional
commitment. . . . Evidently this principle can be exploited according
to two dierent orders of explanation: moving from a prior notion of
truth to an understanding of asserting (or judging) as taking, treating,
or putting forward as true, or moving from a notion of asserting to a
notion of truth as what one is taking, treating, or putting forward a
claim as. (Brandom, 1994, p. 202)
Davidson has made it very clear that he thinks we should start from
a prior notion of truth (Davidson, 1990, p. 314), whereas for Brandom,
truth has an expressive role rather than an explanatory role.
That is, using the concept of truth permits us to say various things
about assertionto make explicit the connections between some
assertions and others, for instance. It is not something that can be
understood in advance of assertion and used to help us understand
assertion itself (Brandom, 1994, p. 202).
For instance: Specically linguistic practices are distinguished as just the social
practices according to which some performances have the signicance of under-
takings of assertional commitment (Brandom, 1994, p. 168).
v.kixo noov ron covr.n.+i\r rnirosornv 93
If Brandom eschews an explanatory role for truth, though, he can-
not appeal to Davidsons Principle of Charity to explain how radi-
cal interpretation gets o the ground: We cannot look to an
antecendent, shared notion of truth to (provisionally) x the beliefs
of interlocutors, in order to work out what their words mean. Instead,
I will argue that Brandom can look to at least three things to explain
how communication might get o the ground: shared circumstances,
shared inferences, and the default attribution of assertion. Before
looking at each of these, let me note where we stand in the argu-
ment. Since Brandom agrees with Davidson that understanding some-
one to be engaging in linguistic behaviorthat the noises she or he
is making constitute a languagecan only be accomplished through
successful interpretation, Brandom is on rm ground to reject the
idea that there might be an untranslatable language. By way of
clinching my account of Brandom as sharing this basic orientation
with Davidson, I now propose to explicate how Brandom would
motivate the idea that successful interpretation could be possible.
To begin with, we share circumstances with those whom we would
interpret. There are things in our world with which we both inter-
act. At least in most cases, assertional practice gets its empirical con-
tent via what Brandom calls language entry and language exit
transitions: that is, assertions we make upon perceiving something
(rather than upon hearing or reading or thinking of something), and
things we do upon hearing (or reading or thinking of ) something
(Brandom, 1994, p. 222). To borrow Quines famous example, a
rabbit runs by and our interlocutor says Gavagai. Does this mean
Theres a rabbit? Initially, of course, any interpretation is dra-
matically underdetermined by the evidence, but if we presume an
assertion has been performed, we can begin to try out assigning
dierent sets of deontic scores. The process will be a familiar one,
and I need not dwell on its details.
Two points are important to make, though. First of all, what about
Quines insistence that his Gavagai example shows how translation
is indeterminate? Wouldnt all situations in which we might trans-
late Gavagai as Theres a rabbit also be situations in which we
could translate it as Theres an undetatched rabbit part? Brandom
has an ingenuous, though quite technical, answer to this worry, based
on the strategy that he has developed to deal with singular terms.
See (Brandom, 1994, pp. 40912). The basic idea is that the natives linguis-
94 cn.r+rn +nnrr
Second, are we just assuming that our interlocutor picks out objects
the same way we do? If so, this starts to sound like Holliss bridge-
head that we rejected above. Mightnt Gavagai refer only to rab-
bits seen on sunny days, or to some sort of rabbit-like god? Certainly
it could. Suppose that we spend some time with our subject, and
begin to see that the assertionand concomitant commitmentswe
initially attributed was not entirely apt. We might come to under-
stand, for instance, that she sorts animals into fair-weather and foul-
weather types ( Ramberg, 1989, p. 84). This still provides the
language-entry transition that Brandom has said is crucial to secur-
ing empirical content for our assertions. In addition, Brandom empha-
sizes that deontic scorekeeping always involves keeping separate track
of what an interlocutor takes to follow from her or his commitments,
on one hand, and what (as we see it) actually follows, on the other
(Brandom, 1994, pp. 185 and 488). Suppose that we have come to
see that our subject does take Gavagai to signal the presence of
a rabbit-shaped deity. We know how to score such an assertion both
for her, and for ourselves. That is, when she says Gavagai, we
attribute to her a commitment to there being a rabbit-god nearby,
while we are entitled to inherit a commitment to a mere mortal rab-
bit being present. This is so even if our backs were turned and we
did not see the rabbit.
One thing that is shared and which undergirds interpretation, in
short, is our circumstanceseven though, as we have just seen, one
can come to see that an interlocutor may interact with those cir-
cumstances dierently than one does oneself. A second thing that
needs to be shared is what Brandom calls sapience, which he
explains in terms of a responsiveness to reasons (Ibid., p. 5). This
does not mean that for us to successfully interpret someone as speak-
ing a language, he must nd convincing all of our reasons, but rather
that the practice of assertion itself essentially involves committing
oneself to the propriety of various inferences, and the notion of infer-
ring is to be understood, says Brandom, as a certain kind of move
tic and other behavior provides us with no reason to attribute complex sortal cat-
egories like undetached part to them. For Davidson on related worries, see
(Davidson, 1984, p. 26n10). For discussion, see (Angle, 1994, pp. 11719).
Brandom gives examples and discusses related issues at (Brandom, 1994, pp.
48090). Particularly important is the way that pronouns and other anaphoric expres-
sions help us to communicate, even when our diering commitments lead us to
mean very dierent things by our words.
v.kixo noov ron covr.n.+i\r rnirosornv 95
in the game of giving and asking for reasons (Ibid., p. 158). Let us
return to someones having uttered There is a red ball on the table.
Unless the speaker is thereby committed to a whole range of infer-
encesand will be responsive to reasoning about themthen she
has not, in fact, spoken, but just made some noises that sounded
like words. She must deny There is nothing on the table and
arm There is a ball on the table. In general, our starting point
in radical interpretation will be our whole inferential apparatus, that
which is made explicit through the use of logical vocabulary but
which is implicit in our everyday linguistic practice.
So far my discussion of Brandom has only aimed to show that
Brandom provides as solid a grounding for comparative philosophy
as Davidson does, by showing that Brandom shares Davidsons abil-
ity to rule out untranslatable languages. To bring home my larger
argument, it now remains to show that the fundamental role played
by social norms in Brandoms account enables him to make good
on what I have been calling the second need of comparative phi-
losophy, namely to show how conceptual dierences can be robust
and important, even if radical incommensurability is not in the ong.
While Brandom argues against talking of linguistic conventions, he
shows that the norms implicit in our practices play a central func-
tion in making possible linguistic interactions.
The simplest place to start is with Brandoms rejection of conven-
tion. The most inuential account of conventionswhich is endorsed
by Ramberg (Ramberg, 1989, p. 113)is that of David Lewis (Lewis,
1969). According to Lewis, conventions are social regularities that
are sustained by various beliefs, intentions, and desires of the par-
ties to the convention. In addition to conforming to the convention,
they must believe that others will do so, to prefer that everyone so
conform, and so on. According to Brandom and Davidson, though,
intentions and meanings arrive together: neither can be prior to the
other. Brandom cites Davidson as follows: Philosophers who make
convention a necessary element in language have the matter back-
wards. The truth is rather that language is a condition for having
Without conventions, though, we seem to be without
(Brandom, 1994, p. 232), citing (Davidson, 1984, p. 280). Davidson does some-
times say that correct interpretation involves hearer matching how the speaker
intended to be understood, which might seem to make intention prior to interpreta-
tion; see, e.g, (Davidson, 1986, p. 442). Brandom points out, though, that Davidson
96 cn.r+rn +nnrr
resources to explain how social norms might structure and constrain
our meanings.
Brandoms solution is to look to norms implicit in our practices.
Rather than looking to conventions that we can dene in terms of
prior intentions, look to proprieties (that is, norms or rules) that we
acknowledge in practice. Brandom shows that Ludwig Wittgenstien
and Wilfrid Sellars both advanced this idea; he cites Sellars as follows:
We saw that a rule, properly speaking, isnt a rule unless it lives in
behavior, rule-regulated behavior, even rule-violating behavior. Lin-
guistically we always operate within a framework of living rules. (The
snake which sheds one skin lives within another.) In attempting to
grasp rules as rules from without, we are trying to have our cake and
eat it. To describe rules is to describe the skeletons of rules. A rule is
lived, not described.
Language, too, is lived rather than described. Dictionaries describe
what we meant by our words yesterday. Most of the time we still
mean the same things today, but if our usage has evolved, then the
lexicographers need to catch up. This is not to say that an individ-
ual can mean anything she or he wants with a given word: usually,
idiosyncratic usages are malapropisms and, if it is socially appropri-
ate to point out the error, the speaker will acknowledge his or her
The point, though, is that we authorize our languages
norms by what we do: what we say, the commitments we attribute
and acknowledge, and so on.
With an understanding of using language as one among the many
things that we do, it is straight-forward to see that specically lin-
does not take it that the contents of these communicative intentions can be made
sense of antecedently, in abstraction from interlocutors interpretation of one another
(Brandom, 1994, p. 670n6).
(Sellars, 1980, p. 135), quoted in (Brandom, 1994, p. 25).
Davidson reects insightfully on the tension between speaker intention and
hearer knowledge in (Davidson, 1991). As his discussion of Joyces Finnegans Wake
shows, radically idiosyncratic uses are not always mistakes, but can be eorts to
provoke the reader into an involuntary collaboration. Davidson notes that coopted
into Joyces world of verbal exile, we are forced to share in the annihilation of old
meanings and the creationnot really ex nihilo, but on the basis of our stock of
common loreof a new language. All communication involves such joint eort to
some degree, but Joyce is unusual in rst warning us of this, and then making the
eort so extreme (Ibid., p. 11). This seems exactly correct to me: our stock of
common lore is, in Brandoms terms, our stock of common commitments, some
of which we will come to violate as we interpret Joyces language. Thanks to Sam
Wheeler for this reference.
v.kixo noov ron covr.n.+i\r rnirosornv 97
guistic practice must be bound up with many other practices. Brandom
in fact emphasizes this when he talks about language entry and
language exit transitions: our words are intimately bound up with
what we perceive and do in our world. Like linguistic practice, our
many other forms of practice also have norms implicit in them: pro-
prieties and improprieties that shape how we interact with others
and with our shared world. We understand ourselves in and through
these practices, both linguistic and otherwise. When we seek to make
explicit these self-understandings, we often advert to notions like
community. A community is a group with whom we share (argue,
play, eat, shop, reason, etc.). We belong to many communities, and
they often overlap and have fuzzy borders. Brandom puts this thought
in terms of how, and with whom, we say we (Brandom, 1994,
p. 3). Sociologists, anthropologists, and historians have explored many
of ways in which our practices dene us. Or rather, they have
explored the ways that we dene ourselves through our practices,
for, once again, we authorize our practices and their proprieties by
engaging in them.
If linguistic practices cannot be neatly separated out from the other
practices through which we dene ourselves, then what we can
appropriately say becomes a complicated matter. Contrary to
Davidsons blas remark quoted above, failures of straight-forward,
word-for-word translations are often not accidents.
Brandom writes
that: When the prosecutor at Oscar Wildes trial asked him to say
under oath whether a particular passage in one of his works did or
did not constitute blasphemy, Wilde replied Blasphemy is not one
of my words (Ibid., p. 126). Wilde recognized, that is, that using
the word blasphemy brought with it certain commitments that
heand other like-minded individualsrejected, even if he were to
deny that a particular passage was blasphemous. Communication,
seen here as the shared eort to understand one of Wildes writings,
temporarily breaks down.
Comparative philosophy rarely involves dramatic, face-to-face
encounters between alternative communities, but such encounters do
help to sharpen the issues involved in comparison and communica-
tion. Mario Biagioli gives us such an instance in his discussion of
Two particularly relevant studiesrelevant because of the ways they examine
the inter-dependence of linguistic and non-linguistic practicesare (Bourdieu, 1974)
and (Biagioli, 1990).
98 cn.r+rn +nnrr
communicative breakdowns between Galileo and his Aristotelian
rivals; Alasdair MacIntyre imagines diculties of comparison when
Confucians encounter Aristotelians, or when a single individual is
torn between two communities.
According to Brandom, we are to
understand the resistance to accommodation and the diculty in
nding common ground experienced by parties to these encounters
in terms of the norms implicit in their (linguistic and other) prac-
tices. At the same time, Brandom assures us that where there is a
will to overcome dierences, we can communicate, for there are no
untranslatable languages. The courtroom scene I described above
was not well-suited to communication about the nature of Wildes
writings, just as both Galileo and his rivals, according to Biagioli,
had something at risk (their socio-professional identities) if they
successfully communicated. Comparative philosophy, though, seeks
to get beyond these barriers.
This is not the place to evaluate the specic recent proposals,
mentioned at the outset of my essay, in terms of which several com-
parative philosophers have proposed we can integrate or chal-
lenge or seek truth across traditions. My goal has instead been
to demonstrate how Donald Davidson and, more completely, Robert
Brandom have shown us both why the synthetic projects of com-
parative philosophy are possible, and why they can be dicult. These
synthetic projects lie at the core of what might more properly be
called an emerging global philosophy, and they are well worth their
(Biagioli, 1990), (MacIntyre, 1991), (MacIntyre, 1989).
My sincere thanks to my colleague Joe Rouse, for discussions on these themes
and for his very helpful comments on an earlier draft. Thanks also to all the par-
ticipants in the conference on Davidson and Chinese Philosophy, especially Sam
Wheeler, Ye Chuang, and Xiao Yang. Sams two-barrelled argument that Davidson
has a satisfactory account of norms, and that Brandoms is hopelessly mired in
the inappropriate logic of obligation, is ingeneous and challenging, though in the
end I am not convinced, as I have tried to explain.
Allinson, Robert E. (2001), The Myth of Comparative Philosophyor the Comparative
Philosophy Malgr Lui, in Bo Mou. ed., Two Roads to Wisdom? Chinese and Analytic
Philosophical Traditions, Chicago: Open Court.
Angle, Stephen C. (1994), Concepts in Context: A Study of Ethical Incommensurability, Ann
Arbor: University Microlms.
(2002), Pluralism in Practice: Incommensurability and Constraints on Change
in Ethical Discourses, in Michael Barnhart, ed., Varieties of Ethical Reection: New
Directions for Ethics in a Global Context. Lanham, Maryland: Lexington Books.
Biagioli, Mario (1990), The Anthropology of Incommensurability, Studies in the
History and Philosophy of Science 21, pp. 183209.
Bourdieu, Pierre (1974), The Economics of Linguistic Exchanges, Social Science
Information 16 (1974), pp. 645668.
Brandom, Robert (1994), Making It Explicit: Reasoning, Representing, and Discursive
Commitment, Cambridge: Harvard University Press.
Davidson, Donald (1980), Essays on Actions & Events, Oxford: Clarendon Press.
(1984), Inquiries Into Truth and Interpretation, Oxford: Clarendon Press, 1984.
(1986), A Nice Derangement of Epitaphs, in Truth and Interpretation: Perspectives
on the Philosophy of Donald Davidson, Oxford: Basil Blackwell.
(1991), James Joyce and Humpty Dumpty, in Peter A. French, Theodore
E. Uehling, Jr., and Howard K. Wettstein, eds., Philosophy and the Arts (Midwest
Studies in Philosophy, vol. XVI), Notre Dame: University of Notre Dame Press.
(1990), The Structure and Content of Truth, Journal of Philosophy LXXXVII,
pp. 279328.
(2001), Foreword, in Two Roads to Wisdom? Chinese and Analytic Philosophical
Traditions, Chicago: Open Court.
Grandy, Richard E. (1973), Reference, Meaning, and Belief , Journal of Philosophy
(1973), pp. 439452.
Hacking, Ian (1982), Language, Truth, and Reason, in Rationality and Relativism,
Cambridge: MIT Press.
Hollis, Martin (1970), The Limits of Irrationality, in Rationality, Oxford: Basil
(1982), The Social Destruction of Reality, in Rationality and Relativism,
Cambridge: M.I.T. Press.
Kripke, Saul (1980), Naming and Necessity, Cambridge: Harvard University Press.
Kuhn, Thomas S. (1983), Commensurability, Comparability, Communicability,
PSA 1982 2, pp. 669688.
Lewis, David (1969), Convention: A Philosophical Study, Cambridge, MA: Harvard
University Press.
Lukes, Steven (1970), Some Problems About Rationality, in Wilson, Bryan R.,
ed., Rationality, Oxford: Basil Blackwell.
(1982), Relativism in its Place, in Rationality and Relativism, Cambridge: MIT
MacIntyre, Alasdair (1988), Whose Justice? Which Rationality? Notre Dame: University
of Notre Dame Press.
(1989), Relativism, Power, and Philosophy, in Michael Krausz, ed., Relativism:
Interpretation and Confrontation, Notre Dame: University of Notre Dame Press, 1989.
100 cn.r+rn +nnrr
(1991), Incommensurability, Truth, and the Conversation between Confucians
and Aristotelians about the Virtues, in Eliot Deutsch, ed., Culture and Modernity,
Honolulu: University of Hawaii Press.
Ramberg, Bjorn T. (1989), Donald Davidsons Philosophy of Language, Oxford: Basil
Sellars, Wilfrid (1980), Language, Rules, Behavior, in J. Sicha, ed., Pure Pragmatics
and Possible Worlds: The Early Essays of Wilfrid Sellars, Reseda, California: Ridgeview
Van Norden, Bryan (2001), Mencius and Augustine on Evil: A Test Case for
Comparative Philosophy, in Bo Mou, ed., Two Roads to Wisdom? Chinese and
Analytic Philosophical Traditions, Chicago: Open Court.
Yu, Ji-yuan, and Nicholas Bunnin (2001), Saving the Phenomena: An Aristotelian
Method in Comparative Philosophy, in Bo Mou, ed., Two Roads to Wisdom?
Chinese and Analytic Philosophical Traditions, Chicago: Open Court.
David B. Wong
1. Interpreting the principle of charity
The principle of charity, under Donald Davidsons inuential con-
strual, says that we must interpret others on the assumption that
they are rational beings, talking about and navigating the same world
as we are. Otherwise, we shall not be able to interpret them at all
as holding beliefs or making intelligible utterances. Charity directs
us to optimize agreement between them and ourselves where ever
it is plausible to do so. The idea is to make them right, as far we
can tell, as often as possible.
Charity, Davidson explained, does
not enjoin us from attributing intelligible error. As our interpretation
of others takes shape, we might nd that it makes better sense to
attribute mistakes to them, given our emerging conception of how
they are interacting with the objects of their beliefs. However, there
is a limit to such attribution. To attribute massive error to them is
to undermine a crucial assumption of interpretation: that they are
forming beliefs about the same world as we are. If we were to
attribute to ancients the belief that the earth is at, and what is
more, virtually none of the other beliefs we have about the earth,
we would undermine the assumption that they have beliefs at all
about the earth.
In his recent writings on interpretation, Davidson noted that he
previously tended to construe charity in terms of maximizing agree-
ment in belief and that a more perspicuous statement of what
he had in mind all along is that agreement in beliefs should be
Donald Davidson (2001), Radical Interpretation, in Inquiries into Truth and
Interpretation, 2nd edition, Oxford: Clarendon Press, p. 136.
Donald Davidson (2001), Thought and Talk, in Inquiries into Truth and Interpretation,
2nd edition, p. 168.
104 cn.r+rn rotn
Rather than the most agreement, we need the right
sort of agreement that enables understanding of others. We should
try to reach agreement as far as possible, subject to considerations
of simplicity, hunches about the eects of social conditioning, and
of course our common-sense, or scientic, knowledge of explicable
Matters become further complicated with the recognition
that it is not just agreement in belief that must be optimized, but
also in the desires, values,
and intentions we attribute to others,
since we must understand their behaviors not only in relation to
what they believe about the world but also in relation to what they
want of it, value in it, and intend to do in it. To make sense of the
actions of others, we construe these actions as stemming from inten-
tions, which in turn stem from certain patterns of beliefs, desires,
and values. Some patterns make others intelligible to us and others
To see why Davidsons move from talk of maximizing to opti-
mizing agreement seems the right move to make but also gives rise
to many questions about how to interpret others, consider some uses
in ethical theory of the earlier maximizing version of charity. David
Cooper holds that We can only identify anothers beliefs as moral
beliefs about X if there is a massive degree of agreement between
his and our beliefs. His conclusion is that the principle of charity
refutes any signicant form of moral relativism. It is dicult to dis-
agree with Coopers point that a moral belief must have for its sub-
ject matter something connected with welfare, happiness, suering,
security, and the good life.
We would have strong reason to sus-
pect our interpretation of another beliefs as moral beliefs if we were
to construe them as having nothing to do with welfare, happiness,
suering, security, and the good life. However, this is not the same
as refuting a moral relativism that asserts signicant dierences are
over what is believed about these subjects. Quite a lot depends on
Donald Davidson (2001), Introduction to the 2nd edition of Inquiries into Truth
and Interpretation, p. xix; Radical Interpretation, p. 136.
Donald Davidson (2001), On the Very Idea of a Conceptual Scheme, in
Inquiries into Truth and Interpretation, 2nd edition, p. 196.
I am going to remain neutral, for the purposes of this paper, on the question
of how to understand a persons values in relation to her beliefs and desires, whether
they are beliefs, desires, or some combination thereof.
David E. Cooper (1978), Moral Relativism, in Midwest Studies in Philosophy 3,
pp. 101, 104. I discuss this kind of argument in David B. Wong (1984), Moral
Relativity, Berkeley: University of California Press, pp. 114116.
vnrnr nroixs 105
how one applies Coopers requirement that there be a massive
degree of agreement between our moral beliefs and those of another
person. It seems arbitrary to say that among competing interpreta-
tions of another persons beliefs, the best interpretation is the one
that simply produces the greatest number of overlapping beliefs, even
if one had condence in ones ability to count beliefs, which I do
not. This is one of the correct reasons, I believe, why Davidson has
corrected his earlier statements of the principle of charity, from max-
imization to optimization.
Michele Moody-Adams gives a more recent formulation of Coopers
argument, starting with the premise that understanding others requires
that there be quite substantial agreement about many of the basic
concepts that are relevant to moral reection.
She then leaps to
the conclusion that ultimate or fundamental moral disagreement
is not possible.
To validate this conclusion, she must hold that dis-
agreement over concepts relevant to moral reection is limited to
nonultimate or nonfundamental concepts. This seems a coher-
ent stance to take, but cannot be derived in a priori fashion from a
principle of charitable interpretation of others, as Moody-Adams
attempts to do. Why must groups have moral beliefs only if they
have precisely the same stock of basic moral concepts as we do?
Why is it not sucient to have agreement on some critical mass,
however the threshold is dened? And with respect to any one basic
concept relevant to moral reection, such as justice,
what counts as
having the same concept? When does one count two very dierent
conceptions of, say, justice, as two dierent interpretations of the
same concept, and when does one say that one has two overlapping
but dierent concepts? And when basic concepts such as justice and
compassion represent values, must other people set precisely the same
priority as we do on these values when they come into conict? If
they dont set the same priority, could their disagreement qualify as
a fundamental disagreement under some circumstances?
Michele M. Moody-Adams (1997), Fieldwork in Familiar Places: Morality, Culture, &
Philosophy, Cambridge, MA: Harvard University Press, p. 55.
Moody-Adams, 1997, p. 56.
Justice and compassion were example of some basic concepts given by Moody-
Adams in a paper read by her, entitled The Idea of Moral Progress, at the
Eastern Meetings of the American Philosophical Association, December 1998, in
Washington D.C.
106 cn.r+rn rotn
A Davidsonian principle of charity, properly interpreted as call-
ing for optimization rather than maximization, cannot possibly resolve
these questions by itself because the requirement of optimization itself
requires interpretation. In illustrating this point, Henry Richardson
has pointed out interpreting a philosophical text requires taking
account of the cognitive aims the authors had in writing what they
did. Is it more charitable, Richardson asks, for a translator of
Machiavellis The Prince to resolve ambiguities and seek to maximize
agreement between Machiavelli and the relevant audience? Or is it
more charitable to set him out as intentionally provocative and delib-
erately cryptic?
In this paper I accept that charity rules out the possibility of oth-
ers having beliefs and desires that are dierent from ours in radical
and sweeping ways. Aside from imposing this constraint against rad-
ical dierence, charity is less a denite principle of interpretation but
rather the assortment of the various ways we have of explaining the
talk and actions of others. I shall discuss a couple of strategies we
employ in interpreting othersanalogy and the attribution of ratio-
nalitywith special reference to the understanding of Chinese thought
and culture from an American perspective. To make others intelli-
gible by likening them to us can frequently involve accepting that
the analogies we use are extended and rough, consistent with accept-
ing signicant dierences between them and us. Often, this is good
enough, and investigating why this is good enough will lead to the
question of who we are, when we interpret others to have beliefs
and desires similar to ours. I argue that the we and the ours har-
bor signicant diversity in belief, and that any plausible interpretive
approach must presuppose a certain range of diversity in belief and
desire over human culture, and in particular in the range of values
that are central to particular cultures.
My strategy of argument consists in working backwards, as it were,
from interpretations of the moral tradition of Confucianism from a
contemporary American perspective. Such interpretations are by no
means uncontroversial, but they are least eligible as illuminating and
plausible interpretations, and that is all I need. The question I want
to address is how analogy and models of rationality operate within
Henry S. Richardson (1997), Practical Reasoning about Final Ends, Cambridge:
Cambridge University Press, pp. 268269.
vnrnr nroixs 107
these interpretations to increase our understanding of Confucianism,
both by likening the Confucians to us and by expanding our own
sense of human possibility.
2. Interpretation of Confucianism through analogy
Let me begin by identifying three signicant and distinctive features
of the Confucian ethic as it is represented in classical works of
Chinese philosophy, primarily the Analects. We understand these fea-
tures through noting their similarity with themes we nd to be famil-
iar and present in our own culture, but as with all analogies, the
similarities co-exist with signicant dierences. In all three cases, I
shall claim, we have little reluctance to accept the dierences.
The rst feature is the centrality of xiao ( usually translated as
lial piety.) It is a common feature of many cultures that one
should honor thy father and mother, of course, and it is not dicult
to nd analogies within American society to Confucian lial piety.
At the same time, the Confucian tradition is unusual in the strin-
gency of its duties to parents. The Analects 2.7 identies the require-
ments of xiao as going beyond providing parents material support
when they are elderly, but more fundamentally showing them its
not just about giving them food (supporting them when elderly), but
jing ( ). Jing originally applied to the attitude one should have when
sacricing to ancestors, an attitude of devotion to carrying out great
responsibilities to ones ancestral spirits. Analects 2.8 amplies the
nature of jing in saying that the young should take on the burden
when theres work to be done and let the old enjoy the wine and
food, but that hardly deserves to be called lial. Its the expression
on ones face or demeanor that is dicult to manage.
The scope of duties to parents includes taking care of what they
alone could have given oneones body. Zeng Zi, one of Confucius
students, is portrayed in 8.3 of the Analects as gravely ill and near
death. He bids his students to look at his hands and feet, and quotes
lines from the Book of Poetry to convey the idea that all his life he
has been keeping his body intact as part of his duty to his parents.
It is only now near death, he says, that he can be sure of having
been spared and thus fullling this duty to parents. This very idea,
that one must keep ones body intact as a duty of gratitude to ones
parents, has remained a central idea in Chinese culture.
108 cn.r+rn rotn
Why is xiao so central a virtue in the Confucian ethic? Part of
the reason seems to be a view about its centrality to the develop-
ment of ethical character. In Analects 1.2, one of Confucius most
prominent students, You Zi, says that being good as a son and obe-
dient as a young man (perhaps reading that one is obedient to ones
elder brothers) is perhaps the root of character, the basis of respect
for authority outside the family. One learns respect for others rst
for those within the family. Another part of the reason for the cen-
trality given to lial piety is the need to express gratitude to those
who have given one life and nurture. In Analects 17.21, Zai Wo
objects to the traditional length of mourning for ones deceased par-
ents. One year is enough to disrupt ones normal life in those ways,
but the traditional period of three years is too long, he says. Confucius
comments on Zai Wos lack of feeling in this regard. All children
are completely dependent on their parents for the rst three years
of life, he observes. Did not Zai Wo receive three years worth of
love from his parents?
The virtue and its rationale have analogues in American culture.
We can certainly recognize the themes of gratitude, the need to rec-
iprocate in some fashion for great gifts received, and the conception
of family relationships as pivotal in the development of character.
Such similarities of theme, however, seem to underdetermine the
centrality of lial piety and the stringency of its duties in Confucianism
and in the larger traditional culture. While we recognize such ratio-
nales for lial piety, we generally do not accord it nearly as central
a place in the catalogue of moral virtues, nor do we conceive its
duties to be so stringent. And the theme that one owes ones body
to ones parents and that it is deep ingratitude not to take care of
such a great gift is something that can be understood from an
American perspective, but is not generally accepted.
Do such dierences suggest that we have not correctly understood
Confucianism or traditional Chinese culture? I submit that we accept
such dierences as part of the normal range of human possibility,
perhaps because we can imagine ourselves having taken a path we
have not taken. The themes of gratitude, reciprocity, and the impor-
tance of the family in moral development are familiar to Americans
and at the same time have the potential for justifying a value hav-
ing to do with respecting parents that is far more stringent than the
one many Americans accept. And indeed, taking the Confucian per-
vnrnr nroixs 109
spective, an American might come to have the strange feeling that
that perspective makes more sense.
Consider another signicant feature of Confucian ethics: its inclu-
sion of an aesthetic dimension in its conception of a good and worth-
while life. Right action in Confucian is tting action. It expresses
appropriate care and respect for others in a manner that bets the
nature of ones social relationship to them and to other, more par-
ticular features of their situation and ones own. There are often
conventional forms for the expression of these ethical attitudes, rit-
ual forms called the li (
. To fashion oneself into a better person
is to become practiced in the performance of such li such that they
become second nature, but they must always express the appropri-
ate attitudes. To attain the proper balance between form and feel-
ing is to ennoble and beautify human nature. Antonio Cuas translation
of a passage from Xun Zi puts the point nicely:
human nature provides raw material, and constructive human eort
is responsible for the glorication and ourishing of elegant form and
orderly expression. Without constructive human eort, human nature
cannot beautify itself.
It is instructive that Cua draws analogies to the perception of qual-
ities in works of art, likening the grace or joy that can be seen an
accomplished ritual activity to the grace of a curve in a painting or
joy in a piece of music.
David Hall and Roger Ames observe that
the Confucian notion of the right action has much in common with
the artists choice of the right brush or the right color in the exe-
cution of a painting.
For the Confucian, doing the right thing means
not only doing ones duty, not just for the right reason, and not just
with the right feeling, with the proper grace and elegance that is
both an aesthetic end in itself but also bespeaks the ease and con-
tentment of one who has attained the virtues and realized ones
humanity. These uses of analogy help to make intelligible the aes-
thetic dimension of Confucian ethics, but they also illustrate how
Antonio S. Cua (2002), The Ethical and Religious Dimensions of Li , The
Review of Metaphysics 55, p. 481, translating from Ti-sheng Li, Hsn Tzu chi-shih,
Taipei: Hseh-sheng, p. 439.
Cua, 2002, p. 483.
David L. Hall and Roger T. Ames (2003), Chinese Philosophy, in E. Craig
(ed.) Routledge Encyclopedia of Philosophy, London: Routledge. Retrieved August 28,
2003, from].
110 cn.r+rn rotn
analogy can help. They take what is familiar to us in one context
(in this case, painting or music) and point to its occurrence in a
dierent context. The analogy illuminates if we can conceive of the
familiar taking place in that dierent context.
The third feature of Confucian virtue ethics is its emphasis on
harmonious relationships as a central part of the ethical life, as illus-
trated by the Analects on an adults relation to his parents. Notice
that in 1.2, the crucial dimension of moral development that is started
in the family is respect for authority. The strong Chinese preference
for harmony emerges in 2.6, where the Master says to give parents
no cause for anxiety except for illness. Consider 4.18, where Confucius
considers occasions on which ones own opinions as to what is right
or best can conict with ones parents wishes. One should remon-
strate with ones parents gently, he says. What he says next about
ones course of action if they are not persuaded is ambiguous: you
jing bu wei ( ), but whether one translates it as D.C. Lau does,
one should not become disobedient and remain reverent, or as
James Legge does, Do not abandon ones purpose to respectfully
persuading them, the value placed on harmony is apparent. The
value placed on it, as indicated by 4.18, does not require silence in
the face of real disagreement with ones superiors. Indeed, the Confu-
cian tradition celebrates the scholar-intellectual who says what he
thinks about the rulers methods and ends, often to the rulers face.
However, the ends served by such moral courage include the end
of harmony. Rulers who fail to govern for the good of the com-
munity, the state, and the nation must be called to account precisely
for the good of all.
Someone must have the authority to settle conicts, if only in the
sense of saying whose view prevails this time. Human beings have
yet to invent a society without having to designate such authority
and to inculcate some degree of respect for it. The reasons for pre-
ferring harmony are quite intelligible, but here again, the reasons
underdetermine the degree of preference for it manifested in the cul-
ture: e.g., informal negotiation involving interaction and reconcilia-
tion between the contending parties is still the traditional way of
resolving business disputes in China; informal mediation committees
operate to resolve disputes at the grassroots rural village and urban
neighborhood levels; and Chinese courts encourage mediation between
contending parties even after litigation proceeds have begun.
That Chinese culture should show this high degree of preference
vnrnr nroixs 111
for harmony does not seem to threaten its intelligibility to those on
the outside. Why? Some preference for harmony exists on the American
side of the comparison, to which analogy can be made, however
major the dierences the analogy leaves in place. The American side,
after all, embraces a signicantly diverse range of subcultures in
which a high degree of preference for harmony is shown. These sub-
cultures include, of course, Chinese-American and other Asian-
American subcultures, as well as Latino and Mexican-American
subcultures. Moreover, the various European-descended subcultures
of American society have in the past demonstrated a stronger pref-
erence for family harmony and cooperation within various levels of
community than they do now. It is in part this internal diversity
that helps to make Confucian values intelligible as a path we our-
selves could and in some cases have taken.
So far, I have been assuming that the relevant reference point for
understanding Confucianism is the contemporary American per-
spective, and up until the previous paragraph I have assumed that
this perspective is more or less unied. Of course, that is false. In
reality, we treat such perspectives as unied only for the sake of cer-
tain comparisons, for the sake of certain comparative purposes. In
other contexts, and for other purposes, we make much ado about
the dierences. The unication is at best relative to the purpose of
understanding a presumably distant culture that is more dicult to
comprehend. The us in the comparison between them and us is
diverse, and such diversity provides some of the analogies we use to
make sense of them. This raises the question of how this diverse
group became us in the rst place.
If we are limited to beginning from our individual selves as mod-
els for understanding others, it seems quite unlikely that we could
get the range of beliefs, desires and values that we take for granted
even within relatively small circles. I accept that some people are
attracted to holding power and exercising it over others, even though
my own psychology does not bear much resemblance to theirs in
this respect. I accept that some of my students believe in the extreme
libertarianism of Ayn Rand, even though there is an inevitable point
where I fail to follow their thought processes when they explain them
to me. I accept that some people believe that they have been abducted
by alien beings, even though there is very little from my own expe-
rience that I could use to illuminate whatever experience and thought
processes could have led to such a belief.
112 cn.r+rn rotn
I suggest that our self-understandings comprehend signicant diver-
sity because the very concepts we use to interpret both ourselves and
others embody diversity. Consider the concept of rationality.
3. The model of rationality and the problem of diversity within ourselves
The attribution of rationality to others is another dimension of the
meaning of charity as an interpretive principle. Not only must be
interpret others to be interacting with the same world as ours, but
we must understand them as processing their interaction with the
world in basically the same way as we do, through the formation
of beliefs, desires, values and intentions. Because certain inferential
patterns among these elements make sense to us and others do not,
as we navigate ourselves through the world, we can make sense of
other peoples actions only if we can attribute to them similar enough
We can certainly be justied in attributing errors of reasoning to
others. We are familiar with errors of reasoning in our own cases,
and more fundamentally, we learn what good reasoning is not only
by being given exemplars of good reasoning but also of typical bad
reasoningovergeneralization from an insucient number of instances
and wishful thinking that conforms beliefs too directly according to
what we desire to be true. It is eective to provide exemplars of
bad reasoning precisely because human beings have shown them-
selves prone to engage in it. Indeed, it would not surprising if the
ratio of bad inferences to good ones made by the human race at
any given time would be a fairly high number. This is not to deny
that our interpretations of each other presuppose a normative ideal,
but we are often best interpreted as deviating from the ideal quite
often for intelligible reasons.
The extent to which human beings do deviate from the ideal of
rationality is an open, empirical question because we are capable of
taking a third-person scientic perspective on ourselves as well as
projecting ourselves into each others places and imagining what pat-
terns of beliefs and desires and values would make sense of each
others actions. A scientic understanding of human beings might
indeed require revision of the model of human agency we presuppose
in commonsense interpretation. Empirical studies indicate that char-
acter traits, dispositions to action, and values might be a lot more
vnrnr nroixs 113
situationally sensitive than we have supposed them to be and in ways
that escape our self-understandings.
Consider the study of seminary
students who were presented with a coughing man slumped in an
alley. The most powerful predictor of whether a student helped was
whether he was in a hurry for an appointment. By comparison,
whether a student had just before been preparing a sermon on the
Good Samaritan (!) or not had no predictive power.
In another
experiment, being in a good mood, prompted by nding a stray
dime, renders most people much more likely to help a stranger.
Reections in the previous section suggest, moreover, that not all
diversity comes under the heading of error and failures of intelligi-
bility. I think there is a good deal of truth in Allan Gibbards sug-
gestion that we are guided in conceiving what makes sense for human
beings to think and to feel by the imperatives of social coordina-
We exert pressure on each other when we propose what makes
sense in the way of thinking and feeling and acting, and this pres-
sure can result in social coordination that might not otherwise be
possible. However, there is no reason why there cant be a variety
of models of what makes sense, encompassing a variety of behav-
iors, patterns of relationships between beliefs and desires, and
congurations of values. Natural selection, after all, works through
genetic variation over time, and it is likely that human beings within
a cooperating group are likely to show such variation in psycholog-
ical drives and temperaments.
The idea that our conceptions of what makes sense has the prag-
matic function of social coordination can help to explain why our
conceptions of we and us embody signicant diversity. Not only
is it true that human beings are likely to show a variety of psy-
chological drives and temperaments, but this variety also can prove
See John Doris (2002), Lack of Character: Personality and Moral Behavior, Cambridge:
Cambridge University Press; Gilbert Harman (199899), Moral philosophy meets
social psychology: virtue ethics and the fundamental attribution error, Proceedings of
the Aristotelian Society 99, pp. 31531; and Gilbert Harman (19992000), The non-
existence of character traits, Proceedings of the Aristotelian Society 100, pp. 2236.
J.M. Darley, and C.D. Batson (1973), From Jerusalem to Jericho: A study
of Situational and Dispositional Variables in Helping Behavior, Journal of Personality
and Social Psychology 27, pp. 100108.
A.M. Isen and P.F. Levin (1972), Eect of Feeling Good on Helping: Cookies
and Kindness, Journal of Personality and Social Psychology 21, pp. 3848.
Allan Gibbard (1992), Wise Choices, Apt Feelings: A Theory of Normative Judgment,
Cambridge, MA: Harvard University Press.
114 cn.r+rn rotn
useful to the group. The more dierentiated social roles become, the
more useful it would be to have dierent kinds of people who are
suited for dierent roles through their distinctive abilities, styles of
social interaction, and preferences. Our conceptions of what makes
sense, then, may be deeply inuenced by the pragmatic need to
include within the scope of the normal and acceptable the variety
of kinds of people that can make social cooperation ourish. Another
way to put the point is that saying to another person that her thoughts
and actions make sense is not only to build a bridge of commu-
nication to that person but lay the groundwork for cooperation with
her, and the need to do these things can be a powerful force for
making the concept of we and us signicantly diverse.
A very common variation in natural temperament with which it
is very important for human beings to deal is precisely variation in
preference for harmony within groups. Arguably, a good part of the
importance of adhering to rites (li ) in the Confucian ethic lies in
taking whatever natural preference for harmony that individuals pos-
sess and strengthening it in relation to competing drives by engag-
ing in customary forms of behavior that express attitudes of respect
and concern for others. 12.1 says, To return to the observance of
the rites through overcoming the self (ke ji constitutes ren ).
An ethic that places a strong emphasis on harmony is one type
of solution to the problem of fostering social coordination among
human beings. However, it is not at all surprising that other ethics
would not place as much emphasis on the value of harmony but
rather, say, on permitting conicts in certain domains and keeping
them under control through restrictions on the manner in which
they are conducted and the measures taken to win. Our models of
rational resolutions to problems of social coordination, then, should
plausibly encompass a variety that is in keeping with the variety in
temperaments, preferences, and values that we have every reason to
expect to nd in human beings given what we have been taught to
expect from those in our immediate vicinity and given the best sci-
ences we can apply to the project of understanding ourselves.
Moreover, we can expect to revise our models of rational delib-
eration upon further scientic understanding of our capacities and
The line of thought in this paragraph was provoked by stimulating comments
from Michael Krausz and Yang Xiao for comments in response to the presented
version of this paper in Beijing.
vnrnr nroixs 115
limits. Models of maximizing expected utility have typically been
held up as ideals for rational deliberation, but doubts about such
models have emerged given the typical limitations on our knowledge
of the consequences of the actions we consider, our ignorance regard-
ing what value we will ultimately attach to those consequences we
do foresee, and our capacity to take into account only some of the
alternatives available in any choice situation.
Alternative models
better suited to human limitations embody the idea of selecting sat-
isfactory means to our ends, rather than the best means, where a
satisfactory means might be the rst option considered that is good
or the option that is not clearly inferior to any other
Perhaps alternative models of rationality are more or less
suited to particular problems and circumstances. Maximizing mod-
els are better suited to problems where the time frame that must be
considered is relatively short and where the relevant information is
largely available and preferably in quantiable form.
4. Conclusion
Understanding others is partly making them like us, but it is more
like nding ways in which we overlap with them rather than nding
identities between them and us. That understanding often turns on
loose analogies should not be surprising when we consider how inter-
nally diverse we ourselves are, and how much we can surprise our-
selves as well as others. We expect that others and ourselves may
sometimes become opaque to us in the sense that their actions do
not make rational sense, and we incorporate some of these patterns
of irrationality into our conception of what it is to be human. Our
very conception of what is rational includes a signicant range of
desires and values, partly because of the mechanism of natural
selection through genetic permutations. And if determining what is
rational is partly a matter discovering what processes of deliberation
H.A. Simon (1976), Administrative Behavior, 3rd edition, New York: The Free
Press, p. 81.
Philip Pettit (1984), Satiscing Consequentialism, Proceedings of the Aristotelian
Society supplementary volume 58, pp. 165176.
Gary Klein (2001), The Fiction of Optimization, in Gerd Gigerenzer and
Reinhard Selten (eds), Bounded Rationality: The Adaptive Toolbox, Cambridge, MA: MIT
Press, pp. 101121.
116 cn.r+rn rotn
and inference are best adapted to limited creatures such as us in
our particular environments, our ideals of rationality must to some
extent always be defeasible and open to the possibility that a range
of solutions to human problems may have equal claim to be called
Yiu-ming Fung
1. A.C. Grahams Sinologists Criticism and the
Myth of Pre-logical Thinking
A.C. Graham, a widely respected Sinologist, may be the rst scholar
in the context of Chinese philosophy to express opinions counter to
Donald Davidsons principle of charity and to his view on the very
idea of a conceptual scheme. As a Sinologist with a comparative
perspective based on a strong British theological background and on
a long-term experience through energetic work in Chinese Studies,
Graham has made signicant contributions to the eld of Chinese
philosophy, especially in his interpretations of Chinese texts and his
explanations of the problems in the eld. Grahams contributions
seem inseparable from his special status in terms of his reading mean-
ings from or into Chinese texts with a comparative perspective from
a double eyea British eye together with a Chinese eye. It seems
to Graham a basic faith that his and his colleagues comparative
studies with bilingual capabilities (or to use my metaphor, with a
double eye) are signicant and that the comparisons between Chinese
and Western thoughts are understandable or intelligible though they
are based on very dierent conceptual schemes which fundamentally
have very little in common. In this regard, Davidsons challenge to
the very idea of a conceptual scheme, I think, is also a challenge to
the Sinologists basic faith.
In his rebuttal to Davidsons thesis, Graham declares: For inquir-
ers into the thought and language of other cultures, the issue is
inescapable. That very idea [of a conceptual scheme] is one of their
indispensable tools, to which Davidsons objections do not directly
apply, since their own tendency is to think of it in terms, not of
propositions, but of classication by naming, and perhaps of syntactic
118 cn.r+rn ri\r
One of the main reasons for Graham to reject Davidsons
objections is that [a]t the roots of the systems of propositions called
conceptual schemes by philosophers there are patterns of naming,
pre-logical in the same sense as patterns of perception are pre-
To illustrate this point more specically, Graham mentions
Le Galls and Bruces failure in using forme (or law) and matire
(or matter) to translate li ( ) and qi ( ) in the texts of Song-
Ming Confucianism as a starting point of his journey to search for
a pre-logical or even pre-linguistic realm.
He claims that to think
of Le Gall and Bruce as making mistakes which we now avoid would
miss the whole point. There are no exact equivalents for li and qi
among our concepts, and there is no way of approaching them except
by breaking out from or awakening to one analogy after another.
Because he believes that all thinking is grounded in analogization,
that the metaphorical root behind Westerners matter and law
are dierent from that behind Chinese qi and li, and, unfortu-
nately, that the outsider, unlike the insider who habitually thinks
with their concepts, are much less conscious of the dierences at the
What are the dierences at the bottom? Borrowing Roman Jakob-
sons ideas of paradigm/syntagm and metaphor/metonym, Graham
argues that, at the bottom of each language or thinking, there is
some kind of pre-logical patterning of names that is a stock of par-
adigms already grouping syntagmatically in chains of oppositions
which at their simplest are binary.
He thinks that, for example,
before entering into sentences, we can have the compound words as
daylight and such formulas as the light of knowledge syntagmat-
ically grouping from a stock of paradigms which consists of binary
oppositions such as day and night, light and darkness, and
knowledge and ignorance, etc. When comparing king with lion
as men with beasts, by metaphor we can have the lion as king of
the beasts and the king as a lion among men. King connects with
throne as chairman with chair, so by metonymy the monarchy is
A.C. Graham, 1992, p. 59.
A.C. Graham, 1992, p. 207.
A.C. Graham, op. cit., p. 61.
A.C. Graham, op. cit., pp. 6162.
A.C. Graham, op. cit., p. 62.
r.\irsoxs 119
called the throne and the chairmanship the chair.
Based on these
ideas coming from what Graham calls Semiology, he believes that
we can nd the beginning of a conceptual scheme in these chains
of oppositions. He calls the thinking in these chains correlative in
contrast with analytic in the sense that the former is conceived
as spontaneous, pre-logical, and operating at the level of the non-
sentential combinations of words, while the latter is discursive, logical,
and operating at the level of propositions.
Based on the idea that a conceptual scheme is not a system of
logically related propositions but a pre-logical pattern of names,
Graham argues that a comparison of Western and Chinese concep-
tual schemes should begin at the level of non-sentential units. Through
his double eye, he seems to be able to see that the structures of
Chinese conceptual scheme are exposed nakedly by the tendency
to parallelism in the classical language, and are overly formulated
in the Yin-Yang ( ) cosmological scheme.
In comparison with
the Western scheme which tends to center on conicting opposites
(truth/falsehood, good/evil), he points out that the Yin-Yang scheme
is focused on complementary polarities. Some of the English chains
of oppositions such as day/night and light/darkness seem to t
neatly into a Yang/Yin scheme, but, as stressed by Graham, in the
latter A and B are interdependent with A only relatively superior,
and the chain does not lead to conicting duality such as good/evil.
So, he concludes, [o]ur conceptual schemes dier, not in assuming
the truth of contradictory propositions, but in including or exclud-
ing dierent pairs of words.
In contrast to the complementary char-
acteristic of Chinese thinking, Graham thinks that David Hall and
Roger Ames are right in demonstrating that the West habitually
treat[s] A as transcendent in the sense that A is conceivable with-
out B but not B without A; for Westerners there could be God with-
out world, reality without appearance, good without evil.
Grahams opinions mentioned above are mainly based on his de-
viant notion of conceptual scheme and his interpretations of Chinese
concepts used in ancient Chinese texts, especially in philosophical
texts. His arguments, if valid, assume that there are two levels of
A.C. Graham, 1992, pp. 6263.
A.C. Graham, op. cit., p. 64.
A.C. Graham, op. cit., p. 65.
120 cn.r+rn ri\r
thinking (analytic and correlative) and, most importantly, that the
latter, in contrast with the former, is spontaneous, pre-logical, and
even pre-linguistic though it is presented as a pattern of oppositions.
Besides the mystical characteristic of the so-called pre-logical and
pre-linguistic thinking, he also assumes that the meaning of a sen-
tence is dependent on the meaning of the words which occur in the
sentence though the truth of the sentence is independent of its com-
ponent words. In this regard, Graham seems to adopt an atomistic
or building-block theory of meaning which is opposite to Davidsons
holism. In the following, I will argue that Grahams criticism is not
accurate both in the sense that his mystical idea of pre-logical think-
ing together with his atomistic theory of meaning is not well argued
and in the sense that his interpretations of the Chinese concepts are
not well grounded in Chinese sources. I will also point out that,
even if we accept, for the sake of argument, his idea of the two lev-
els of thinking, his explanation of the dierences of Chinese and
Western conceptual schemes is more consistent with Davidsons prin-
ciple of charity than his idea of bilingual but distinct understanding
would suggest.
Why do I think Grahams mystical idea of pre-logical thinking is
not sustainable? One of the major reasons is that the so-called pre-
logical chain of oppositions is a self-contradictory description (How
can we make sense of oppositions without logic?). Although Graham
borrows his idea of binary oppositions from Jakobson, unlike Graham
who understands them as pre-logical, Jakobson stresses their logi-
cal structure.
Jakobsons idea of markedness is based on the log-
ical nature of oppositions applied both at the level of the signier
and at the level of the signied.
In Semiology or Semiotics, we
know that there is a semiotic square, which is adapted from the
logical square of oppositions, to characterize dierent kinds of oppo-
sitions such as contradiction, contrariety, and complementarity or
implication. It appears unintelligible to say that these binary rela-
tions are oppositions without any logical sense. How can Graham
Roman Jakobson, 1978, p. 115. He and his student Morris Halle also remark
that the binary opposition is a childs rst logical operation. See Jakobson and
Halle, 1956, p. 60.
Jakobson mentions that [e]very single constituent of any linguistic system is
built on an opposition of two logical contradictories: the present of an attribute
(markedness) in contraposition to its absence (unmarkedness). See Jakobson, 1985,
p. 85, cited in John Lechte, 1994, p. 62.
r.\irsoxs 121
claim that they are pre-logical without any supporting argument?
If Grahams idea is not mere speculation, there is still a question
whether this chain of oppositions in peoples mind is innate or
obtained by learning. If it is innate, people from dierent cultures
should have the same chain reected in their spontaneous reactions
unless there is some kind of biological ground or genetic evidence
to explain why two dierent races or two dierent groups of peo-
ple from dierent cultures innately have dierent chains of opposi-
tions. If dierent chains of oppositions reected in dierent cultures
are basically grasped by learning, it is impossible for us to under-
stand a chain of oppositions as pre-logical. Because what items
selected and correlated as opposites from learning cannot be under-
stood in a space without any logical relations. How can we know,
from learning, if two items (linguistic or ontological entities), such as
truth/falsehood and good/evil (in English), are conicting, or other
two items, such as Yin/Yang and day/night (in Chinese), comple-
mentary? If all these are grasped by learning and not innate, a native
speakers seemingly spontaneous reaction in using his or her words
in a chain of oppositions should not be considered as dierent in
nature from a skillful swimmers or a mature drivers reaction. Some
Westerners viewing day and night as conicting may be based on
a perspective emphasizing the dierent characteristics between them,
while some Chinese perception of them as complementary is prob-
ably based on a dierent perspective focusing on their alternation.
I think both people can easily understand each others opinion without
presupposing the necessity of going back to Grahams pre-logical
While the idea of pre-logical discussed above can be considered
as self-refuting, the idea of pre-linguistic can be understood as mys-
tical. It is interesting that Graham interprets human beings reaction
of correlation as pre-linguistic on the one hand, and treats this reac-
tion as similar to Pavlovs dogs conditioning on the other.
In dis-
cussing the problem whether animals have thought as humans have,
Davidson also mentions the dog, in addition to his example of the
y. In this regard, Graham seems to be totally ignoring Davidsons
arguments about the concept of belief or the belief of belief . It
is obvious that the idea of pre-linguistic correlation is inconsistent
A.C. Graham, 1992, p. 207.
122 cn.r+rn ri\r
with Grahams other description of the same thing, a pre-logical
patterning of names (How can we have names by pre-linguistic cor-
relation?). However, lets put aside the problem of inconsistency and
focus on this dog-like correlation. According to Davidsons view in
his Thought and Talk, a dog probably knows that its master is
home but does not know that Mr. Smith (its master) is home and
the president of a bank is home.
A dog, unlike a rational creature,
cannot have a thought in such an intensional context. So, a speech-
less creature may be able to discriminate something from other and
may be conditioned to make similar reaction to similar stimulus, but
it does not mean that it entertains a distinction between what is
believed and what is the case. In other words, a dog does not have
a belief which is known to be either true or false. In order to have
a belief known to be either true or false, a creature must have the
belief of a belief, i.e. the concept of a belief. In Davidsons words,
[t]he reason neither a dog nor any other creature can have a sin-
gle belief, such as that it is seeing a cat, is that what identies a
belief is what we loosely call its propositional content. Thus, to have
a belief about a cat, one must have mastery of the concepts that
are involved in this judgement of belief. A creature does not have
a concept of a cat merely because it can discriminate cats from other
things in its environment. Mice are very good at telling cats apart
from trees, lions, and snakes. But being able to discriminate cats is
not the same thing as having the concept of a cat. You have the
concept of a cat only if you can make sense of the idea of misap-
plying the concept, of believing or judging that something is a cat
which is not a cat. To have the concept of a cat, you must have
the concept of an animal, or at least of a continuing physical object,
the concept of an object that moves in certain ways, something that
can move freely in this environment, something that has sensations.
There is no xed list of things you have to know about, or associ-
ate with, felinity; but unless you have a lot of beliefs about what a
cat is, you dont have the concept of a cat.
A dog doesnt have
a single belief (a rst-order belief ), because it doesnt have the belief
of a belief (a second-order belief ). If Westerners thinking or Chinese
peoples thinking at the so-called correlative level is dog-like as
Graham describes, it would be reasonable to treat the thinking as
Donald Davidson, 2001, p. 163.
Donald Davidson, 2002, p. 124.
r.\irsoxs 123
pre-logical without being true or false. Nevertheless, is it possible
for Chinese people to have a chain of oppositions, such as day/night
and Yin/Yang, without some beliefs about daytime and sunshine and
some other beliefs about nighttime and the Moon? Is it possible for
Westerners to have another pattern of naming, such as good/evil
and true/false, without some intentions about behaviors and some
beliefs about sentences? If our conceptual thinking really operates at
the very beginning on Grahams underground level, Chinese think-
ing (in ancient times) on this level, for example, would have day/night
opposition without backing up by any beliefs about day and night.
In other words, they would think of day/night or Yin/Yang as
complementary without believing that There is sunshine at day-
time, The moon appears at night, and so on. Can we imagine
that Chinese people (in ancient times) are thinking of day/night
correlatively rst without any beliefs and then later thinking about
them analytically with propositional attitudes? What is the rationale
for this learning process? How can Graham identify these correla-
tive concepts or ideas without assigning some background knowledge
to the speakers?
If we accept, for the sake of argument, that the schemes them-
selves are patterns of names which are neither true nor false, and
that factual statements depend on them for their meaning but not
for their truth, Grahams building-block theory would still have more
trouble than the ordinary version of the theory. His words or names
used at the pre-logical level (of course, it is impossible for them to
be used at the pre-linguistic level) must be context-free, otherwise
they would appear in a context that connects them to other words
for identifying their meaning or sense, and this context must be sen-
tential. He thinks that learning Chinese words through guess from
the entries of Mathewss dictionary would never be at home; instead,
to understand the meanings of these words, such as an ( ) and its
opposite wei ( ), is not to analyze them in a logical way but to
correlate them within dierent contextual patterns.
However, these
contextual patterns are not patterns in sentential context, but pat-
terns of naming at pre-logical level. This may be one of the rea-
sons why he considers his pre-logical terms as having no other
content than the oppositions themselves.
But, when he uses some
A.C. Graham, 1992, p. 67.
A.C. Graham, op. cit., p. 209.
124 cn.r+rn ri\r
examples to illustrate the meanings of these sentential context-free
terms or names, it indicates clearly that the terms contents are more
than the oppositions themselves. He says, for example, to be yang
not yin is nothing else but to be light not dark, or male not female.
It is obvious that the example is not consistent with his no more
content thesis, because to be yang not yin is not the same as to be
A not B or to be A not ~A. In order to know the similarity
between yang and light or male and the dierence between each pair,
we have to know some contents more than just oppositions. If the
terms have no other content than oppositions, then to be yang not
yin would have no dierence from to be A not B or to be A not
~A. If so, how can this abstract and logical idea be understood at
the pre-logical level?
Since the dierences in meaning between (ancient) Chinese and
English words and sentences, for Graham, are to the bottom, at
the pre-logical level, it seems natural for him to claim that Mao-
wo-zai-xi-zi-shang ( ) cannot be translated into The cat
sat on the mat, Cao-qing ( ) into Grass is green, and yang
( ) into sheep. To the bottom, it seems that Chinese and English
correlate things or divide up the world dierently. But Grahams
explanations for these examples are not digging into the bottom if
there is one. Although he mentions that wo ( ) does not express
the same posture as sit and that the classication of oor cover-
ings for xi-zi ( ) and that for mats are dierent, this expla-
nation is denitely not about something underground but about
contents, more than just correlations and oppositions. Grahams spon-
taneous thinking seems to exclude the possibility of having background
beliefs about xi-zi in naming xi-zi. However, as he describes, in
Chinese xi-zi is used of [sic] straw mats. In other words, it is something
not made of cloths but straw only. This indicates that, when ancient
Chinese people using the word xi-zi , it is impossible for them not
to know that xi-zi is made of cloths is false and xi-zi is made of
straw is true. Grahams bottom-claim is also not supported by his
example of the unintertranslatability between Cao-qing and Grass
is green. He mentions that the concept of cao ( ) has a wider scope
than that of grass, qing ( ) as blue-green is based on Chinese pri-
mary Five Colors which are dierent in division or classication from
A.C. Graham, 1992, p. 209.
r.\irsoxs 125
English primary Seven Colors. However, his term wider scope means
wider extension and to identify the scope or extension of a con-
cept is to identify the applicability of the concept, to test whether it
is true or false in applying a concept to some object. Although
Graham stresses that [n]aming is by contrast within the scheme
rather than by adequacy to the object
and that to correlate Chinese
and English words need not assume coincide in extensions,
it is
obvious, in the explanation of this example, that he has to use the
idea of scope or extension to identify the dierences.
Grahams example about the unintertranslatability of yang and
sheep is also not consistent with his bottom-claim. When reading
the explanation in two languages of a vocabulary dierence between
them, Graham thinks that one is positively grateful that they do
not say exactly the same thing, much as when collecting informa-
tion about an incident one wants photographs taken from dierent
angles at dierent moments. It means that from a Chinese angle
we can see that the scope of yang includes both things English peo-
ple call sheep and goat; but from an English angle we can see
that the scope of sheep is exclusively dierent from that of goat.
Like the case that seeing or photographing an accident from dierent
angles presupposes there is one and the same accident, this case also
requires that there is one and the same object in plain sight; oth-
erwise, we cannot have dierent perspectives or dierent descrip-
tions of the (same) object in plain sight.
Graham identies his idea of correlative thinking or naming as
pre-logical, pre-linguistic, spontaneous, and also mystical. It is
mystical in the rst sense that Grahams idea presupposes some kind
of mentalist meaning which is independent of analytic thinking; it is
mystical in the second sense that Graham treats his pre-logical pat-
tern of names as the products of a classifying act of naming with-
out understanding them as the singular terms of logic. Nevertheless,
in what sense could this kind of naming be understood as a clas-
sifying act? Isnt it necessary to have some sense in which something
can be named rightly under a classifying act while others cannot?
If Chinese classifying yang is dierent from English classifying sheep
in the sense that the English word sheep cannot be used to refer
to a goat but the Chinese word yang can be used to refer to both
A.C. Graham, 1992, p. 212.
A.C. Graham, op. cit., p. 68.
126 cn.r+rn ri\r
a sheep and a goat, Chinese people should know at the same time
that it is right to say yang to cover both sheep and goats and wrong
to refer to sheep only. It seems to be the same situation as indi-
cated by Davidson that we can master the distinction between erro-
neously applying the concept cow to bulls when faced by a bull and
correctly applying a concept that covers both cows and bulls through
a test of learning to explain errors.
To compare a foreign language
with home language, I think we can also use this kind of test for
learning to explain the dierences. For the same reason, if someone
says Thats yang when faced by a sheep consistently, and says the
same thing when faced by a goat at only one time, a Westerner
may not know at that time whether she is erroneously applying the
word yang (if he has the impression that it is equivalent to sheep)
or correctly applying the word that refers to both sheep and goats.
Until some day he knows more about her background beliefs through
a triangulated interaction as described by Davidson, he could be sure
that her concept yang is not the same as his concept sheep. It is
unintelligible that without the concept of applicability of names we
can have an act of classication. The job of classifying names can-
not be done merely at the level of names, especially not at the so-
called pre-logical level, which is separated from or independent of
the context of sentence or proposition. Graham is not only neglect-
ing Davidsons holism, but also arguing for his mystical version of
meaning atomism in a self-refuting way.
According to Davidson, Because of the fact that beliefs are indi-
viduated and identied by their relation to other beliefs, one must
have a large number of beliefs if one is to have any. Beliefs support
one another, and give each other content. Beliefs also have logical
relations to one another. As a result, unless ones beliefs are roughly
consistent with each other, there is no identifying the contents of
beliefs. A degree of rationality or consistency is therefore a condi-
tion for having beliefs.
In other words, to identify the content of
a belief, a large number of other beliefs with a high degree of con-
sistency are necessary: the principle of charity with the holistic char-
acteristic. This point is not only true for a single belief, but also true
for a single word. Davidson says, [W]ords, like thoughts, have a
familiar meaning, a propositional content, only if they occur in a
Donald Davidson, 1998, p. 7.
Donald Davidson, 2002, p. 124.
r.\irsoxs 127
rich context, for such a context is required to give the words or
thoughts a location and a meaningful function.
For example,
Grahams binary oppositions of day and night for a child cannot
be understood as no other content than the oppositions themselves,
unless the child could know something so abstract as the formal rela-
tionship between A and ~A at the very beginning of learning. To
understand the word night requires someone to know that there is
no sunlight when the word is used in a context related to what he
or she sees. He or she is also required to have the belief that some-
times he or she can see the moon and stars at night. These rele-
vant knowledge or beliefs about the general features of the event (or
object) are the criteria for people to apply their concepts correctly,
and thus to identify the content of the word. Davidson has an exam-
ple with a similar explanation of the holistic character of having a
single concept. He says that we would deny that someone had the
concept of a man who did not know something about what distin-
guishes a man from a woman, who did not know that fathers are
men, that every man has a father and a mother, and that normal
adults have thoughts.
So, generally speaking, according to Davidson,
the meanings of the words that refer to these features, and the con-
tents of the concepts the words express, depend as much on the nat-
ural history of how the words and concepts were acquired as was
the case for porcupine and echidna. There are no words, or con-
cepts tied to words, that are not understood and interpreted, directly
or indirectly, in terms of causal relations between people and the
world (and, of course, the relations among words and other words,
concepts and other concepts).
He also mentions that, in the sim-
plest cases words and thoughts refer to what causes them, it is clear
that it cannot happen that most of our plainest beliefs about what
exists in the world are false. The reason is that we do not rst form
concepts and then discover what they apply to; rather, in the basic
cases, the application determines the content of the concept.
put in another way, [w]e can give the meaning of any sentence (or
word) only by giving the meaning of every sentence (and word) in
the language.
Donald Davidson, 2002, p. 127.
Donald Davidson, 1998, p. 2.
Donald Davidson, 2002, pp. 5051.
Donald Davidson, op. cit., pp. 196197.
Donald Davidson, 1984, p. 22.
128 cn.r+rn ri\r
Graham agrees with Chad Hansen that classical Chinese nouns
in general are closer to English mass than to English count nouns,
and thus accepts Hansens argument that Western thought is pre-
disposed by number termination to conceive the world as an aggre-
gate of distinct objects while Chinese by the mass noun to conceive
it as a whole variously divisible into parts. He also declares that
Chinese thought being conditioned to divide down rather than add
up is in any case suggested by other features of the language and
that Chinese thinking is in terms of process rather than of static
entitiesindividual objects. However, some serious diculties of
Hansens explanations based on his mass noun hypothesis and his
misinterpretations of the Chinese texts seem to be totally unknown
to Graham. Among other criticisms, I have argued about Hansens
issue elsewhere that, based on this hypothesis, his interpretations and
explanations of the problem in Chinese texts are basically focused
on very few paragraphs mainly in Mohist Canon ( ) and Gong-Sun-
Long-Zi ( ). But unfortunately, even limited to these para-
graphs, some terms used in Mohist Canon obviously referring to
individual object or abstract entity seem to be totally ignored by
Hansens treatment. More unfortunately, even though he claims to
use Davidsons principle of charity to interpret Gong-Sun-Long-Zi, he
has to interpret Gong-Sun Long as committing inconsistency, par-
ticularly in the case that he interprets Gong-Sun Long as using the
term fei-ma ( ) in two dierent senses: one meaning as English
not (identical with) horse and the other as non-horse (in the con-
text of mentioning cow-horse collection, it means cow).
Lets put
aside the problem whether ancient Chinese has count terms or not
(it is clear that there are a lot of evidences provided by Chinese lin-
guists to indicate that count terms do exist in ancient China), there
is still a question of how the ancient Chinese people could have a
language which lacks count terms to express individual objects. Is it
possible that there is a language used by them whose words can
only be used to refer to mass stu instead of individual object? I
dont think so. I dont think that there is a language without expres-
sions referring to dierent individual objects. Claiming that Chinese
thinking is in terms of process rather than of static entities is not
only distorting Chinese thinking, but also misleading the real issue
Yiu-ming Fung, 2000, ch. 8.
r.\irsoxs 129
in Chinese texts by such a groundless hypothesis. Even if we accept,
for the sake of argument, the hypothesis, it is not necessary for us
to accept Grahams (or Hansens) pseudo (or real) linguistic deter-
minism of thinking and ontology. Graham seems to be totally ignor-
ing Davidsons view on convention (and on the prior theory and the
passing theory). According to Davidsons view, what interpreter and
speaker share, to the extent that communication succeeds, is not
learned and so is not a language governed by rules or conventions
known to speaker and interpreter in advance; but what the speaker
and interpreter know in advance is not (necessarily) shared, and so
is not a language governed by shared rules or conventions. One of
the reasons is that a suciently explicit framework could be dis-
credited by a single malapropism. There is some evidence of a more
impressive sort that internal grammars do dier among speakers of
the same language. He concludes that [w]e must give up the idea
of a clearly dened shared structure which language-users acquire
and then apply to cases. Because we have discovered no learn-
able common core of consistent behaviour, no shared grammar or
rules, no portable interpreting machine set to grind out the mean-
ing of an arbitrary utterance. We may say that linguistic ability is
the ability to converge on a passing theory from time to time.
It is quite strange that Graham recognizes Wittgenstein, Ryle, and
Derrida as his heroes in ghting for his correlative thinking covered
by analytic thinking. Nevertheless, it is evident that what these impor-
tant gures of contemporary philosophy have done is not struggle
and search for something at the bottom as Grahams mentalist and
mystical root. Wittgensteins or Ryles disenchantment work is to
try to explain away any mystical element which is supposedly believed
by the mentalists as other than something identied in a social con-
text; and one of the aims of Derridas deconstruction is to subvert
the binarism of Structuralism by demonstrating the instability of
binary oppositions. It is clear that these heroes cannot be under-
stood as supporting Grahams idea; on the contrary, they clean up
his bottom.
Let us imagine that a person, who is retarded and fails to recog-
nize the words sheep and goat, invents by himself or herself a
word shoat (not meaning a young pig just after weaning) to refer
Donald Davidson, 1986, pp. 444446.
130 cn.r+rn ri\r
to either sheep or goats based on his or her perception of some of
their shared characteristics. With respect to his or her new word, I
dont think people using normal English as their home language can-
not understand and interpret what he or she says about shoat; and
I also dont think people using ancient Chinese as their home lan-
guage cannot understand yang as the equivalent of shoat. It is
unnecessary to treat the triple sides as having dierent conceptual
schemes. They share a lot even though there are dierences in vocab-
ulary and grammar among dierent cultures or at home. Semantic
and syntactic exibility and expandability are not only possible for
interpretation at home, but also possible for translation of words and
sentences in a foreign language. As indicated by Davidson, Dierent
speakers have dierent stocks of proper names, dierent vocabular-
ies, and attach somewhat dierent meanings to words. In some cases
this reduces the level of mutual understanding; but not necessarily,
for as interpreters we are very good at arriving at a correct inter-
pretation of words we have not heard before, or of words we have
not heard before with meanings a speaker is giving them.
As to
a foreign language, for example, we know that, before silk or china
transported from China to Europe several thousand years ago, there
was no equivalent word for each of these things in English. But it
is not reasonable to say that the word china was not translatable
between Chinese and English at that time but is translatable now.
Literally speaking, there was no exact word in English to translate
the Chinese word of china at that time; but in principle, it was not
untranslatable. Even though the English word dragon in a Westerners
mind may have dierent mental image, emotional color, or opinion
from the Chinese word long ( ) in a Chinese mind, if they both
use their words in most of the propositional contexts with the same
truth condition and with the same content-xing cause, we would
have no ground to say that most of the sentences about long in
Chinese and dragon in English are not intertranslatable. As indi-
cated by Davidson, there is no need to have a word-by-word trans-
His criticism of Thomas Kuhns idea is based on the fact
of reason, if we could use the Kantian description in this context,
that incommensurability, if making sense, presupposes a co-ordinate
to identify a dierence; and the dierence can be explained and
Donald Davidson, 2001, p. 277.
Donald Davidson, op. cit., p. 72.
r.\irsoxs 131
described using the equipment of a single language.
It means that
Davidsons notion of translation does not mean literal translation
but interpretative translation which allows explanation and descrip-
tion in addition to semantic and syntactic expansions. In the past
the English words mutton and sheep dened one another by con-
trast of their being cooked vs. on the hoof. French word mouton
does not have this contrast; so, we cannot nd its English equiva-
lent in translation. However, when subtracting mutton from English,
it would lead sheep to expand its extension into something like
French mouton. I do not think there is any person who would be
so unthinking as to claim that the literal translation among dierent
languages is always workable without semantic and syntactic expan-
sions. If there are people in this list, Davidson is denitely not.
2. A Methodological Reection on the Principle of Charity
If I am right in the above discussion, Grahams Sinologists criticism
of Davidsons principle of charity together with his view on the very
idea of a conceptual scheme can be judged as groundless and self-
defeating. I believe Davidsons principle of charity can survive in its
application to the Chinese or Sinologist context. However, there are
quite a few criticisms of this principle which are not based on con-
crete contexts in examining its applicability, but are focused on the
controversial issue of its methodological character. Most of the crit-
icisms in this regard are targeted on its seemingly Kantian nature
of transcendental argument or transcendental deduction (hereafter,
simplied as TA and TD respectively). However, Davidson seems
to be reluctant to reject this label in the beginning and unwilling to
accept it later. In his In Defense of Convention T, Davidson writes:
Tarski is right, I think, in proposing that we think of natural lan-
guages as essentially intertranslatable (although I dont see why this
requires word-by-word translation). The proposal idealizes the exibility
and expandability of natural languages, but can be justied by a
transcendental argument (which I will not give here).
In a foot-
note Davidson refers the reader to his articles On the Very Idea
of a Conceptual Scheme and The Method of Truth in Metaphysics
Donald Davidson, 2001, p. 184.
Donald Davidson, op. cit., p. 72.
132 cn.r+rn ri\r
for this TA. But later when responding to Andrew Cutrofellos crit-
icism, he replies that I dont know if my arguments for the prin-
ciple of charity are transcendental or not. Andrew Cutrofello does
not quote me as saying so. In any case, the arguments he says are
mine are not transcendental, good, or mine.
In response to A.C.
Genovas interpretation of his idea, he mentions that [p]eople sug-
gested that what I had hit on was a transcendental argument, and
I didnt reject the idea. But was it? and gives a more concrete
explanation that [i]f you accept the steps that lead to my version
of externalism, what Genova calls semantic realism, then you can-
not, I think, be a skeptic about the existence of an external world
much like the one we all believe we share, nor about the existence
of other people with minds like ours. But the considerations in favor
of semantic realism seem to depend in part not on purely a priori
considerations but rather on a view of the way people are.
in response to Barry Strouds discussion, he also stresses that [w]hat
is not trivial is to show that we know enough about the world to
be able to say that it is pretty much as we think it is.
On another
occasion, Davidson mentions that the principle of charity in inter-
pretation is not a policy: we might do better to think of it as a
way of expressing the fact [my italic] that creatures with thoughts,
values, and speech must be rational creatures, are necessarily inhab-
itants of the same objective world as ourselves, and necessarily share
their leading values with us.
Although Davidson agrees that whether
a creature subscribes to the basic principles of rationality, includ-
ing the principle of continence, the basic principles of logic, and the
principle of total evidence for inductive reasoning, is not an empir-
ical question,
it seems he is only referring to the coherence part
of the principle of charity; as to the correspondence part, it is obvi-
ous that the question is of a factual nature. It seems he is express-
ing the same point when he replies to Thomas Nagel in emphasizing
that [t]he conclusion [of the reasoning of charity] that I know that
the world, both in general and in many particular ways, is as I think
Lewis H. Hahn, ed., 1999, p. 342.
Donald Davidson, Reply to A.C. Genova, in Lewis H. Hahn, ed., 1999, pp.
192, 194.
Donald Davidson, Reply to Barry Stroud, in Lewis H. Hahn, ed., 1999,
p. 163.
Donald Davidson, 1984, p. 18.
Donald Davidson, 1985, p. 352.
r.\irsoxs 133
it is, depends on the fact [my italic] that I have just the beliefs I
In other words, the argument for the impossibility of massive
errors of our beliefs or for the necessity of massive truth shared by
the speaker and the interpreter is based in part on empirical evi-
dence; in this sense we cannot label Davidsons argument for the
principle of charity transcendental.
With respect to the arguments for the condition of attribution of
thought and for the condition of language-hood, it is obvious that
Davidson uses many similar key terms or phrases which are the indi-
cators of using TA or TD in a Kantian sense or a popular sense
since P.F. Strawsons use. In Davidsons third book which deals with
three kinds of knowledge, we can easily nd a pattern of terms or
phrases used in his sentences, such as X is possible only if Y, X
is impossible unless (there is a presumption that) Y, X requires
Y, X is essential to the possibility of Y, The very possibility of
X demands Y, X is intelligible only on the supposition that Y,
X takes its content from a background of Y, In order to have
X, we must entertain Y, X is a necessary condition of Y, and
so on. These expressions are undoubtedly also used in the discourse
of TA or TD. So, I think the popular perception of transcenden-
tal attributed to Davidsons arguments is not groundless. That
Davidson lacks detailed explanation may be in part responsible for
the resultant misunderstanding.
In order to give a supplementary explanation for this problem, I
will make a distinction based on Kants two kinds of TD or TA in
the following. We know that in Kants three Critiques there are at
least two kinds of TD or TA. In Kants rst Critique there is a par-
adigmatic use of the TD, i.e. the transcendental deduction of the
categories. It starts from a question How is experience (or empiri-
cal knowledge) possible? and answers with the following chain of
expressions: From the premises that there is experience and that
the categories are necessary conditions of its possibility, it proves or
legitimizes the objective validity of the categories. In his second
Critique Kant seems to use the same kind of TD. Its question is:
How is (autonomy) morality (or moral judgment and action) possi-
ble? Its argument runs like this: From the premises that there is
Donald Davidson, Reply to Thomas Nagel, in Lewis H. Hahn, ed., 1999,
p. 209.
134 cn.r+rn ri\r
(autonomy) morality and that its possibility is based on the postu-
late of the freedom of will, it proves or legitimizes the objective valid-
ity (or practical necessity) of the freedom of will. In this regard, the
rst premise in each deduction seems to Kant to be empirically true,
but the second premise in each deduction obviously includes some
transcendental concept or transcendental entity, i.e. a priori category
or the freedom of will, and thus it cannot be considered as empir-
ically true. To assert that having some transcendental concept or
transcendental entity is a necessary condition of, or is presupposed
by, having an empirical fact is clearly not an empirical claim, but
some kind of a priori claim or stipulation. In this sense we are legit-
imated to say that this kind of argument is a TD or TA in its proper
sense or paradigmatic use. Another sense of TD or TA used in the
rst Critique is about the objective validity of (empirical) objects. As
for the same question as above, How is experience possible? Kants
reply is this: From the premises that there is experience and that
(physical) objects are necessary conditions of its possibility, it follows
that there are (physical) objects. We can see in this case that the
second premise does not have any transcendental concept or tran-
scendental entity; it seems not to be an a priori stipulation but an
empirical claim of a necessary condition, just like the claim that
having food is essential to being healthy, good teaching requires
good learning, stopping smoking is a necessary condition of pre-
venting lung cancer, and having an event of bell ringing presupposes
having a bell. All these are based on some empirical fact. In com-
parison with the rst kind of TA, a paradigmatic use of TA, this
second kind of argument has also been classied as TA by many
philosophers but is obviously lacking a transcendental sense. The
other dierence between these two kinds of TA is related to the
problem of objective validity. With respect to the rst kind of TA,
if there are dierent TAs based on dierent a priori stipulations for
the possibility of the same empirical fact, what can be justied or
legitimized would be dierent transcendental concepts or transcen-
dental entities; and the so-called objective validity of the target con-
clusion would be relativized and thus the argument could not be
sound. On the other hand, each example of the second kind of TA
is an empirical claim, each argument can be valid or invalid, sound
or unsound, as long as we can have the relevant empirical evidence
to conrm or disconrm (or to verify or falsify) each sentence in the
argument. This kind of TA may have Strouds problem with
r.\irsoxs 135
superuous or redundant in relation to verication;
but the rst
kind is denitely irrelevant to the problem. Although these two kinds
of TA are very dierent, I think it is unfortunate that some people
like to put these two together under the same label. To make them
devoiced may be good for our understanding of many related but
distinct arguments in the literature, and especially for our under-
standing of Davidsons principle of charity.
The impression of the transcendental sense given to people by
Davidsons principle of charity may be due to its expressing in a
very general way and its seemingly a priori characteristic. However,
if we pay ample attention to some of the concrete examples which
are usually used by Davidson to illustrate the applicability of the
principle, this impression may be dierent. For example, a dierent
impression would emerge if we read carefully the paragraph about
the concept of a cat I quoted from Davidsons article in the previ-
ous section.
In this particular case, but not in the general form of
the principle, it is obvious that this claim for the condition of hav-
ing a concept and a belief, or attributing a concept or a belief, is
supported by empirical evidence; this is not an a priori stipulation.
If we use Grahams seemingly counter-example as our example, it
also conrms with empirical evidence that massive true beliefs shared
by Chinese and Westerners are essential for interpreting and under-
standing the dierences between their languages. Someone who fails
to interpret or to translate the Chinese word yang into the English
word sheep at the beginning would eventually discover that the
scope or extension of the former is larger than that of the latter
through a test of Davidsons triangulation. In this case, even though
they classify things dierently and each side may have dierent men-
tal images, emotive colors, and opinions attached to their words,
they do have a lot to share. They both recognize that there are two
sorts of objects, though English, unlike Chinese, does not put them
together as belonging to the same kind. They both know that the
objects either referred to by the Chinese yang or referred to by the
English sheep or goat are not plants but animals, that these ani-
mals are physically visible, freely moving, do not eat other animals,
and so on. The list of shared beliefs and thoughts is too long to
Barry Stroud, 1968, p. 256.
See the quotation in footnote 15.
136 cn.r+rn ri\r
enumerate; but it forms an empirical background for each to inter-
pret and to understand the others language. Based on this and other
examples, I think Grahams argument against Davidsons view is
actually self-refuting; Grahams explanation of the dierences between
the two languages, in order to make sense, is also necessarily based
on some kind of background beliefs shared by both sides.
If a teacher of biology wants to persuade his or her students to
stop smoking, he or she may ask the students a question like this:
How is preventing lung cancer possible? Some smart students would
give an answer from a health magazine that It is the case that
many people have no lung cancer, and (we know from statistical evi-
dence) that a necessary condition of preventing lung cancer is to
stop smoking; therefore, stop smoking. This Modus Ponens argu-
ment is clearly based on empirical evidence. Let us take another
example to illustrate the same point. We can ask an experienced
teacher, like Davidson, a question How is it possible to be a good
teacher? I think he would answer like this: There are many good
teachers in the world, and (we know from experience that) to be a
good teacher requires someone to be a good learner; therefore, to
be a good learner. So, based on this valid argument, we should try
our best to be a good learner if we want to be a good teacher. This
argument starts from a question of possibility, its premises do not
involve any transcendental concept or transcendental entity, and its
assertion of the necessary condition is based on empirical evidence.
Hence, it is not an a priori claim. The general form of Davidsons
principle of charity seems to suggest an a priori claim, but its par-
ticular examples are not. I think this principle is in accord with our
intuition of the ordinary use of language (speaking or interpreting)
and, more importantly, is generalized from a lot of particular cases
with empirical evidence as its inductive base. This base is very sta-
ble and seems unshakable, because we have not yet found any con-
crete and obvious counter-example to the principle. In this sense, I
conclude that the principle of charity, Maximizing true beliefs is a
condition of having thought, is based on a huge amount of empir-
ical evidence, as is the principle of nutrition, Maximizing nutriment
is a condition of having energy.
r.\irsoxs 137
3. Zen Buddhism Interpreted as Deconstructive Skepticism
Davidson, like other contemporary philosophers, wants to kick out
skepticism. What Davidson can do in this respect, I think, is to kick
out some version of skepticism in a constructive or explicit form, but
not a deconstructive or implicit form of skepticism. As demonstrated
by Putnams model-theoretical argument, metaphysical realism and
skepticism are two sides of the same coin.
In this sense we may
say that metaphysical realism or Platonism is a constructive form of
skepticism. This kind of skepticism, if expressed explicitly, is not only
claiming that our existing empirical knowledge of the world is an
illusion or delusion, but also establishing its own position and assert-
ing a view of reality. Since it is not only negatively rejecting what
we have known as illusive or delusive appearance, but also positively
arming an ultimate truth of reality, there is trouble for this kind
of skepticism to justify itself. It has trouble both because the burden
of proof is now on its shoulder and because its non-empirical stance
is obviously no better than our empirical stance even though it would
try to legitimize or justify itself on its own stance. In this regard,
Davidsons arguments based on triangulation are obviously more con-
vincing than the transcendental or metaphysical arguments of skep-
ticism such as Platonic realism; and in this sense, it is fair to say
that Davidson can kick out this kind of constructive skepticism.
However, Davidson believes that to show we know enough about
the world to be able to say that it is pretty much as we think it is
is to show that it [i.e. skepticism] is false, though he agrees that
we cannot prove it false in particular cases.
Or more moderately
speaking, I set out not to refute the skeptic, but to give a sketch
of what I think to be a correct account of the foundations of lin-
guistic communication and its implications for truth, belief, and
knowledge. If one grants the correctness of this account, one can
tell the skeptic to get lost.
It seems he is claiming that he can
kick out skepticism in general. But I do not think this is true; I think
the case is reversed. There are many brands of skepticism, con-
structive or deconstructive, metaphysical or mystical, or in particu-
lar, the thesis of Platos cave or Descartes demon. In general, it is
Hilary Putnam, 1984, ch. 1.
Donald Davidson, Reply to Stroud, in Lewis H. Hahn ed., 1999, p. 163.
Donald Davidson, 2002, p. 157.
138 cn.r+rn ri\r
not easy to kick out all of them; but when facing some particular
cases of constructive skepticism, it is quite easy to compare its power
of justication with that from our empirical stance and thus not
dicult to kick it out.
Although all brands of skepticism cannot justify themselves, some
of them cannot be falsied by empirical evidence. This is just like
the case that although all brands of theism cannot justify there is
God, some of them cannot be falsied by empirical evidence. Let
us imagine that, if a theist claims that in reality God is in the world
though in appearance you do not know that it is the case, what
could an anti-theist say? She could say that Please tell me how to
identify or locate God in this world? Then, the burden of proof
would be put on the theists shoulder, because what he claims is
more than a negative statement that the world is not what it appears,
a mere logical possibility, but he also has a positive claim about
something (i.e. God) which does not appear in the world but is really
in the world. Just like the case of constructive skepticism, this can
be easily judged as false unless the theist can identify or locate the
non-empirical entity in the empirical world in a non-mysterious way.
Nevertheless, if some slippery theist claims that God does exist but
He and His Action are ineable and unintelligible in a rational
way, it seems that we cannot directly prove the claim is false. When
you try to prove its falsity, the theist would reply that Hey, it can-
not be thought in a rational way, so, it cannot be falsied in a ratio-
nal way either. My point is that, when facing this deconstructive,
implicit or hidden kind of skepticism, we cannot kick it out directly,
because it rejects the whole idea of rationality, either used for prov-
ing or for disproving the issue: it refuses to prove itself in this way
and also refuses any disproof in the same way. It is totally out of
reason. But, is it irrational?
Davidson is right in saying that we could not understand some-
one whom we were forced to treat as departing radically and pre-
dominantly from all such [rational] norms. This would not be an
example of irrationality, or of an alien set of standards: it would be
an absence of rationality, something that could not be reckoned as
He is also right in making his conclusion that If what
we share provides a common standard of truth and objectivity,
Donald Davidson, 1993, p. 297.
r.\irsoxs 139
dierence of opinion makes sense. But relativism about standards
requires what there cannot be, a position beyond all standard.
the deconstructive skepticism or slippery relativism mentioned above,
according to Davidson, can be understood as not irrational, but non-
rational (except its criticism of the rational view of the world). This
means that it is not about thought, therefore it cannot be either true
or false. Irrational thinking is false because it is qualied to be false;
but non-rational thinking (if it can be called thinking) is not false
because it is not qualied to be false. From Davidsons point of view,
non-rational thinking is just not making sense. Here, I totally agree
with Davidson on this point. However, I do not think the story ends
here. Because the deconstructive skepticism would reply that based
on your rational standards, you have your own right to judge that
my thinking is not making sense; but based on my non-rational or
super-rational standards, your knowledge of the world is illusive or
delusive. My thinking not making sense to you is tantamount to my
thinking not understood by you, because it cannot be understood by
rational thinking. Why do I have to give up my own standards and
then to share your rational standards and thoughts? Could I ask you
to do the reverse? Your request of my giving up my standards and
my request of your giving up your standards seem to be of the same
weight of legitimacy. At this point, I do not think Davidson and
any other can kick out this deconstructive skepticism without more
Davidson seems to have smelt this kind of skepticism; so he men-
tions that there is the idea that any language distorts reality, which
implies that it is only wordlessly if at all that the mind comes to
grips with things as they really are . . . Yet if the mind can grapple
without distortion with the real, the mind itself must be without cat-
egories and concepts. This featureless self is familiar from theories
in quite dierent parts of the philosophical landscape.
The thesis
based on the idea of the featureless self in general, or the idea of
Zen (or Chan ) Buddhisms no-self in particular (which will be
discussed later), is denitely rejected by Davidson, because he thinks
that the mind is divorced from the traits that constitute it. However,
the reason of Davidsons rejection is still based on rational standards
Donald Davidson, 1993, p. 307.
Donald Davidson, 2001, p. 185.
140 cn.r+rn ri\r
which are not welcome to this kind of skepticism; so, I do not think
it can persuade the skepticism to get lost.
In Asian philosophy, especially in Chinese philosophy, the idea of
featureless self , conceptless subject, contentless consciousness,
absolute mind, or no-self was very popular a few thousand years
ago. Ancient Daoism, Zen Buddhism, and New Confucianism (includ-
ing Song-Ming and contemporary Confucianism) are the main repre-
sentatives with this characteristic in the eld. In the following, our
discussion will mainly focus on Zen Buddhism, with special refer-
ence to D.T. Suzukis interpretation. We know that Suzuki was very
successful in promoting Zen Buddhism in the West. In order to
attract Western peoples attention by some salient dierences, some-
times Suzuki exaggerates the contrast between Western and Asian
thinking. One of the major dierences emphasized by him is the
contrast between rationality and irrationality, or logical and illogical
thinking. For example, Suzuki sometimes says that Zen is the most
irrational, inconceivable thing in the world, that it dees all concept-
making, and that the essence of Zen is satori ( ), the experi-
ence of sudden enlightenment, which is irrational, inexplicable, and
In his popular book, entitled Living by Zen, he
stresses that If we are to judge Zen from our common-sense view
of things, we shall nd the ground sinking away under our feet. Our
so-called rationalistic way of thinking has apparently no use in eval-
uating the truth or untruth of Zen. It is altogether beyond the ken
of human understanding. All that we can therefore state about Zen
is that its uniqueness lies in its irrationality or its passing beyond our
logical comprehension.
In response to the Chinese historian Hu
Shihs ( ) criticism, he says, Zen is not explainable by mere
intellectual analysis. As long as the intellect is concerned with words
and ideas, it can never reach Zen.
Therefore, to know Zen one must
give up his or her rational thinking and dualistic logic, and then he
or she could be enlightened with praj-intuition ( ), an
unknowable knowledge. Why do you have to give up rational think-
ing and dualistic logic? It is because people without Zen enlighten-
ment are living in the world of samsra ( ) with the suerings
issuing from dualistic thinking. If you want to be emancipated from
D.T. Suzuki, 1996, pp. 13, 103.
D.T. Suzuki, 1994, p. 20.
D.T. Suzuki, 1953, p. 26.
r.\irsoxs 141
these suerings and to enter Zens non-dualistic world, you have to
go beyond rational thinking. To be free from the dualistic cage and
enter this beautiful world, one must know nothing; because to fall
into the dualistic abyss, one is forced to know something conceptu-
alized. Zen or the insight of sunyata ( emptiness) is nothingness,
because it is the undierentiated totality, there is nothing in it which
can be conceptualized. Suzuki thinks that the dualist view of real-
ity has been a great stumbling block to our right understanding of
spiritual truth,
and thus Zen is decidedly not a system founded
upon logic and analysis. If anything, it is the antipode to logic, by
which I mean the dualistic mode of thinking.
He even condemns:
According to the philosophy of Zen, we are too much of a slave
to the conventional way of thinking, which is dualistic through and
through. No interpenetration is allowed, there takes place no fus-
ing of opposites in our everyday logic. Zen, however, upsets this
scheme of thought and substitutes a new one in which there exists
no logic, no dualistic arrangement of ideas.
In order to deconstruct dualistic logic, Suzuki sometimes stresses
the signicance or necessity of Zen masters using incoherent or para-
doxical statements to express their insight. He explains the reason
why they [i.e. the masters] make those apparently incoherent state-
ments. Their inclination is to set the minds of their disciples or of
scholars free from being oppressed by any xed opinion or preju-
dices or so-called logical interpretations.
More theoretically speak-
ing, Paradoxical statements are . . . characteristic of praj-intuition.
As it transcends vijnana ( ) or logic it does not mind contradicting
itself; it knows that a contradiction is the outcome of dierentiation,
which is the work of vijnana.
One of the paradoxical statements
frequently used by Suzuki is the following example: We generally
reason: A is A because A is A; or A is A, therefore, A is
A. Zen agrees or accepts this way of reasoning, but Zen has its
own way which is ordinarily not at all acceptable. Zen would say:
A is A because A is not A; or A is not A, therefore, A is
It seems that the way of Zen that Suzuki describes is the way
D.T. Suzuki, 1994, p. 28.
D.T. Suzuki, 1991, p. 38.
D.T. Suzuki, 1996, p. 112.
D.T. Suzuki, [1991], pp. 7879.
D.T. Suzuki, 1955, pp. 9495, also in Charles A. Moore, ed., 1951, p. 24.
D.T. Suzuki, 1953, p. 37.
142 cn.r+rn ri\r
to subvert, generally, the duality of A and ~A; and specically,
the dichotomy of subject and object. He believes that in praj this
dichotomy no longer exists, because, [ p]raj is not concerned with
nite objects as such; it is the totality of things becoming conscious
of itself as such. And this totality is not at all limited. An innite
totality is beyond our ordinary human comprehension.
It means
that not only the extension but also the content of Zen wisdom is
beyond ordinary human comprehension. So he concludes, Satori
(emptiness) may be dened as an intuitive looking into the nature
of things in contradistinction to the analytical or logical understanding
of it.
Although Suzuki puts too much emphasis on the irrational and
illogical character of Zen, it is basically a strategy. In reality, he is
not rejecting rational thinking and logic for its own sake; in order
to demonstrate their irrelevance to or distortion of the ultimate real-
ity, he has to use this strategy to help people to go beyond them.
In other words, Zen is a realm of pure land inhabited by people
with pure consciousness without any concept and logic. For Suzuki,
to ght against rational thinking and logic is not to justify their fal-
sity, but to throw them away altogether in order to enter Zen enlight-
enment. In this sense we can say that Suzukis Zen is not irrational
but non-rational. To understand irrational thinking, as demonstrated
by Davidson, we have to put it into the framework of rational think-
ing. I think Suzuki would agree with this point. However, to under-
stand non-rational or super-rational thinking, for Suzuki, we have to
throw away rational thinking altogether, there is no place for it to
stand. In other words, giving up rationality is essential to attaining
Zen enlightenment. Based on this point, I argue that, if Zen Buddhisms
view on the secular world can be interpreted as a form of decon-
structive skepticism, its non-rational web can be considered as very
defensive to any attack from the rational army.
4. Locating Problem and Landing Problem
Zen as a form of deconstructive skepticism is both like and unlike
the relativism dened by Davidson and others. It is like relativism
D.T. Suzuki, 1960, p. 57.
D.T. Suzuki, 1996, p. 84; and 1991, p. 88.
r.\irsoxs 143
in the sense that it takes an unorganized reality for granted; it is
unlike it in the sense that it throws away the so-called conventional
scheme and denes its understanding or enlightenment as concept-
less and speechless. As interpreted by Suzuki, Zen masters can grasp
their ultimate reality through a special and private access, while ordi-
nary people still in the ignorant state distort Zens truth and mis-
understand reality as appearance. This picture of the two kinds of
truth (Zens truth and ordinary truth) or of the two levels of world
(Zens world and ordinary world) may be understood as a special
sort of relativism; but it seems not appropriate to apply Davidsons
argument against the very idea of a conceptual scheme in this con-
text, because, as indicated by Suzuki, Zen has no conceptual scheme.
Suzuki would probably address the question in this way: Zen mas-
ters would reject applying Davidsons identifying or locating crite-
rion by saying that there is no conceptual content, clear or unclear,
in my state of enlightenment, and one can reach this state only if
he or she gives up rational thinking and thus is free from the con-
ceptual cage. We have no language at all when enlightened. We are
not to use language to organize or to x anything. For Zen, there
is nothing to be organized or to be xed. What we have is neither
an object nor a subject, because the dualistic contrast between sub-
ject and object has been merged into an undistinguishing unity, an
absolute subjectivity. If we have to describe what it is, it is the sub-
ject without opposite to object, the subject in itself, or more appro-
priately, no-self or emptiness (sunyata). Since Suzuki stresses that only
through each individuals private access can one enter Zens world,
we may consider Zens world as solipsistic and, according to Davidson,
judge it as having no size or as not a world at all.
But Suzuki
would not agree with this accusation, because he believes that one
who enters Zens world with enlightenment is absolutely convinced
of its universality in spite of its privacy.
It is private because it
cannot be grasped by concepts and can be entertained or enjoyed
only in each individuals mind. It is not solipsistic because it is enjoy-
able by all people though not enjoyed by those who are still in the
realm of illusion. Furthermore, even using Davidsons idea of the
compartmentalization of the mind would be unhelpful, because
Donald Davidson, 2002, p. 119.
D.T. Suzuki, 1953, p. 33.
144 cn.r+rn ri\r
Zens problem is not the division of the mind, but the deconstruc-
tion of the mind. This deconstructive or hidden skepticism implies
a special sort of relativism which is very slippery and is not easy to
get rid of.
Based on the principle of charity, of course, Davidson would ask
Suzuki to identify or locate Zens thinking in a human language. As
mentioned by Davidson, We do not understand the idea of such a
really foreign scheme. We know what states of mind are like, and
how they are correctly identied; they are just those states whose
contents can be discovered in well-known ways. If other people or
creatures are in states not discoverable by these methods, it cannot
be because our methods fail us, but because those states are not cor-
rectly called states of mindthey are not beliefs, desires, wishes, or
intentions. The meaninglessness of the idea of a conceptual scheme
forever beyond our grasp is due not to our inability to understand
such a scheme, nor to our other human limitations; it is due sim-
ply to what we mean by a system of concepts.
I think Suzukis
Zen or other skepticism painted with a similar color of mysticism
would respond to Davidsons request of identication of their inner
state in this way: Yes, they are not usual or ordinary mental states.
You have to be trained or to practice by yourself (in a special reli-
gious or moral way of practice or gong-fu ) in order to trans-
form your self from illusion to enlightenment. And Davidson might
answer in this way: Oh, thats all I can say of your so-called inner
states. If they cannot be discovered and identied by these meth-
ods, I dont know what they are. I only know that they are not what
we call mental states in the common usage, period. Nevertheless,
I do not think Suzuki and his comrades are satised with Davidsons
answer on this point, and there would be no further dialogue between
Davidson and them to follow. However, I prefer a strategy of putting
aside the locating problem and ask them another question: We
agree, for arguments sake, that you can be enlightened without ratio-
nal thinking, but, how can you transform your self and know that
you are crossing the gap after you have been trained or practiced
in your special way and then arrived at the realm of the so-called
inner experience of enlightenment? How can you know, without
any ordinary concept, the very secular concepts are all wrong in
Donald Davidson, 2002, p. 40.
r.\irsoxs 145
grasping your ultimate concern? You only say that your inner world
cannot be expressed in discursive language or even in any language
(i.e. ineability thesis), and sometimes say that your transcendental
entity (i.e. absolute subjectivity) is not an ordinary object, such as
table, tree, apple, or an event in the physical world; but that is not
enough for us and even for you to understand what it is. I think
this is not to ask for an answer of the locating problem, but another
Let us suppose there is a game of twin boxes. Both boxes are
similar in the cover, but one of them is empty and the other either
has something inside or has not. We do not know which one is
empty and whether both are empty. However, the master of the
game gives us some hints. He tells us that inside the boxes both
include no ordinary object, such as apple, pencil, or cake, and that
the only dierence between these two boxes is that one of them is
empty and the other one has a non-ordinary object inside. I think
what Suzuki tells us in his game is similar to this box game: Based
on our standard of identifying ordinary object, we are not only unable
to identify the so-called non-ordinary object inside the transcenden-
tal box, we also cannot understand whether there is any dierence
between having this very thing in the transcendental box and hav-
ing nothing in the other ordinary box. It is no use to characterize
it based only on some negative descriptions, such as non-ordinary
object, emptiness or no-self , because these can also be applied to
the case when there is nothing inside the transcendental box (or
applied to the ordinary empty box), the truth condition for both
cases is the same. Suzukis example of tea drinking also indicates a
similar game. He says: You and I sip a cup of tea. That act is
apparently alike to us both, but who can tell what a wide gap there
is subjectively between your drinking and my drinking? In your drink-
ing there may be no Zen, while my [sic] is brim-full of it. The rea-
son for it is: you move in a logical circle and I am out of it.
time the twin boxes seem to have the same kind of object inside,
but the game master stresses that only he who moves beyond a log-
ical circle can know they are dierent. But I do not think Suzuki
can make the distinction in a signicant way. When Suzuki de-
clares that the aim of Zen is the unfolding of a new world hitherto
D.T. Suzuki, 1991, p. 88.
146 cn.r+rn ri\r
unperceived in the confusion of a dualistically trained mind,
is his standpoint? If he declares from the standpoint of the old world,
this declaration is self-defeating; if he declares from the standpoint
of the new world, it would be nonsense. It is self-defeating because
the dualistically trained mind rejects itself in the former case; it is
nonsense because there is no point in contrast in the latter. I do not
think he can explain his standpoint. He sometimes stresses no other
method than that of casting away this intellectual weapon and in all
nakedness plunging into sunyata itself,
because he thinks that, [a]s
long as conceptualization goes on, there will be no discovery of the
real self
and [t]he self escapes from all these meshes of concep-
tualization or objectication.
However, on the other hand, in order
to explain the foundational meaning of sunyata, he on another occa-
sion emphasizes that sunyata is not a negative term but a positive
concept [my italic], and is not arrived at by abstraction or postula-
tion, for it is what makes the existence of anything possible.
is wandering in between the conceptual and the conceptless world.
My conclusion is that not only can we not understand that, but also
the Zen masters are unable to distinguish their own inner experi-
ence from ordinary things, because their Zen wisdom is totally
beyond logical and rational thinking and thus they cannot think of
any dierence between Zen and other things. In other words, they
give up their right to make any distinction and they become the
slave of their non-dualistic thinking.
The problem of locating or identifying is to ask the question of
how to locate the position of an entity under investigation in a ratio-
nal or public space. If we consider the idea of Zen as having this
problem, it would be a question about the location of Zens tran-
scendental self (or no-self ). However, as mentioned above, Suzuki
does not take this problem seriously, because he rejects the ratio-
nal doctor to treat his problem and thinks that Zen has its non-
rational immutability which is free from the disease cured by the
rational doctor. Although Suzukis interpretation of Zen could be
recognized as immune from the locating problem, it still has another
D.T. Suzuki, 1996, p. 84; and 1991, p. 88.
D.T. Suzuki, 1951, p. 5.
D.T. Suzuki, 1954, p. 174.
D.T. Suzuki, op. cit., p. 168.
D.T. Suzuki, 1951, p. 4.
r.\irsoxs 147
problem which may be thought as troubling him. This other prob-
lem, which I call the landing problem, is about bridging the Zens
inner world with the secular outer world, or about landing the
Zens transcendent or transcendental self into empirical life. More
particularly speaking, it is the problem of how to transform from
the illusive life of attachment to the secular world into the enlight-
ened life of freedom from such attachment. As indicated by Suzuki,
what makes Zen unique as it is practiced in Japan is its systematic
training of the mind. Ordinary mysticism has been too erratic a
product and apart from ones ordinary life; this has Zen revolu-
tionized. What was up in the heavens, Zen has brought down to
However, he also believes that [w]hen conceptually under-
stood, the lifting of a nger is one of the most ordinary incidents in
everybodys life. But when it is viewed from the Zen point of view
it vibrates with the divine meaning and creative vitality. This would
appear to be dicult to land on earth. Our question is: how could
one shift from the ordinary view to the Zen view? He mentions that
[t]hose who have only read the foregoing treatment of Zen as illog-
ical, or of Zen as a higher armation [i.e. armation without nega-
tion as antithesis], may conclude that Zen is something unapproachable,
something far apart from our ordinary life, something very alluring
but very elusive; and we cannot blame them for so thinking. Zen
ought, therefore, [to] be presented also from its easy, familiar and
approachable side. Life is the basis of all things; apart from it noth-
ing can stand. With all our philosophy, with all our grand and
enhancing ideas, we cannot escape life from as we live it. Star-gazers
are still walking on the solid earth.
But, how can Zen-gazers land
on the solid earth?
If Zen cannot land on the earth of our ordinary life, it is irrele-
vant to our ordinary life and thus is unable to criticize our ordinary
view of the world as illusion and delusion, or as rational attachment
and distortion of reality. If I am right on this point, I think Zens
situation would be no dierent from that of a speechless dog in its
inability to criticize the illusion and delusion of humans attachment
to the secular world. However, Suzuki thinks that Zen must never
be confused with naturalism or libertinism, which means to follow
D.T. Suzuki, 1991, p. 45.
D.T. Suzuki, op. cit., pp. 8081.
148 cn.r+rn ri\r
ones natural bent without questioning its origin and value. There
is a great dierence between human action and that of the animals,
which are lacking in moral intuition and religious consciousness. The
animals do not know anything about exerting themselves in order
to improve their conditions or to progress in the way to higher
Even though he shows us a demarcation between humans
intentional action and animals natural reaction, this picture is over-
simplied. If humans knowing something about exerting themselves
in order to improve their condition or to progress to higher virtues
is based on rational thinking, it is denitely not the characteristic of
Zen; if it is based on Zen wisdom, Zen wisdom would be the rea-
son of an intentional action and thus could not be separated from
rational thinking. As indicated by Suzuki, all that we can do in
Zen in the way of instruction is to indicate, or to suggest, or to show
the way so that ones attention may be directed towards the goal.
If the transcendental goal is not separated from the empirical world
in the sense that what is done in our ordinary life is intentionally
related to the goal, then the verbal and nonverbal actions done for
Zen can be understood and interpreted in our rational language. If
Zen masters verbal and nonverbal actions are intentional, they can-
not be either totally irrational or totally non-rational. My question
is that, without rational thinking, how can a Zen master show his
way towards the goal of enlightenment? Should he and his disciples
not be conscious of whether it is a right or wrong direction when
they intend to go ahead? It is clear that, as an intentional action,
making eort to attain Zen enlightenment cannot be separated from
rational thinking.
The locating problem seems inescapable for any religious thought
which claims that their world of enlightenment or transcendence is
radically or rigidly dierent from the secular world of common sense.
But some special religious thought, such as Zen Buddhism as inter-
preted by Suzuki, which stresses its practical wisdom as necessarily
supervened on the Aufheben of rational thinking and logic, may
have a good excuse for rejecting the request of locating. However,
the locating problem is only performed as the rst check-point of
rationality, even though putting aside this problem, there is still a
D.T. Suzuki, 1991, pp. 8586.
D.T. Suzuki, op. cit., p. 92.
r.\irsoxs 149
problem of landing, the second check-point of rationality, to which
this kind of skepticism is required to respond. In other words, the
masters may have some excuse to reject or to ignore the request of
identifying or locating its non-conceptual and speechless entity; but
they cannot escape the request of explaining the possibility of bridg-
ing the gap between the two worlds and the feasibility of trans-
forming the ordinary self into the transcendental self (or no-self ). It
is obvious that, without landing the transcendental mystery on empir-
ical life, entertaining the private experience of the mystery would be
non-sense or irrelevant to actual life. If it has to land on actual life,
the Zen master has to explain how his qi-zhi ( natural dis-
position) can be transformed into the ideal state of Zen living, how
he can know, from a view point without rational thinking and logic,
that the ordinary understanding of the world from a view point of
rational thinking and logic is illusive and delusive, and how he can
know what he has attained is not the wrong thing. These are all
questions of landing problem they have to answer.
Suzukis Zen Buddhism shares with New Confucianism and other
schools of Buddhism one of the major characteristics of mysticism
and pantheism, i.e. the ultimate reality is both transcendent and
immanent. Similarly, he also recognizes there are two levels of world
(satoris own world and a world of multitudes), two kinds of expe-
rience (Zen experience and ordinary experience), and two sorts of
truth (Zen truth and rational truth). Furthermore, he also claims that
Zens suchness, experience, or truth is both transcendent of and
immanent in the ordinary world. However, we know that tran-
scendent immanence or immanent transcendence is not a coher-
ent concept,
how can Suzuki bridge these two contrary terms in a
coherent way? It is obvious that Suzuki encounters the same kind
of diculty as most of the mysticists and pantheists confront. Although
Suzuki thinks that these two contrary terms are essentially incom-
, he still strongly believes that [i]n satori what is imma-
nent is transcendent and what is transcendent is immanent.
it is merely a claim without argument. Of course, Suzuki would say
that if we provide argument for the claim, satori would be murdered
See my article on the three dogmas of New Confucianism, Yiu-ming Fung,
2001, pp. 245266.
D.T. Suzuki, 1994, p. 119.
D.T. Suzuki, op. cit., p. 47.
150 cn.r+rn ri\r
by this dualistic or rational thinking and it would not be genuine
satori. But how can Suzuki bridge these two extreme terms and thus
solve the landing problem if no rational thinking is permitted? It
seems that there is no other way to the end except appealing to
satori itself. Actually, Suzukis strategy of persuasion is appealing to
satoris absolute and private authority. He thinks that [t]he gap
between satori and rationality could never be bridged by concept-
making and postulation, but by an absolute negation of the reason
itself, which means an existential leap.
According to Suzuki, an
absolute negation or absolute armation can only be made by
someone who has a satori-eye. It is therefore clear that appealing
to satoris authority can solve the landing problem; so, eventually,
Satori is an existential leap.
Nevertheless, mysticism and pan-
theism also claim the absolute authority of their inner, private,
absolute, transcendental, mystic, and holy experience, and appealing
to private access is obviously not a privilege of Suzukis Zen Buddhism.
I think the landing problem is not easy for him to solve.
Even if the Zen masters are able to escape the locating problem,
they still have to solve their landing problem, otherwise they would
face the following dilemma: On the rst horn, if they really argue
against conceptualization and duality, it would presuppose the con-
ceptualization of its anti-thesis and make a duality between the the-
sis under attack and their own anti-thesis. It means that their anti-thesis
would be self-refuting. On the second horn, if they want to transcend
but not reject conceptualization and duality, their transcendental
journey would be irrelevant to the daily world, a journey which is
not only logically impossible to be identied but also substantially
unrelated to actual life. It means that they cannot make any thesis
or anti-thesis except sleeping on their private bed of mystery.
5. Conclusion
In the previous sections, I have tried to discuss some problems about
Davidsons principle of charity. The rst problem is the question of
whether the principle can survive in the context of comparison
between Chinese and English language and philosophy. In this regard,
D.T. Suzuki, 1994, p. 69.
D.T. Suzuki, op. cit., p. 125.
r.\irsoxs 151
A.C. Grahams criticism has been sorted out as a salient example
of the reaction issued from most Sinologists basic faith. It seems
that one of the major academic interests of most Sinologists is moti-
vated by their expectation of nding some alien characteristics in
Chinese culture and language, or by the goal of so-called Chinese-
ness. Even though there might be something called Chinese-ness,
such as Grahams idea about a distinct root thinking in Chinese
language and philosophy, this Chinese-ness cannot be merely and
basically described and explained as Chinese-ness, i.e. understood
only in a Chinese way of thinking and interpreted only in a Chinese
way of expression. So Grahams criticism is fundamentally self-
defeating; it cannot be understood as a real challenge to Davidsons
principle of charity.
The second problem is the methodological character of the prin-
ciple of charity. Although most of the discussants on the problem
are inclined to identify Davidsons arguments as transcendental, I
have argued that this is not accurate and that this misunderstand-
ing is due to their not taking seriously some of the concrete exam-
ples provided by Davidson and not realizing a crucial distinction
between two kinds of TA in the literature. The general form of
Davidsons principle may give people a perception that it is an a
priori claim; however, there are a lot of concrete examples which
constitute a strong inductive base of empirical evidence to support
the principle.
If Davidsons principle is defendable either in a concrete context
or through a methodological examination, and thus qualied as indi-
cating a necessary condition of understanding and interpretation, can
it be used to kick out all kinds of skepticism and relativism? As to
this third problem, Suzukis interpretation of Zen Buddhism has been
selected as a challenge to Davidsons principle. I have argued that
the idea that giving up rational thinking and logic as a necessary
condition of attaining Zen could escape the locating problem required
by Davidsons principle. However, there is still a problem of land-
ing. For the transcendental Zen to be relevant to the empirical world,
it should solve its landing problem, for example, the problem of
explaining how people can transform their self from an empirical
into a transcendental state and how the Zen masters can criticize
dualistic thinking without committing to dualistic thinking. While the
locating problem is a request for identifying a location in a rational
space, the landing problem, on the other hand, is a request for
152 cn.r+rn ri\r
making sense of the relation between two kinds of truth or two lev-
els of world. My strategy is not to use Davidsons locating request
to force the Zen masters to give up their non-dualistic thinking for
identifying Zens location in a rational space, but to use the land-
ing request to force them into a dilemma: either to explain the rela-
tion with our rational language and logic or to fall into an abyss of
total isolation from the real world.
In order to illustrate the detailed dierences to the bottom of think-
ing in dierent cultures, Graham gives some examples to explain
why English and Chinese sentences are not fully intertranslatable. In
his rst example, he says that if an English speaker says, The cat
sat on the mat, and a Chinese, Mao wo tsai hsi-tzu-shang [Mao-wo-
zai-xi-zi-shang ], only the cat is satisfying these con-
ditions. For the English its posture is similar to a man sitting in a
chair, for the Chinese to a man lying (wo) whether face forward or
on his back. As for the mat, we cannot expect an unrelated lan-
guage to share precisely our classication of oor coverings as mats,
rugs, carpets; hsi-tzu [xi-zi ] is used of straw mats. In addition,
the verb is tensed in English but not in Chinese.
Therefore, he
thinks that the sentences do not express the same proposition, because
Mao wo tsai hsi-tzu-shang is true even if cat has never before now
sat on the mat, false if it sat on a cloth mat. More generally speak-
ing, the reason is that Chinese and English divide up and organ-
ise the world dierently.
The second example mentioned by Graham
is that when using Grass is green to match with Tsao ching
[Cao-qing ], we would nd that the meaning of tsao depends
on a division of vegetation into tsao mu [Cao-mu ] grass and
tree, implying a wider scope than our grass. Ching is one of the
Five Colours, the blue-green which contrasts equally with red, yel-
low, white and black. If grass were blue Grass is green would be
false but Tsao ching would be true.
So, he considers these two
sentences not translatable to each other.
Based on these examples, Graham pushes to the point that the
unintertranslatability between these two languages fundamentally lies
in word level, not in sentence level; in correlative thinking, not in
analytic thinking; in pre-logical process, not in logical operation. In
A.C. Graham, 1992, p. 65.
A.C. Graham, op. cit., pp. 6566.
154 cn.r+rn ri\r
this regard, to explain the dierences to the bottom is tantamount
to understand the words by correlation within the scheme, a pre-
logical process which analysis assists but cannot replace. He believes
that a native speaker is in command of his or her home language
only when he or she stopped analyzing and applying grammatical
rules to the chain of oppositions; and the gap of the chain, for exam-
ple, dog/dogs, tree/trees, house/ . . ., can be spontaneously lled
with houses. So, he thinks, if I learn the words primarily by cor-
relating them, with analysis secondary even if employed at all, I
understand the Chinese as I understand the English, and can conrm
the truth of either Tsao ching or Grass is green by looking at grass
and other herbs without bothering about translatability. If in a par-
ticular context a Chinese reports what he saw by Mao wo tsai hsi-
tzu-shang, I am oriented towards what he saw as towards things I
have seen myself, possibly but not necessarily by visually imagining
as in my own case I visually remember. I can then say, The cat
sat on the mat, as I might say, You still have that cat then, respond-
ing to the event which he observed without concern for whether I
am saying what he said. I do have to co-ordinate the Chinese and
English sentences, but will do so most accurately by correlation sen-
sitive to more dierence and similarity than I can analyse; there is
no need to relate them logically because if I want to infer from one
of them it will be in the same language.
Names, for Graham, cannot be understood as singular terms of
logic; he treats names as the products of the classifying act of nam-
ing in a pre-logical pattern. He thinks that some of the Chinese and
English names seem to match each other but actually dier in their
dierent patterns of classication. An interesting example is the trans-
lation of the Chinese word yang. He says, there is a single goat in
plain sight, and X says Yu yang There is a yang (conventionally trans-
lated sheep) and Y There is no sheep, I may be startled if I fail
to appreciate that yang include goats as shan yang mountain yang
[ ]; but for anyone who has fully correlated the Chinese and
English words the observation conrms both sentences. Even though
Davidson oers a similar case of agreement disguised by dierent
usages of yawl and ketch, Graham thinks that his account diers
from Davidsons in not having to assume (even if it is indeed the
A.C. Graham, 1992, p. 67.
r.\irsoxs 155
case) that the extensions of yang and sheep and goats precisely
coincide, that they are intertranslatable like the dates of the two cal-
endars. The distinct picture of his example is that [w]hen read-
ing explanation in two languages of a vocabulary dierence between
them one is positively grateful that they do not say exactly the same
thing, much as when collecting information about an incident one
wants photographs taken from dierent angles at dierent moments.
On this basis, therefore, he judges Davidsons criticism of Whorf s
relativism unsound, because there is no paradox here; Whorf would
hardly have denied that bilingual readers would be clearer about the
divergence with an equally sophisticated Hopi account to compare
with his.
Unlike most of the views of justication held by the people in the
eld of epistemology, Graham thinks that to conrm or to refute
requires not only logic and observation but also checking whether
words have the same sense.
In contrast to the justication by logic
and observation, this checking is not on the level of proposition,
but on the level of pre-logical patterns of names. Hence, he con-
tinues, [t]o escape the conclusion that all truth is relative to incom-
mensurable conceptual schemes it is enough to show that the schemes
themselves are patterns of names neither true nor false, and that fac-
tual statements depend on them for their meaning but not for their
truth; we need not bother about what lies between these extremes.
In addition to the syntagmatic connextions of pairs in a chain,
Grahams notion of conceptual scheme also includes the syntactic
structures which organize words in sentences. To compare English
with Chinese grammar, he believes the dierences in conceptual
schemes are also obvious. With respect to this point, Graham rec-
ommends Hansens claim that classical Chinese nouns in general are
closer to English mass than to English count nouns. He says, Hansen
argues that Western thought is predisposed by number termination
to conceive the world as an aggregate of distinct objects, Chinese
by the mass noun to conceive it as a whole variously divisible into
parts. Le Galls inability to understand the chi [qi ] except as a
collection of atoms would be a good illustration. The hypothesis sur-
vives Harbsmeiers classication; most or all philosophical terms would
A.C. Graham, 1992, p. 68.
A.C. Graham, op. cit., p. 69.
A.C. Graham, op. cit., p. 72.
156 cn.r+rn ri\r
presumably be not count but generic nouns. This is plainly the case
with chi; the yi chi one chi [ yi-qi ] divides into the erh chi
two (sorts of ) chi [er-qi ] down to the wan w myriad (sorts
of ) thing [wan-wu ]. That Chinese thought would be condi-
tioned to divide down rather than add up is in any case suggested
by other features of the language.
Graham, therefore, thinks that
[t]here would be no such compulsion to assume the primacy of
individuals if English, on the analogy of Classical Chinese, lacked
number termination, and we said the closet and its shoe and shirt
as we say its dust or its smell. The eect of number termination
is such that we cannot even make the simple statement that lan-
guage classies things as similar or dierent without implying in
advance that the thing-s are dierent.
Graham believes that Chinese thinking is in terms of process rather
than of static entities. Although Davidson recognizes that dierent
languages may individuate dierently over a certain range of words,
Graham does not agree that this is only a local diculty for trans-
lation. The impossibility for Davidson to explore such dierences
without sharing concepts that individuate the same objects, for
Graham, is now magically transformed into a situation of possibil-
ity. Graham mentions, for example, that a Chinese student of English
has been assuming that yang and sheep are synonymous but begins
to doubt it. He points out a sheep and a goat, asks of both Is that
a sheep?, and in the second case I answer No, a goat. He has no
need to guard against the danger that I might take him to be point-
ing at the horn; I cannot answer No, a horn because horn, unlike
goat, is not on the same paradigmatic level as sheep. It would be
less appropriate to his problem to ask the [sic] What is that? which
allows me to answer A horn, forcing him to introduce a shared
concept by narrowing his question to something like What is that
Graham thinks that introducing syntactic structure into the con-
ceptual scheme does not bring us nearer to epistemological rela-
tivism, because truths of fact are independent of the scheme. Here
the word independent, he stresses, does not mean that factually true
statements are translatable into any natural language, but that to
A.C. Graham, 1992, p. 4.
A.C. Graham, op. cit., p. 75.
r.\irsoxs 157
conrm or refute a factual statement by reason and observation you
have only to understand its place in the appropriate conceptual
scheme, you do not have to share the scheme. Nor does it suggest
that if schemes could be perfectly corrected by logic and observa-
tion they would all become the same.
Graham interprets Wittgenstein, Ryle, Kuhn, and Derrida as his
heroes who can make an important move to dig below the surface
of our supposedly exact knowledge to nd the correlative at its foun-
dations. He thinks that this recognition is the same as the Sinologists
when in searching for metaphorical roots of a Chinese concept he
discovers that to compare and contrast it with Western concepts he
has to explore their roots as well. In the end, the solution, he sug-
gests, is to accept and come to terms with the thought that analy-
sis starts from the results of spontaneous correlation.
A.C. Graham, 1992, p. 78.
A.C. Graham, op. cit., p. 81.
The landing problem in Suzukis idea of Zen can be focused on two
questions: (1) How can a Zen master and his disciples transform
their self from a delusive state into an enlightened state, how can
they communicate with each other, and how can they know that
they are enlightened? (2) How can a Zen master criticize the dual-
istic thinking from his non-dualistic thinking and how can he make
sense of his anti-logical thesis without logic?
One of the Zen masters dialogues (mondo ) with their disci-
ples quoted by Suzuki can be used to illustrate his idea of giving up
the duality of logic for entering into Zen enlightenment. This can
also be used for explaining our rst question about landing prob-
lem. Wei-kuan ( ), a Zen master of Tang dynasty, was asked by
one of his disciples about the question Where is Dao? The dia-
logue continues as follows:
Kuan: Right before us.
Monk: Why dont I see it?
Kuan: Because of your egoism you cannot see it.
Monk: If I cannot see it because of my egoism, does your Reverence
see it?
Kuan: As long as there is I and thou, this complicates the situation
and there is no seeing Dao.
Monk: When there is neither I nor thou is it seen?
Kuan: When there is neither I nor thou, who is here to see it?
In comparison with Robinson Crousoes physical isolation from other
people, the mental state of the Zen master stated above is absolutely
and logically isolated from the other minds. We can call this state
Robinson Crousoe in an isolated mental state, instead of Robinson
Crousoe in an isolated island: the person entering into a mental
state which is impossible to communicate with others. The terms I
and thou mentioned above may have two alternative interpreta-
The translation is taken from Suzuki, 1996, p. 209; similar dialogue appeared
in Suzuki, 2000, p. 342.
r.\irsoxs 159
tions: one is the conventional use in ordinary language; the other a
special use in Zen language. If it is the rst option, i.e., Wei-kuan
uses the term I to refer to I and thou to refer to thou as no
dierent from his disciples ordinary use, then he is still in the com-
plicated situation in which there is no seeing Dao as mentioned
by himself. In other words, the requirement he states for seeing Dao
is self-defeating, and his dialogue cannot help his disciples enlight-
enment. If it is the second option, i.e., Wei-kuan uses the terms I
and thou not in the conventional way, but in Zens way, it is clear
that his speaking is not against the conventional use of I and thou,
and thus not against the duality of the rational discourse, though we
do not know what they really mean. In this regard, we can say that
Wei-kuans mind falls into a lonely place; as a master with private
Zen experience, he can be called Robinson Crousoe in an isolated
mental state. Although he reaches the level of enlightenment, it is
logically impossible for him to communicate with others. Furthermore,
it is also logically impossible to explain how he could transform him-
self from the state of puzzlement to enlightenment, because he would
not know whats wrong with the conventional thinking for attaining
enlightenment without an analysis of the conventional thinking.
As regard to the second question of the landing problem, Suzukis
idea is also inescapable from the predicament of self-defeat. If the
negative sentence The ower is not red, nor is the willow green
can be regarded by Suzuki as the same as its armative The ower
is red and the willow is green when they are understood from two
kinds of perspectives or seen by two kinds of eyes: the Zens eye
and the ordinary eye, then what Suzuki declares is denitely not
really anti-logical or a violation of logic. Furthermore, if we agree,
for arguments sake, that there is Zens eye and Zens experience,
the Zen masters or Suzuki himself still cannot escape from the trap
of duality. In order to illustrate the self-refuting characteristic of
Suzukis thesis of anti-dualism, specically, and anti-logic, generally,
let me pose the following hypothetical dialogue between Suzuki
and me:
Suzuki: The absolute armation must not be the one accompanied
or conditioned by a negation.
Y M: Could you give us an example?
D.T. Suzuki, 1991, p. 68.
160 cn.r+rn ri\r
Suzuki: Certainly! For example, Zen would say: A is not A, there-
fore A is A and A is A, therefore A is not A.
Or more
concretely, I is not I, therefore I is I and I is I, therefore
I is not I.
Y M: The two examples you have just said can be formulated as
~Q, therefore Q and Q, therefore ~Q. If so, does your
sentence I is not I, therefore I is I mean or imply the sen-
tence It is not the case that I is not I therefore I is I?
Suzuki: Absolutely, it doesnt mean that!
Y M: If we simplify the above sentence ~Q, therefore Q as P,
and ~(~Q, therefore Q) as ~P, would you reject the sen-
tence P means or implies ~P or the sentence P, therefore
Suzuki: If my answer is yes, so what?
Y M: If you reject the sentence P, therefore ~P, how can you
assert the sentence Q, therefore ~Q? Here they share the
same form, and P and Q can be substituted by each other
in the above sentences. So, I think your absolute armation
or great armation is self-refuting.
Suzuki: If so, shouldnt I . . . I assert the paradoxes Q, therefore ~Q
and ~Q, therefore Q? Eventually, do I have nothing to
Y M: Absolutely, you do have something to assert! The Buddha or
Zen teaches us no attachment. So long as you are not attach-
ing to the idea of anti-logic or the transcending of logic,
and not excluding the duality in our ordinary way of think-
ing and in your Zen thinking, you will naturally and happily
live in a world without attachment.
If someone wants to hold a thesis of anti-logic or the transcending
of logic for freedom from the attachment of (logical) dualism, he or
she would be inevitably trapped in the attachment to another dual-
ism: the logical and the anti-logical. In reality, what Suzuki claims
is eventually self-refuting. This is his predicament related to the sec-
ond question of the landing problem.
D.T. Suzuki, 1991, pp. 5960; and 1996, pp. 115, 269.
D.T. Suzuki, 1971, pp. 3031; and 1961, pp. 272273.
I would like to express my gratitude to my colleague Professor Angelina Yee
for her comments on the English writing of this paper.
Davidson, Donald (1984), Expressing Evaluations, the Lindley Lectures (monograph),
University of Kansas.
(1985), Incoherence and Irrationality, Dialectica, 39:4.
(1986), A Nice Derangement of Epitaphs, in E. LePore and B. McLaughlin
(eds.), Truth and Interpretation: Perspectives on the Philosophy of Donald Davidson, Oxford:
(1993), Locating Literary Language, in Reed W. Dasenbrock ed., Literary
Theory after Davidson, University Park, PA: Pennsylvania State University Press, pp.
(1989),The Conditions of Thought, in J. Brandl and W. Gombocz eds., The
Mind of Donald Davidson, Amsterdam: Editions Rodopi, Grazer Philosophische
Studien Band 36, pp. 193200.
(2001), Inquires into Truth and Meaning (second edition), Oxford University Press.
(2002), Subjective, Objective, and Intersubjective, Oxford University Press.
Fung, Yiu-ming (2000), Gong-Sun-Long-Zi: A Perspective of Analytic Philosophy, Taipei:
Tung Tai Book Company.
(2001), Three Dogmas of New Confucianism: A Perspective of Analytic
Philosophy, in Bo Mou ed., Two Roads to Wisdom?Chinese and Analytic Philosophical
Traditions, Chicago and La Salle: Open Court.
Graham, A.C. (1992), Unreason within Reason: Essays on the Outskirts of Rationality,
LaSalle: Open Court.
Hahn, Lewis H. ed. (1999), The Philosophy of Donald Davidson, Chicago: Open Court
Publishing Company.
Jakobson, Roman and Halle, Morris (1956), Fundamentals of Language, Hague: Mouton.
Jakobson, Roman (1978), Six Lectures on Sound and Meaning, Cambridge Mass.: MIT
(1985), Selected Writings, ed. by Stephen Rudy, Hague, Paris: Mouton.
Lechte, John (1994), Fifty Key Contemporary Thinkers: From Structuralism to Postmodernity,
London: Routledge.
Moore, Charles A. ed. (1951), Essays in East-West Philosophy: An Attempt at World
Philosophical Synthesis, Honolulu: University of Hawaii Press.
Putnam, Hilary (1984), Reason, Truth and History, Cambridge, Mass.: Cambridge
University Press.
Stroud, Barry (1968), Transcendental Arguments, The Journal of Philosophy, 65, pp.
Suzuki, D.T. (1951), The Philosophy of Zen, Philosophy East and West, 1:2, pp.
(1953), Zen: A Reply to Hu Shih, Philosophy East and West, 3:1, pp. 2546
(1954), Zen and Pragmatism: A Reply (Comment and Discussion), Philosophy
East and West, 4:2, pp. 167174.
(1955), Studies in Zen, New York: Dell Publishing Co., Inc.
(1960), Lectures on Zen Buddhism, in D.T. Suzuki, Erich Fromm, and
Richard De Martino eds., Zen Buddhism and Psychoanalysis, New York: Harper and
Row, Publishers.
(1961), Essays in Zen Buddhism (First Series), New York: Grove Press, Inc.
(1971), What Is Zen? London: Buddhist Society.
162 cn.r+rn ri\r
(1991), An Introduction to Zen Buddhism (revised edition), London: Rider & Co.,
An Imprint of Random House Group Ltd., 1969.
(1994), Living by Zen, York Beach, Maine: Samuel Weiser, Inc.
(1996), Zen Buddhism: Selected Writings of D.T. Suzuki, ed. by William Barrett,
New York: Doubleday & Company, Inc.
(2000), Essays in Zen Buddhism (Third Series), New Delhi: Munshiram Manoharlal
Publishers Pvt. Ltd.
Samuel C. Wheeler
Chinese ethical thought has long been a paradigm for the West of
a way of thinking that is sophisticated, highly developed, advanced
on any reasonable scale of advancement, but very dierent in many
judgments from the ethical thinking of the West. Stock examples
include diering degrees of respect for authority, dierent views about
the responsibility of groups for the actions of individuals, and the
like. This essay investigates whether such divergence shows that ethics
is less objective than natural science. This essay concerns the abstract
question of the objective truth-values of ethical sentences, rather than
any particular dierence that might be claimed to hold between these
two cultures ethical thinking.
1. Kant and Founding Ethics on Rationality
Many ethical theories have attempted to show that it is rational to
do the good. The idea is that the relatively uncontroversial norma-
tivity of correct reasoning can found the normativity of morality.
The most ambitious of these theories was that of Kant. Kant attempted
to show that it was irrational to do anything other than the right
thing. The argument for the categorical imperative is an argument
that the very concept of acting on purpose requires that a perfectly
rational agent do only what has a coherent general principle. The
least a perfectly rational agent could do, given that he was acting
for reasons, would be to act so that he was consistent, because act-
ing for reasons requires acting according to considerations that apply
generally. If the considerations are universally applicable, it ought to
be possible that everyone follow them. Thus acting for reasons con-
sistently requires acting in a way that anyone could act.
166 cn.r+rn six
Kant hoped that this criterion of consistency would suce to give
both necessary and sucient conditions for moral actions. Kants
program was quite ambitious. He not only tried to supply a ratio-
nal ground for doing good when a person knows what the good is,
he also required that there be a rational basis by which one could
always know what the good is. Rationality would not only provide
a reason for doing the good, but would also provide a criterion for
the good that would enable a careful person always to do the right
Kants conception of doing the right thing is doing ones duty
or fullling ones obligations. An ethics based on obligations and
duties is essentially deductive, since principles of obligation yield argu-
ments that can be known to be sound on the basis of limited infor-
mation. Thus, a Kantian theory requires that ethics have the structure
of a system of principles, so that ethical reasoning is primarily deduc-
tive. Given adequate principles, a person is equipped to know the
right thing to do in every circumstance. Thus the person must be
able to deduce the correct thing to do on limited information.
2. Davidsonian Kantianism in Outline
A combination of several ideas in Donald Davidsons philosophy sug-
gests a way to reconstitute something like Kants project. For Davidson
the normativity implicit in reasonable interpretation of another as
an agent supplies the basis for understanding the normativity of eth-
ical concepts. Interpretation, though, uses a much richer notion of
rationality, one that includes inductive logic. The conception of
ethics that results is not deductive. No algorithm will allow even a
perfect rational agent to always do the right thing. Davidsons ideas
connect the normativity of rational interpretation to the normativity
in the truth-conditions of ethical language. On this account, the nor-
mative is not dierent in ontological or epistemic status from the
Let me sketch some of these Davidsonian ideas:
a) Interpretation as Maximization of Agreement:
Davidson, following the suggestions of Quine, uses the relatively
testable and empirical notion of what we do in understanding another
r.\irsoxi.x n.+iox.ri+v .xr r+nic.r ris.onrrvrx+ 167
to analyze the concept of rational agent. He determines the con-
tent of rational agent or entity acting for reasons by seeing what
constraints we apply in interpreting another. Davidsons account of
interpretation yields an account of rationality that is rich enough to
give hope that one could show that doing good is reasonable.
Interpretation maximizes agreement, so that interpretation is con-
strained by the necessity for treating the other as mostly believing
the true and wanting the good. Maximize agreement is a formula
for a set of probabilistic constraints whose content is, roughly, Interpret
on the supposition that the entity being interpreted is an agent.
Maximize agreement says, Interpret the agent by using the fea-
tures of a paradigm agent, namely oneself. To interpret according
to a probabilistic t with our own case of rational agency is to apply
the term agent on the basis of a hypothesis that the other is like
us, an agent. Thus, since this is the general procedure for applying
predicates, the reasonableness of such constraints is a priori.
Familiar examples of the operation of maximization of agreement
discuss maximization of agreement in beliefs. But action interpreta-
tion must also maximize agreement in desires, that is, conceptions
of the good, as well. When someone drops a rock on his foot, the
interpretation that the person wanted pain and believed this was an
eective way of bringing it about is reasonably rejected, barring very
special circumstances,
and interpreted as unintentional. Action inter-
pretation always maximizes agreement in both desire and belief.
Given only that the other entity is an agent, and given only that x
is a belief or desire, if we have that belief or desire x, so, probably,
does the other entity.
A broadened conception of the rational that constrains the con-
tent of desires is required for action interpretation. Without such
constraints on content, interpretation could not get started, since any
behavior is consistent with any beliefs, if that any desire whatsoever
For instance: Aliens have landed on campus and are choosing whom to abduct
for their experiments. They wish to take only those without already existing pain,
so as to have an accurate baseline. In this situation, seeking pain would be an
From this thesis alone, some degree of agreement on ethical matters follows.
Other things being equal, a value one individual holds is a value the individual
being interpreted holds as well. Some agreement on values, conceptions of the good,
follows from another entity being a rational agent.
168 cn.r+rn six
is as likely to be present as any other.
So, such constraints as that
a person ought not to want pain are part of the concept is a rational
Speech interpretation is a special case of action interpretation.
What a person is doing is, for example, asserting that the cat is on
the mat. Speech interpretation, though, requires hypotheses about
the speakers intention. For instance, in interpreting an utterance as
a sincere assertion, one applies maximize agreement in beliefs in
the light of the hypothesis that the person desires to say the truth.
Thus we seek a truth-denition for the language the person is speak-
ing that will make the assertion true. Given that the utterance is a
sincere assertion, i.e. that the intention is to tell me what is the case,
and given that the speaker is by and large a believer of truths, the
utterance ought to be true. The ought, which is close to a prob-
ably, falls out of the very structure of the constraints on interpre-
tationthat they are maximization constraints and not absolute
constraints. Briey, principles of interpretation are ought principles.
b) Ought Sentences
Davidsons discussion of weakness of the will
presents a brief account
of conditional ought-sentences (hypothetical imperatives) that sug-
gests a connection of ethics with interpretation-theory. Roughly, con-
ditional ought sentences are formally akin to conditional probability
sentences, in that they do not detach when the antecedent is true.
In logical form, conditional ought-sentences are relative to some-
thing like an consideration-base, akin to an evidence-base to which
a conditional probability claim is relative.
That is, the apparent antecedent of a conditional probability
claim, an If A, then probably B claim, is really the evidence rela-
tive to which the consequent is probable.
By analogy, conditional
Rather than try to give a general proof of this, I will give an example. Consider
a person who is tying his shoe. Given an overwhelming desire that a comet crash
into Connecticut and an utter lack of concern about shoe-wearing safety, this could
reect the belief that tying this shoe now will bring that about.
In Davidson, How is weakness of the will possible?, Essays on Actions and Events,
Oxford 1980.
This point about conditional probability is from Hempel (1960), Inductive
Inconsistencies, Synthese 12, pp. 439469.
Is probable is related to the probability of A, given B is 0.5 in something
r.\irsoxi.x n.+iox.ri+v .xr r+nic.r ris.onrrvrx+ 169
ought sentences give the bearing of a consideration on whether a
person ought to do something. Ought, like probably, is funda-
mentally inductive.
In contrast, obligation sentences, the focus of Kants ethical the-
ory, have the logic of deductive argument, as outlined in various
deontic logics. But ought, not obligation, is the primary ethical
concept, since one can always ask whether one ought to fulll an
obligation. If ought is the primary ethical concept, and Davidson
is right about the form of ought-sentences, then we get a very
dierent picture from Kants of the connection between rationality
and ethics. Ethical arguments and ethical thinking are properly under-
stood on the model of induction rather than deduction. In an induc-
tive argument, a set of true premises {p1 . . . . . pn} that strongly
leads to the conclusion B is compatible with the existence of a true
premise pn+1 which, together with {p1 . . . pn}, strongly leads to
the conclusion not-B. In an ethical consideration, a set of truths
{p1 . . . . . pn} may strongly lead to the conclusion that you ought
to do B, while there can be another truth pn+1 which, together
with {p1 . . . pn}, strongly leads to the conclusion you ought not to
do B.
If ethical argument is like induction, then ethical reasoning does
not proceed according to Kantian universal moral principles. Briey,
an ethical actor will not in general know that he is doing the right
thing, since he is always operating on limited information, using
guidelines that allow for the possibility that further information can
undermine the conclusion so far reached. The principles that can
be applied are not universally quantied commands, but rather rules
of thumb. No algorithm will determine, given a description of a sit-
uation, what in fact a person should do in that situation. Furthermore,
the rationality that might be the foundation of ethics would require
much more than formal consistency. Ethical conclusions are more
like scientic conclusions than like logical consequences. Just as there
is no algorithm for the construction of a new theory in the light of
novel experience, so there will be no algorithm for the correct eth-
ical judgment, given that every situation is novel in some respects.
like the way that tall is related to Is 2 meters in height. Is probable and prob-
ably, the adverbial form, are attributives on a dimension of the degree of support
the antecedent gives to the consequent. If A, then probably B means, roughly,
the probability of B given A is pretty high.
170 cn.r+rn six
Davidsons account on which ought-sentences have a form sim-
ilar to that of conditional probability sentences suggests a theory of
ought-sentences. If ought-sentences are essentially like conditional
probability sentences, then their categorical form has an implicit rel-
ativization. Just as you may say It will probably rain, so one may
say, We ought to help those people. In both cases, there is an
implicit relativization, represented more explicitly by an All things
considered clause.
Such relativizations suggest a unied account of the distinctions
between the various senses of ought. Consider the general princi-
ple that one ought to believe the logical consequences of what one
believes. This principle has obvious counter-examples. If Fred believes
that Susan is honest and that Susan has been embezzling his funds,
he should not conclude that honest women sometimes embezzle, but
rather ought to give up one of his premises.
The various senses of ought are to be understood as further rel-
ativizations to backgrounds. For oughts applied to human intentional
actions, various implicit consideration relativizations yield dierent
oughts. The prudential ought might be relativized to an implicit
set of a persons self-concerned interests, the logical ought might
be relativized to considerations of consistency, and so on.
A Davidsonian account treats all ought sentences as having the
same basic structure and semantics. The methodological principle
operative is that, other things being equal, a theory should minimize
homonyms. Could the ought in You ought not to inict pain
and in If you turn the key, the car ought to start be the same
word? The two occurrences may be dierent words. However, a
plausible theory that makes the two oughts the same word is prefer-
able to a theory that multiplies homonyms.
3. Relativizations
a) Two Kinds of Relativization
A conditional ought utterance, just like a conditional probability
utterance, has two general kinds of relativities. These are further
complicated by the fact that they are sometimes implicit rather than
r.\irsoxi.x n.+iox.ri+v .xr r+nic.r ris.onrrvrx+ 171
(1) The rst kind of relativity is often overtly specied in the
antecedent of the utterance. If you want a nice dinner, you should
go to Caveys recommends an action relative to a desire. On the
normal understanding of such an utterance, the utterance is a pru-
dential ought. The prudence indicated is relative to having the
desire specied in the antecedent. Given that desire, and given infor-
mation that you have, and so that the other should have, the action
that the other should (prudentially) take is to go to Caveys. When
the utterance is a prudential ought, the antecedent gives a special
circumstance which, given prudential considerations, make it rea-
sonable for a person to choose Caveys.
These conditional relativizations describe what, relative to the
sense of ought (the considerations described below), a person
should do given all the surrounding circumstances. That is, in the
light of all the information available,
if you want a good dinner,
you should go to Caveys. The bearing of the information on what
you should do is by and large inductive. Features of Caveys, infor-
mation about your tastes, information about alternatives, and so forth
reasonably lead to that conclusion, albeit not by an algorithm. In
eect, such reasoning is a practical syllogism with the syllogism
expanded to include more adequate forms of reasoning.
(2) The other relativity distinguishes the various senses of ought.
This is relativity to considerations intended to be taken to bear on
the choice. To illustrate, consider a player in a chess game, in a
position where she can force mate in four with a bishop sacrice at
f7, while every other move loses. In one sense of ought, it is clear
that she should go bishop f7 check. But her opponent is her boss,
who will be angry if this employee defeats him in chess. So, pru-
dentially she ought to make some other move than bishop to f7
check. But her loss to the boss may demoralize her colleagues so
that their lives are worse if she loses, even though her own life is
worse if she wins. Or, it might be dishonorable to lose a game on
A question here is whether the ought is relative to all the information thats
available or to all the truths that are relevant. If, unbeknownst to all, a medium-
sized meteor is in fact headed for Caveys, and meteor-impacts greatly diminish
dining pleasures, then what is reasonable to do given all available information may
dier from what it is reasonable to do objectively.
172 cn.r+rn six
purpose. The noble thing for her to do, and perhaps what she ought
to do, is win.
These dierent understandings are dierent restrictions on the con-
siderations relevant to what to do. In the rst case, the restriction
might be called chessic. Only the goal of chess, mating the oppo-
nents king, is taken to be relevant. What the person ought to do is
taken relative to just those considerations. Considerations that con-
cern only her personal welfare would yield a prudential ought. If
the recommendation that she move bishop to f7 were interpreted
prudentially, the claim would be that, taking into account only con-
siderations relevant to the self-interested desires of the agent, she
should go bishop to f7.
The sentence itself does not determine this relativization. Rather
the intent of the utterance does. So, even though the unqualied
recommendation in a book of chess problems that bishop to f7
is the best move, it may not be, given another intended set of
Such restricted background considerations can be regarded as a
set of interpreted sentences,
but the conditions of set membership
are too vague to make that a precise proposal. In particular, the
notion of personal welfare and self-interest will turn out to be irre-
deemably vague. The background considerations will have to remain
an unanalyzed primitive, akin to the backgrounds that are presup-
posed in remarks that it will probably rain, or in counterfactuals.
Analogous relativization occurs with antecedents in the logical ought. If you
believe its raining and you believe its cold, you ought to believe its cold and
rainy. Nothing recommends the belief apart from the antecedent, combined with
the limitation of considerations at play to the purely logical.
In the case of the logical ought other considerations that over-ride are even
more apparent. If the evidence is very much against rain and against cold, then a
person who believes it is raining and believes it is cold should abandon both beliefs
rather than believe their conjunction.
A Davidsonian will need a theory of propositions that will t the roles of propo-
sition. The details of how one does this by application of Davidsons On Saying
That (in Inquiries into Truth and Interpretation, Oxford UP, 1980, pp. 93108) are not
r.\irsoxi.x n.+iox.ri+v .xr r+nic.r ris.onrrvrx+ 173
4. Questions of Logical Form:
How Do These Two Kinds of Relativizations Work?
As Davidson showed long ago, the relativizations cannot be thought
of as having the form of conditionals that detach. Thus a condi-
tional ought-sentence cannot have the form of a connective join-
ing two independent clauses, one of which has an ought. This
implies that the conditional ought is basic, and that the absolute,
categorical ought is something like ought, all things considered.
This accords with recent results in probability theory that argue that
conditional probability, rather than absolute probability, is the fun-
damental probability notion.
So, an ought-sentence must be a two-
place modality, a construction that depends on more than the
truth-values of the two clauses.
Modality brings in a complicated set of diculties about logical
form and ontology on its own. A diculty, then, is that a serious
Davidsonian account of the form of ought-sentences will have to
include a general account of modalities, propositions (or their sur-
rogates), and conditionals generally. This is a daunting task, and one
that by no means complete.
Let me sketch some of the features of a completed Davidsonian
account, which features I will not address in any detail here:
First, the account should treat if . . then so that if must be doing
the same thing in if . . then . . ought sentences as in other occurrences.
Second, some account must be given of the objects that take the
place of A, B and C. If ought is a predicate, the considerations C,
the antecedent A, and the consequent B must be construed as
some kind of entities. A Davidsonian would assimilate such objects
to the general type of demonstrable linguistic objects, along the lines
of On Saying That.
Third, some general account of modalities will be required. For
a Davidsonian, it is arguable that modality must be primitive.
See Alan Hajeks What Conditional Probability is Not, Synthese, Vol. 137,
December 20, 2003.
In radical interpretation, having a theory of anothers speech is having an
account of what alternative utterances would mean. Counterfactuals are admitted at
the very core of the theory. Since an account of what a person means on an occa-
sion of utterance is central to the Davidsonian account of meaning, there is no
hope for a non-circular reduction of the modal to the linguistic, since the linguis-
tic presupposes the modal.
174 cn.r+rn six
dierent senses of the usual modal operators
will be dierent rel-
ativizations to considerations, as sketched above for ought. Consider-
ations, on this account, are construed as sets of things said that,
for the case of cant, function as limitations on the principles from
which the negation of the sentence is a consequence.
The modalities typically treated, necessary, possible, and the like,
are based on logical consequence, and so would have deductive
underpinnings, relative to dierent considerations. Given that there
is no hope of reducing induction to deduction, analogous modalities
that depend on inductive connections would use a primitive conse-
quence relation about which there is only the beginning of a the-
ory. If . . then . . probably and if . . . then . . ought express such
So, a working proto-theory would be that the form of conditional
ought-sentences is Ought (C, A, B) rather than (A->Ought
where C stands for considerations, and where ought is a modal-
ity whose rough content is that Ought (C, A, B) is true if and only
if, given A, B is reasonable on considerations C. Rather than being
some kind of conditional with a consequent containing a sense of
ought, a conditional ought sentence is a three-place modal pred-
ication among an antecedent, a set of considerations, and a con-
sequent. In order to ll places in a predication, A and B and C
must be surrogates for propositions or sets of propositions, rather
than names of truth-values. Thus ought is a predicate, with the
relativizations specied by A and C and the action by B.
5. Connections between ought and conditional probability
Ought often amounts to probably. If you turn the key, your car
ought to start is usually true if and only if If you turn the key,
your car will probably start is true. The relativization in a proba-
bility sentence is not merely to the contents of the if -clause, but
to background conditions as well. The assumed overall conditions
For instance, the senses of cant in which I cant be at the meeting (because
I have a dentists appointment), in which I cant get to China in one billionth of
a second, and in which I cant nd a ratio of integers whose square is two. I am
supposing that the modal operators can ultimately be treated as predicates of
things said or sets of things said.
r.\irsoxi.x n.+iox.ri+v .xr r+nic.r ris.onrrvrx+ 175
that the laws of nature obtain, the material in the tank is gasoline,
and so forth would also be implicit in the antecedent or relativization
of the probability sentence.
While if then probably sentences and if then ought sentences
are connected, a sharp denition of one in terms of the other is nei-
ther likely nor necessary. ought and probably are each semanti-
cally independent modal predicates. A rough account of oughts
connection with probably would rely on the probabilistic, maxi-
mization character of agreement. Given any feature that correct inter-
pretation maximizes, if an individual being interpreted is indeed an
agent, then that individual, relative to being an agent, probably has
the feature. For any truth, then, relative just to an entity being an
agent and the truth being a truth, the agent probably believes it.
ought is then roughly a chain of such probabilistic connections. A
person who believes as he ought, given his information, makes every
inference that he would probably make.
The picture of ethical reasoning is that any complex logical or
practical inference can be broken down into a series of smaller steps.
Each of the beginning premises is a desire or belief the person prob-
ably has. Each of the smaller steps is one that a person would prob-
ably make, (and so ought to make), just in virtue of being a rational
agent. So, ethical reasoning would arrive at ought-sentences by con-
structing what a person would do if that person always did what,
on principles of reasonable interpretation, the person would proba-
bly do.
For most real-life cases of an ethical decision, the chain of cal-
culations and inductive judgments in the ideal determination of what
a person ought to do will be enormously complex. Although some
actual ethical reasoning is something like ideal reasoning, most of
the time various short-cuts are appealed to. The short-cuts, or eth-
ical principles, are rules of thumb that by and large give the right
results. Given the information-costs of thoroughly calculating every
decision, it pays to have devices that give the right result much of
the time. Such short-cuts on occasion give the wrong result. In that
respect they are analogous to the irrationalities that Twersky
See my Inference and the Logical Ought , Nous, Volume VIII, No. 3,
September 1974, pp. 233258.
See Twersky, Amos and Kahneman, Daniel (1974), Judgment under Uncertainty:
Heuristics and Biases, Science, 185, pp. 11241131.
176 cn.r+rn six
others have researched. So, the ethical principles that Kant took to
constitute the core of ethical reasoning, and that allow deductive
determination of what a person ought to do on a given occasion,
are not, from a Davidsonian perspective, central at all, even though
they are important and largely true.
6. The Moral ought
a) Outline
According to this theory-sketch, the moral ought is the minimal
restriction of considerations, the relativization of relevant considera-
tions just to is a person or is a rational agent. The moral ought
is ought, all considerations considered (considered in the light of
what it is to be an agent). The rationality that is built into inter-
pretation is not just formal, but substantive: What a person desires
at bottom is interpreted by maximize agreement. The thought is
then that, given rational desires (in the content as well as structural
sense) and given rational practical inferences, what a person ought
to do will turn out to be the output. In eect, this yields the famil-
iar idea that, given that people at bottom want the same things and
are reasonable, there is a basic moral agreement. Disputes, one might
hope, can be settled by reasoned discussion rather than by brute
For this idea to yield morality as we understand it, the rational
desires will have to be strongly shaped by an argument that other
peoples desires rationally count, just as a persons own desires do,
that it is irrational to show exclusive concern for ones own interests.
Kant tried to show this by arguing from the very notion of purpo-
sive action, and its connection with universalizability. More recently,
Thomas Nagels book, The Possibility of Altruism
tried to establish
this, likewise by reecting on what it is to do something for a reason.
Unless a connection between rationality and counting others desires
as rationally motivating holds, the moral ought is a chimera. I have
There is something called obligation that has to do with contracts and
promises. The long-standing tradition of identifying morality with obedience to law,
specically the covenants with a divine being, have misled Western moral thought.
Nagel, Thomas, The Possibility of Altruism, Princeton UP, 1970.
r.\irsoxi.x n.+iox.ri+v .xr r+nic.r ris.onrrvrx+ 177
no arguments better than Kants or Nagels, although my discussion
of preference below will try to make the Kant-Nagel thesis seem
Even given the soundness of the Kant-Nagel argument that oth-
ers desires count, can an adequate moral theory resolve ethical issues
and arrive at ethical truth? I will argue that the answer is, Sometimes,
to exactly the extent that physical discussions can resolve disagree-
ments about fact. This weaker result leaves us with a substantive
morality, but without the idea that every ethical question has a truth-
value we can determine by calculation or discussion. Sincere people,
perfectly rational from their own points of view, can have irresolv-
able moral disagreements. But this does not mean that all moral
questions lack truth-value. The presumption that irresolvability means
relativity for ethics but not for physics rests on a mistaken view of
The lack of resolvability will be most apparent between cultures
with dierent histories and dierent traditions, as we will discuss
below. However, we should expect the same irresolvability, perhaps
on a smaller scale, between any two individuals. Just as for a
Davidsonian the fundamental language is the idiolect, so in cultural
terms the elementary culture is the individual. Same culture, like
same language, has no strict sense. Nevertheless, of course, I can
loosely say that I speak English and am a Westerner.
The dierence from Kants hoped-for result derives from the nature
of the constraints on interpretation that give a Davidsonian con-
ception of rationality. One important aspect of these constraints, and
the rationality that is constructed on their basis, indicates that irre-
solvable dierences will not disappear: The constraints are maxi-
mization constraints. The other shares ones beliefs and values only
probabilistically. Even when the values are shared, the ranking of
any pair of goods is only probabilistically the same. Furthermore,
our inductive inference-practices are only probabilistically identical.
So, in eect, just as each of us has his own idiolect, each of us has
his own rationality.
This feature is part of the cost of a notion of rationality that goes
beyond mere structure, but rejects the analytic-synthetic distinction.
Allowing that there can be rational constraints on content, so that
an interpretation that ascribes a desire for pain is probably rejected,
would still give us a sharp notion of the rational if there were a
sharp line between desires that are essential to being rational and
178 cn.r+rn six
desires that are not. However, for a Davidsonian, there is no sharp
line between a part of the theory that is essential and a part that is
contingent. Of course, there are important dierences between parts
of theories, and important dierences between desires. Some desires
can only be ascribed given overwhelming evidence; others are ascribed
failing strong evidence that the person lacks the desire. But there
are neither desires every rational agent has to have nor desires that
no rational agent could have.
b) A Problem with Objectivity?
Both of the above considerations, though, apply to physical theories
as well as to opinions about what one ought to do. Why do ques-
tions of ethics appear more subject to unresolvable disagreement than
questions about truth-values of non-ethical claims? The usual answer
is that, whereas agreement in physical theories involves only getting
beliefs in agreement, ethics involves beliefs and desires. Desires, it is
claimed, unlike beliefs, do not correspond to anything objective, so
that people dier in preferences about which there is nothing like a
true or false to supply objectivity. Unlike dierences in belief,
dierences in desires, it is held, reect nothing but internal dierences
in persons. There is nothing like triangulation for desires. There
appears to be no common world of value to which both an inter-
preter and the interpretee are related to when an individual expresses
a preference. So the interpretation of actions lacks one of the ground-
ings in a common world that interpretation of belief has. There
appears to be no common world of correct desirability by which
people learn what is better than what. I argue below that this appar-
ent lack of a common ground that functions as a common world is
c) A Quinean Hedonistic Solution to Objectivity
One might hope that physiological states such as pleasure and pain
could be the stimulations that ground all desires by supplying some-
thing analogous to the common world in which belief-triangulation
takes place. A Quinean version of hedonism could then say that the
relation between the good and pleasure and pain is like that between
external stimulations and a physical theory. Just as there is no sen-
tence-by-sentence reduction of physical sentences to patterns of exter-
r.\irsoxi.x n.+iox.ri+v .xr r+nic.r ris.onrrvrx+ 179
nal stimulation, even though the theory does nothing but organize
stimulations, so there is no good-by-good reduction of valuations to
pain and pleasure. An account of valuations would thus be grounded
in internal stimulations, but holistically. A web of belief and desire
would distribute the sensory meanings of sentences as degrees of
belief and valuation among the sentences of the language. This would
be a kind of pragmatism, since the organism would desire whatever
maximized its welfare. Dierences in preference would be explained
by dierences in physiology, leading to an innocuous relativism.
If the arguments that it is rational to value other agents welfare
are sound, the theory would be a grounding for morality as well as
for prudential thinking. Such a theory would say that, just as it is
unreasonable to take ones own perceptions as the only reliable ones,
so it is unreasonable not to take the preferences of others as count-
ing in the determination of the objectively valuable. Quinean hedo-
nism could thus be a universal utilitarianism of a sort.
The diculty with such a theory is analogous to the diculty with
the Quinean Web of Belief. The theory supposes that sensations are
given unconceptualized desires, the periphery of desire-stu that is
organized via beliefs into particular preferences for, e.g., volleyball
over badminton. Absent such a given, the special tie to pleasure and
pain as physiological phenomena would drop out. But something
akin to the above theory could be a Davidsonian account of ethical
7. Davidsonian Ethical Objectivism
A Davidsonian can treat ethics as objective, and hold that ethical
sentences have truth-values, even without a basis in pre-conceptual-
ized desire-stu. Consider the apparent dierence between beliefs
and preferences that marked a dierence between ethics and physics
above. We need not regard dierences in preferences as brute, just
as we need not treat idio-rationalities as making rationality rela-
tive to persons.
Just as we do not generally identify beliefs with truth, so there is
no reason to identify preference with our good or with the good.
Given our past experience, we know that some of our beliefs are
likely to be mistaken. In the same way we nd that some of our
desires were ill-considered. Thus, just as we do not identify the true
180 cn.r+rn six
with our beliefs, so we do neither identify the good with what we
want nor identify the rational with our total theory. We can realize
that we are less than perfectly reasonable, even though we hold, of
each particular view, calculation, and valuation, that it is reasonable.
The decision-theoretic tradition has taken preferences as just brute
facts about which nothing is rationally required but coherence, whereas
beliefs are true or false. But in interpretation, we clearly take some
preferences to be irrational, such as the simple preference for pain
over lack of pain. Some preferences can only be assigned to an agent
on the basis of very strong evidence. Pain is worse than no pain,
other things being equal, is true. So, prima facie, some better than
sentences are objectively true. If some better than sentences are
true, the reasonable supposition is that better than sentences usu-
ally have truth-values.
The conception of preferences that suggests itself is that a pref-
erence that A rather than B is a belief that A is better than B. The
causes that give rise to preferences would then be construed as akin
to the causes of the involuntary utterances
that are sensory judg-
ments. On this conception, pain and pleasure would, in motivating
action, be involuntary utterances This is bad and This is good.
Such involuntary utterances, of the speaker to himself, as it were,
are incorporated into the web of belief and desire. Sometimes they
are over-ridden by other considerations, just as sudden apparent
ashes of light are over-ridden by the consideration that you may
have an eye problem. That is, we voluntarily get our separated shoul-
der replaced in its socket by a very painful procedure. Such over-
riding is not the discovery that the pain is good, but rather that,
while the pain is bad, the conjunction of the pain and the cure is
good. Given that the person is reasonable, and knows that the same
event will be both a pain and a cure, the person takes steps to bring
about the event.
But this does not mean that the persons evalua-
tion of the pain itself was mistaken. A kind of practical failure of
reason can explain how that a person cannot bring himself to bring
about the event, even though it is better and, intellectually, known
On an account on which there is no given, sensory beliefs can be construed
as utterances that occur involuntarily, and have to be interpreted. In adapted
Quinean terms, such beliefs have to be incorporated into our web of belief and
This is essentially Aristotles account in the Nichomachean Ethics.
r.\irsoxi.x n.+iox.ri+v .xr r+nic.r ris.onrrvrx+ 181
to be better. In that event, the immediate evil, the involuntary utter-
ance that is the pain, is over-rated relative to the future good. I
argue below that we can regard the wide-spread disposition to choose
for actions favoring ones own interests as another example of the
same kind of failure of reason.
8. Better for Me, Self-Interest, and Objectivity
If agreement on the Good is indeed what we maximize in inter-
pretation, we should not take into account only our own involun-
tary utterances about the Good and the Bad. If triangulation is to
apply to the Good, the objectively good must be prior to both what
I think is good and what is good for me. Only if that is the case
could we claim that everyones desires reasonably ought to count as
reasonable for me to take into account. I will make this plausible in
two stages: First, I will argue that self-interest is by no means as
clear a notion as it has often been taken to be. Second, I will explain
how it is that most valuations can be correct while almost everyone
fails to take the moral point of view. I will then conclude the sec-
tion by examining what sort of ethics a Davidsonian will have.
I will argue that we should not expect that a Davidsonian ethics will
be a utilitarianism. Whether and to what extent ethical sentences
have non-relative truth-values will depend on what is the case about
a) What is Self-interest?
In outline, the argument of this sub-section will be as follows: The
particular interests of any normal agent go far beyond anything
identiable as particular states of the particular organism that is the
agent, or self-interest as usually conceived. Usually the identication
of ones interests with the interests of others is limited to a subset
of the others, and so would be termed a partiality, a discrimination.
Ascribing such partialities is part of interpreting action and speech,
and so part of the rationality that is part of interpretation. Interest
and lack of interest in the welfare of others is thus subject to rational
evaluation. Reasonable attachments are maximized in interpreta-
tion. But, if Kant and Nagel are right that it is rational to extend
ones concerns to the interests of every entity that has interests, then
182 cn.r+rn six
it is reasonable to extend the scope of ones concerns with the inter-
ests of others to everyone.
If Xs self-interest is what is better for X, the question is how
to understand the for in better for X. The relativization for is
relative to the interests of . But better relative to the interests of
does not correspond to the notion of self-interested as usually con-
ceived. The pains and pleasures of an individual are not usually all
that an individual seeks. The interests a person may have can include
things like world peace or the success of the Red Sox. Without a
reductive tie to given sensations, self-interest is not separate from
interests in very many things intuitively separate from the self.
There is more to peoples preferences than just partiality to one-
self, narrowly conceived. Not only do people often take their inter-
ests in their own welfare more thoroughly to heart than they take
the interests of others, but their interests include interests in the wel-
fare of a select group of others. Parents have special bonds to chil-
dren, friends to friends, and many Americans to other Americans.
Such special bonds that A has to B include the interests of B, (where
B can be other people or groups), in the interests of A.
There is no sharp line between the prudential interests of a father
and the interests of his child. The fathers happiness and well-being
is as closely tied to the daughters sensations and well-being as it is
to any of the sensations that originate in his own internal reports of
pleasure and pain. The line between self-interest and the interests
of others is not sharp in any individual other than a sociopath.
The identication with other peoples interests that ows from
human relationships is part of a normal life and part of a good life.
If every relationship were contingent on the continuing merit of the
participants, life would be lonely indeed. More or less unconditional
love is important. An interpretation of a persons behavior that attrib-
uted no such attachments would require substantial background. It
would be interpreting a pathological case.
Among the features that are maximized in interpretation are emotional responses
and emotional attachments. For a Davidsonian, emotions are subject to evaluation
by reason, and therefore part of the content of rationality, since we maximize appro-
priateness of emotional response in interpretation, just as we maximize appropri-
ateness of distaste or positive preference in the case of sensations. A person who
enters his burning house to rescue his Tupperware may be evaluated as having
behaved irrationally. So Kants move of leaving emotion out of morality is not
r.\irsoxi.x n.+iox.ri+v .xr r+nic.r ris.onrrvrx+ 183
Every cultures ethical theory makes some relationships reasons
for special concern. Interests in others interests, then, are partial
identications of the interests that motivate an individual with the
interests of others. They are also a kind of deconstruction of the pri-
macy and clarity of the purely personal preferences that have bedev-
iled ethical theories since the beginning. People routinely care about
a lot more than their own pleasures and pains, crudely speaking.
Partialities can be evaluated as rational or irrational, and we do
so routinely in interpretation. A partiality can be excessive, as some
allegiances to states and sports teams often are. Degrees of partiality
are also subject to reasoned evaluation. A man who is just as attached
to the Red Sox as to his children is being unreasonably loyal to the
Red Sox. The basis for such judgments is that there is no reason
for the partiality. Partiality can be inadequate. A lack of favoritism
towards ones children, for example, calls for explanation by point-
ing out that the mother is den mother of the group of Cub Scouts
of whom her child is a member. So interests in the welfare of others
can be rational or irrational.
The maximal scope of concern is concern for the interests of every
entity that has interests. Is this maximal partiality rational? If Nagel
and Kant are right, it is. If we are motivated by partialities that it
is reasonable to have, then this one, the moral point of view, is
rational, and reasonably ought to motivate any agent with interests.
The question is why it motivates so rarely, and so much in dispro-
portion to judgments of what is better for me and mine.
b) Why We Make This Mis-Judgment
Given that there is no well-dened subject matter for self-interested
desires, and given that people in fact take into account the desires
of at least a subset of others, the disposition to focus on limited per-
ceptions of the good must be akin to limited perceptions of the true.
That is, just as many people over-rate their own experience in judging
This is a very limited result, since there is no argument that concern with oth-
ers should be equal. It is implausible that a mother rationally should care about
all children to the degree that she cares about her own, for instance. And this
implausibility is built into interpretation.
So, large questions about genuine conicts of interests, when you want some-
thing for your child that I want for my child, are being ignored.
184 cn.r+rn six
what is true, so many people over-rate their own perceptions of the
Good in judging what is better than what.
How is this compatible with their valuations being mostly correct?
Here is an analogy: Most perceptual judgments are correct; error
tends to be in the inferences drawn from those judgments. In the
same way, most judgments of what is better than what are correct,
but the inferences are mistaken.
The fundamental selsh error is to judge that A and B is better
than not-A and not-B because A is good, where A is an involuntary
judgment that something is good. That is, while it is correct that A
is good, and better than not A, it may not be that A and B is worse
than not-A. In eect, people over-rate the goods that are immedi-
ate to them or that are experientially connected to goods that are
immediate. In the case of immediate judgments of value, involuntary
utterances This is good That is bad carry undue weight in calcu-
lating the value of compounds of which this and that are components.
A is better than B for C
can be treated as a perspectival notion,
that is, relative to Cs local valuational beliefs. The analogy with
points of view in perception should be apparent. Consider a pyramid
with a square base. The pyramid from my angle is square; from
another angle it is triangular.
So we can coherently take preferences
to be beliefs about what is objectively better, and reasonably take
narrow views on what is better to be due to failures of reason.
Good, better than, and similar words have many complexities. Good pole-
vaulter is not a conjunction, since a good pole-vaulter who is also a shot-putter
need not be a good shot-putter. A theory of good and better than also needs to
take into account good at and good as and well as two ways of using good for.
The theory that most attracts me is that of Richard Larsen, begun in his essay,
Olga is a beautiful dancer. For purposes of the present essay, I will suppose that
some theory of good and better than is in place, since little hangs on what the-
ory is applied.
I may make a mistake about what aspect the pyramid presents from my angle,
by careless observation. In the same way, Cs interests are not just Cs desires. Cs
interests are his objective goods, about which he can be mistaken, if he is misin-
formed about what the objective situation is. There is a dierence between what
C thinks is better for him and what really is. A correct perspective on the good
requires correct information about the true.
r.\irsoxi.x n.+iox.ri+v .xr r+nic.r ris.onrrvrx+ 185
c) Davidsonian Ethics
c1) Davidsonian ethics need not be utilitarian
What is better, all things considered, might or might not be what
is better for everybody, and might or might not be a summation of
everyones preferences. While an identication of what is better, all
things considered, with what is better for everyone is a candidate
Davidsonian ethical theory, it is not the only theory that could arise
from a Davidsonian approach. It could be that what is really bet-
ter requires something more. Roughly, the relation between what is
objectively better and the preferences of agents is analogous to the
relation between the truths and the summation of beliefs of agents.
While the beliefs have to be largely true, truth is not reducible to
consensus. In the same way, while most preferences have to corre-
spond largely to what is really better, the Good is not necessarily
the maximal satisfaction of preferences.
c2) Partiality and relativism
Partiality is perhaps the deepest puzzle about morality and its ratio-
nality and objectivity. Call a system of partialities a list of partialities
that a person ought to have and can have, with a ranking of which
partiality should trump which. One troubling fact about partialities
is that dierent cultures have dierent systems. While (almost) every-
one takes certain partialities to be normal and reasonable, there is
no way to argue for one system of partialities over another. Partialities,
by their very nature, are relative in the sense that they provide rea-
sons for preferential treatment of your son for you but not for me,
but not based on any intrinsic properties of your son. One source
of irresolvable dierences among cultures is the dierences in the
partialities that are normal in the culture. What, for instance, should
be the relation between ones partiality towards ones parents and
ones partiality towards ones children? Which should be greater?
A Davidsonian can take one of four positions about partiality:
(1) Kant is right, and all partialities are irrational, and therefore
morally mistaken. In eect this says that the moral point of view
requires not only taking everyones interests into account, but tak-
ing everyones interests into account equally. Perhaps something of
186 cn.r+rn six
parent-child relationships could be salvaged on utilitarian grounds,
perhaps not.
(2) One system of partialities, with their rankings of relative impor-
tance, is objectively correct. So, given that an American, Fred has
promised both his daughter and his father to spend the afternoon
with them, and he cannot spend the afternoon with both, there will
be an objective answer as to which partiality takes precedence. As
observed above, cultures dier in the partialities they regard as nor-
mal or obligatory. This position would be that, wherever that situ-
ation occurs, like-placed individuals should do the same thing.
(3) A system of partialities is correct only relative to a culture.
This option would be a form of relativism that allowed that there
were some objective truths about what was better than what, but a
whole range of states of aairs that were only good or bad relative
to the cultures in which individuals lived. Partialities would be val-
uations that were essentially subjective, about which there was no
right answer as to whether the given state of aairs, say that Fred
chooses to spend the afternoon with his father, was good or not. It
is hard to see how this could be compatible with preferences being
judgments that a state of aairs is good.
(4) A system of partialities is objectively right for a culture. Given
that Fred is an American, an application of partiality may be objec-
tively right even though for a Chinese in precisely the same situa-
tion, doing what Fred did would be objectively wrong. This option
takes into account that actors are situated in cultures.
This fourth position, to which I am most drawn, is not a relativism.
Except in rare cases in which it is indeterminate which culture an
individual inhabits, there is an objective answer to what action is
All four positions allow that there is a moral point of view in
which every being that has interests is taken into account, and that
one should take that view. The options that allow partiality to be
rational, however, allow that people ought not to, or that it is morally
acceptable not to, treat everyones interests as of equal weight.
9. The Ethics of Cultures
There are several grounds that should lead us to suspect that there
will be irresolvable dierences between distinct cultures. Such irre-
r.\irsoxi.x n.+iox.ri+v .xr r+nic.r ris.onrrvrx+ 187
solvable dierences are compatible with a great deal of agreement
on ethical matters. Some such dierences may be cases of one cul-
ture getting something wrong, others may be indeterminacies, and
others may be case of option (4) above. Some of the sources of irre-
solvable disagreement are the following:
(1) Dierence in systems of partiality is one reason that there can
be profound irresolvable dierences between cultures, as discussed
(2) Even though two cultures agree that, for instance piety and
courage are virtues, they may disagree on their relative ranking.
In such disagreement, one may be right, or the question may be
(3) Ethical mistakes may become imbedded in cultural practices.
If ethics is the science of what is really good, and of what is really
better, then ethics is very dicult in application to many concrete
cases in which many factors are relevant.
The familiar analogy might be with weather-prediction. Some weather
predications are correct and easy. If dark clouds are to the West, it
is likely to rain. If its January in Connecticut, it will be less than
90 degrees Fahrenheit. But determining in detail whether and how
much it will snow in Storrs is beyond even dedicated and sophisticated
specialists. However, what the weather in detail will actually be in
Storrs on a given day is almost certainly in principle predictable.
In the same way, if something like the above account is right,
very many choices about what ought to be done are likewise decid-
able in principle but practically very dicult to determine. The
dierence in the case of ethical questions is that questions in ethics
demand answers in a way that questions about the weather do not.
Since we have to do something, we decide.
Since we have some
rules of thumb, and some intuitions that must be right, we use those
rules and intuitions. Much of the time, just as a farmers weather
predictions are, we are right. Another dierence is that, whereas
everyone can determine when the weather-prediction is wrong, no
A moral skeptic may claim that the better analogy is with astrology, where
the King demands an astrological forecast, since the stars must be presaging some-
thing. The skeptic will say that just as the right response there is to deny that the
stars control destiny, so here the right response is that there is no objectively better
in general.
188 cn.r+rn six
such clear-cut evidence is forthcoming when the ethical theory embod-
ied in a cultures practices yields wrong results.
The necessity to have answers to ethical questions such that many
people can agree on the answers generates the ethical theories that
are embodied in cultures and which are part of the socialization of
members of a culture. When the answers generated by the necessity
to have an answer are mistaken, entire cultures can be wrong.
What should we expect when two dierent cultures have an eth-
ical disagreement on a kind of hard case? No matter how sincere
and rational the negotiators are, there is no general algorithm for
determining what is the better course of action. Each culture will
perforce be using its own conception of the rational. That does not
mean that there is no right answer. Sometimes both are wrong,
sometimes one is right, but some cases may just be indeterminate.
Yujian Zheng
1. Varieties of Normativity
People are familiar with notions of rules and norms at the group or
societal level, such as laws, moralities, religious imperatives, cultural
conventions, and more local regulations such as the dress code in a
school or corporation. Many of us are also used to certain personal
rules or policies, some of which are self-imposed, or voluntarily
adopted (and adjusted) as a result of ones own deliberation at a
particular time over questions such as how to solve the dynamic
conicts between ones short and long-term interests.
Most people, however, are perhaps unaware of other important
and fundamental types of norms, such as logical or rational principles
involved or deeply embedded in our ordinary practice of reasoning,
of understanding or interpreting the speech or behavior of others,
of forming beliefs and judgments of ones own, etc. There seem to
be natural enough reasons for such unawareness of the normative
nature of these principles: on the one hand, there is no clear consensus
as to what kind of thing is possibly involved in those forms of prac-
tice. In other words, whatever underlies them is largely implicit, and
so is not easily recognizable as a matter of explicit principles. On
the other hand, even those explicit logical rules we sometimes apply
consciously tend to be associated with empirical laws of nature, which
seem more readily recognizable as descriptive rather than normative.
This consideration leads to our last type of normativity, i.e., that
necessarily involved in our understanding or explaining natural events
or phenomena. In an obvious sense, the hypotheses in science, no
matter how well established they are, are products of, or results
ltered by, rational procedure, which ultimately is nothing but human
decisions in a collective and evolutionary form. This procedure is
explicable in terms of a certain set of rules and norms usually regarded
190 cn.r+rn sr\rx
as epistemic or methodological in nature (e.g. Ockhams razor). More-
over, the function of a scientic hypothesis is to explain (and predict)
certain natural phenomena by subsuming them under its nomological
structure and, to a no less degree, to serve as a (though provisional
or tentative, perhaps) normative constraint for other related hypothe-
ses, as well as for our ordinary beliefs or inferences concerning those
phenomena. In this respect, humanly discoverable and presentable
(thus understandable) laws of nature share the normative status of
methodological rules and principles including the most basic ones,
such as those found in formal logic.
The distinction between the descriptive and the normative appears
to be so fundamental to our conceptualizing the world that the rel-
ative scarcity of direct philosophical discourse about its use as well
as its foundation is striking. Corresponding to each level or type of
normativity mentioned above, for instance, there is an issue con-
cerning its source, which issue is distinct from the question of what
possible roles or manifestations this type of normativity may have.
A rough, yet very helpful expression of the distinction might be
formed in terms of direction of t: the direction of t for the descrip-
tive is from the mind to the world, whereas for the normative it is
from the world to the mind. In other words, the criterion for descrip-
tivity lies in external reality, while that for normativity is internal.
Anything that satises such minimalist expressions could count as
descriptive or normative (at least in the context of this paper), no
matter how remote it may appear from ordinary perspectives.
2. An Integrated Approach to Normativity Underlying and Unifying
Davidsons Philosophical Contributions
It is well recognized that Davidsons thought has a remarkably uni-
tary character, even though it ranges over a wide array of problems
concerning knowledge, action, language, and mind. Although what
best characterizes this unitary character is perhaps something short
of agreement, Id like to demonstrate here that it is a distinctive
sense of normativity that can serve as a unifying thread to Davidsons
whole set of apparently interconnected and mutually reinforcing ideas
which have been developed separately in disparate contexts tradi-
tionally belonging to distinct domains.
. r.\irsoxi.x +o xonv.+i\i+v 191
Let me start with Davidsons well-known principle (adopted from
Quine) in his theory of interpretation, i.e., the principle of charity.
What the principle says is that you should optimize the overall truth
and consistency of beliefs of the person being interpreted when you
try to understand his or her speech or behavior. The normativity of
this principle is obvious. But the question is how to justify it. On
what basis can you be so sure that an interpretees beliefs are not
false or inconsistent, as most of us could surely nd ourselves in pos-
session of such unfortunate beliefs from time to time? Davidsons
answer is this: although you can be never sure that any particular
belief(s) of the interpretee must be true or coherent with his other
beliefs, his belief system as a whole cannot deviate largely from ratio-
nality and truth. For otherwise you as an interpreter would not even
be able to recognize or identify him as a being, like yourself, capa-
ble of believing, speaking or thinking. Any diagnosis of false or irra-
tional beliefs of his, therefore, could only take place against the
background of granting him many more true and rational beliefs.
The very possibility of interpretation, as one crucial form of human
interaction, and just like any other form, depends upon agents mutual
recognition of intentions or beliefs. The distinctive social existence
of human beings would be impossible without such recognition.
Moreover, interpretability governed by charity is not merely a method-
ological issue of understanding others, but also an ontological or
genetic issue concerning how our own intentionality (either as the
rst person interpreter or as a self-interpreted agent) is possibly formed
or constituted. Roughly, human intentionality, typically in the form
of reasons consisting of beliefs and desires, emerges from the same
processes of interpersonal interaction centrally involving interpreta-
tions. Such interactions, be they a matter of individual human infants
learning their mother tongues or of the social evolution of our ances-
tors, must, at least in their initial stages, involve some kind of tri-
angulation of causal relationship among two human subjects and one
shared object. The causal eects from the same object (whose loca-
tion is identied by the intersection of visual lines of the two sub-
jects) on each subject are compared or coordinated through mutual
Davidsons ideas on charity-governed interpretation can be found mainly in the
section of Radical Interpretation in Davidson, Donald (2001), Inquiries into Truth
and Interpretation, 2nd ed., Oxford: Clarendon Press.
192 cn.r+rn sr\rx
responding both to the object and to each others responses. The
basic causal connection between the object and whatever simple or
primary linguistic response of a subject somehow guarantees the opti-
mal condition of truth for the latter while the interpersonal (and
instantaneous) comparing and checking provide for rudimentary nor-
mative force of either a (well-established) rule-following (in the case
of an adult-baby or teacher-student interaction) or some norm-for-
mation or meaning-determination (in the case of two equals facing
a new or unprecedented situation). Davidson seems to rely increas-
ingly on such a triangular nexus of causal-cum-normative relation-
ship (as opposed to, say, his earlier notion of the omniscient interpreter)
in explicating the application of the concept of truth in his later
According to Davidson, were there no such a triangular rela-
tionship, not only would other people not be able to know my beliefs,
but even I myself would not be able to know what beliefs I have
because the very basis on which any and all beliefs depend and
become individuated and thus identiable would not exist. Without
such a normative basis, neither belief (and other propositional atti-
tudes) nor meaning, truth, and rationality could possibly take place.
Given these considerations it is not hard to see the holistic feature
of Davidsonian approach.
It does not stop at describing a static pic-
ture of interdependence among various beliefs, but further extends
to explicating the dynamic structure in which the formation of beliefs
is necessarily bound up with the emergence of truth, meaning (inter-
pretability), and justication (rationality). More concretely, Davidson
borrows Tarskis Convention T, a certain recursive linguistic device in
his semantic theory of the predicate true for formalized languages,
and converts its function into something like a foundation for a truth
theory of meanings (of every sentence in an object language), i.e.,
taking truth as the most primary and self-evident concept for con-
structing a theory of meaning. On the other hand, Davidson relates
See Davidson, Donald (2001), Subjective, Intersubjective, Objective, Oxford: Clarendon
Davidson explicitly claims that [w]hat I think is certain is that holism, exter-
nalism, and the normative feature of the mental stand or fall together. See Davidson,
Donald (2004), Problems of Rationality, Oxford: Clarendon Press, p. 122. What he
fails to address, however, is the further question about the logical relationship
between the three, e.g., whether normativity has any priority status in some orderly
account of the mental over the others.
. r.\irsoxi.x +o xonv.+i\i+v 193
truth to the basic propositional attitude hold true (or believe) in
a broader theoretical framework, i.e., his doctrine of radical inter-
pretation with the principle of charity at its core. The normativity
of this doctrine, apart from its methodological and ontological aspects
mentioned above, also lies in its obvious epistemological implications:
Davidsons criticism of so-called the third dogma of empiricism
(i.e., the distinction between empirical content and conceptual schemes),
his rejection of skepticism and relativism,
and his emphasis on the
interdependence of three kinds of knowledge (i.e., about my mind,
other minds, and the external world)
all manifest the theoretical
power of a holistic approach in its eorts to dig up some deep and
integrated sense of normativity.
Aside from the integration of theories of knowledge, truth, and
meaning, another very inuential and integrated bunch of theories
for which Davidson is famous is in the area of studies on action,
events, psychological explanation, irrational behavior, and the meta-
physical status of the mental.
Due to space limitations, I shall only
concentrate on two central Davidsonian themes in this area in order
to exhibit their underlying commitment to normativity.
The rst theme concerns the relationship between events and
actions. The distinction between actions and events does not lie in
any ontological dierence between the two, but in what kinds of
description each is covered by. Purely physical events can only have
physical descriptions, whereas actions (or mental events) can have
both intentional descriptions and physical ones. Any event describ-
able as intentional must be in principle also subject to physical (neural
or physiological) description, but not vice versa. In brief, actions are
events, but events are not necessarily actions. A necessary condition
for an event to become action is for it to be under some particular kind
of description.
E.g., The room was dark and I ipped the light switch
See On the Very Idea of a Conceptual Scheme, Inquiries into Truth and
See ibid.; A Coherence Theory of Truth and Knowledge and Afterthoughts,
Subjective, Intersubjective, Objective.
See Three Varieties of Knowledge, Subjective, Intersubjective, Objective.
This area is mainly covered in his rst anthology, see Davidson, Donald (2001),
Essays on Action and Events, Oxford: Clarendon Press. And more discussions on irra-
tionality appear in the last section of his Problems of Rationality.
The phrase under a description, though frequently used by Davidson, origi-
nally appears in Anscombe, G.E.M. (1957), Intention, Oxford: Basil Blackwell.
194 cn.r+rn sr\rx
describes an action of mine by suggesting an obvious reason why a
physical event, i.e., my nger movement or ipping, occurred. By
contrast, the description I ipped the light switch and the prowler
lurking in the bushes outside the window was alerted does not cor-
respond to any intentional act of mine if I did not know there was
a prowler out there, even though the ipping seemed to be causally
responsible for the alerting. Yet there was only one event (or one
causally continuous process) here: the event of unintentionally alert-
ing the prowler is the same event as my intentionally ipping the
light switch which, again, is the same event as the physical move-
ment of my nger (and any other relevant parts of my body) in a
certain way. Alternatively, it might be argued, despite the fact that
ontologically speaking there is only one unied causal chain or nexus
within a certain spatiotemporal domain (or one zone as Quine
would say), its possible manifestation as dierent events or types of
event, which correspond to dierent cognitive meanings, depends
upon dierent descriptions from selectively adopted perspectives on
the same nexus of physical elements. Whether a perspective is mean-
ingful or signicant depends in turn on whether the pattern identiable
from the perspective has some explanatory or justicatory role to
play in the larger context, or ts with some larger patterns which
nonetheless are only recognizable by ascending to higher-order per-
spectives. So, for instance, the life-world meaning of my turning on
the light will be lost if we replace it by shifting entirely to the lower-
level descriptions of the minute neurological facts underlying my
above action, even though the latter descriptions may still have cer-
tain cognitive meanings for all physiologists. I shall come back to
the implication of this theme for normativity presently.
The second Davidsonian theme concerns the relationship between
reasons and causes. To qualify as an action, some reason rather than
a mere cause of it must be shown. Generally speaking, causes and
reasons belong to dierent logical spaces: i.e., the space of the descrip-
tive vs. that of the normative. What explains an action is the so-
called primary reason, which consists of a pair of propositional
attitudes, i.e., a belief and a desire. As indicated above, the consti-
tution and recognition of a belief must rely on the holistic applica-
tion of normative principles; similar arguments apply to the motivational
part of a primary reason (and, moreover, many desires in real con-
texts are mediated or activated by beliefs of various kinds). A unique
contribution of Davidsons here lies in his both acknowledging the
. r.\irsoxi.x +o xonv.+i\i+v 195
above distinction of logical spaces and emphasizing the fact that rea-
sons and causes refer to the same ontological entities or events. In
short, the (primary) reason for an action must be its cause (i.e., what
actually mobilizes it). Therefore, the explanation of action by appeal
to its reason must also be a causal explanation in the sense in phys-
ical science. If this Davidsonian thesis is tenable, the relationship
between normative and descriptive elements will become much more
interesting as well as more complex than may otherwise be supposed
(as in some Wittgensteinian thesis). One may get a sense of the com-
plexities such an approach may lead to by looking at Davidsons dis-
cussions of how irrational behavior (such as weakness of will) is
possible, especially with regard to his treatment of unavoidable para-
doxes involved in the attempts to explain irrationalities. On the other
hand, nonetheless, the rst theme above, i.e., description determines
the meaning of an event (or action) also reveals, from another
aspect, the coupling between the natural event interpreted as a mean-
ingful object on the one hand, and the normative constitution of its
meaning by the subject who is doing the describing on the other.
What is description? This is anything but a question merely for
linguistics or rhetoric. I take it as a fundamental and far-reaching
question for Davidson and his integrated project. Although he oers
no direct and systematic account, one may still perceive the central
status of this (potential) question in his various discussions of topics
related to actions and events. What interests me most, however, is
how description in the Davidsonian sense relates to normativity at
this or that level. Here are just a few tentative observations of mine.
First of all, description, as a kind of (linguistic) action issued from
the subject, must have some kind of intentionality, or must be com-
panied by a certain intensional context the subject is aware of. Thus
it must already involve the aforementioned normativity connected
with the individuation of beliefs. Secondly, the means of description
can only be language, while the birth (and xation) of linguistic
meanings can only be eected via logical relations among a massive
amount of sentences as well as the Davidsonian triangular relations
against which certain basic sentences are situated. Both of these two
sorts of relations have already embedded or embodied various nor-
mative principles. Thirdly, description qua propositional content that
is supposed to t the (external) object in the world, must obey the
factual links or natural constraints in the space of causes. That is
exactly the objectivated sense the term the descriptive possesses.
196 cn.r+rn sr\rx
Such a sense of objectivity makes a perfect dichotomy with the nor-
mative force or eort issued from the subjectivity of free will. Thus
it can be seen that the notion of description, just like the notion
of truth, is subject to pulls or constraints from two opposite direc-
tions (for Davidson, truth has two contrary features at the same time,
i.e., external correspondence and internal coherence); so it can play
some indispensable, crucial role between the space of causes and
that of reasons, or between nature and human rationality.
Following this line, one may easily understand why Davidson makes
another seemingly strange combination in philosophy of mind: he
subsumes mental items under the ontological category of physical
items and yet insists that the mental descriptions can never follow
the nomological model of the physical oneshence the label anom-
alous monism.
In summary, since mental events are endowed with
normativity in their very genesis or constitution, mental description
or interpretation can only be governed by rational norms such as
the principle of charity, which nd no exact counterparts in our
descriptions of pure physical events. But, on the other hand, descrip-
tions/interpretations of mental events also have truth values, and
their truth conditions are rooted in the causal interactions between
mental and physical event types. As all causal relations can be sub-
sumed under strict forms of natural laws, each mental event must
therefore be token-identical to a physical event. Otherwise, some
non-physically-describable events would enter into strict forms of
physical lawswhich seems to be a contradiction. In other words,
there cannot be strict laws connecting the mental domain to the
physical one, or the space of reasons to that of causes; neither can
there be strict causal laws in the space of reasons, i.e., among men-
tal items themselves. Thus it seems clear that the idea about the
normative space of reasons as well as the normative distinction
between dierent spaces play some key role here.
See Mental Events and Psychology as Philosophy, Essays on Actions and
. r.\irsoxi.x +o xonv.+i\i+v 197
3. Can There be Interpretative Ruptures in Comparative Philosophy or
Cross-Cultural Understanding from the Davidsonian Perspective?
What is the possible implication of the Davidsonian approach to nor-
mativity for our understanding of the status of Chinese philosophy
or culture at large in relation to world philosophy or (Western-dom-
inated) global culture? Many peoples intuition seems to suggest that
dierences between Chinese and Western mind-sets cut much deeper
than supercial comparisons or translation practices might reveal.
But could we really pin down where these presumably deep dierences
lie? Could they, once identied, count as dierences between two
incommensurable paradigms in the Kuhnian sense?
Davidson rejects the relativist incommensurability thesis.
His rea-
sons might be summarized as the following: Whichever culture or
language you may be born into, the physical world (in its main,
structural features as well as basic elements) is the same for you as
is for anyone in any other culture; the type of triangular causal
nexuses you must face in order to form beliefs, understand others
and the world under various contexts is also the same for you or
anyone who can manage to survive; and thus the overall optimiza-
tion of truth and rationality, warranted by the triangle and expressed
by the charity principle, is necessarily applicable to you. Therefore,
there should be no room left for incommensurable or untranslat-
able conceptual schemes or cultural models.
If we accept such an argument, will that mean that cross-cultural
comparative philosophy cannot have independent signicance, as
opposed to comparisons between personal opinions or theories of
two individual philosophers in one culture? In other words, will it
imply that Chinese philosophy, just like any other non-Western phi-
losophy, cannot distinguish itself from the Western or world philos-
ophy in any non-trivial sense, trivial here meaning that any theoretical
position can be regarded as distinctive from others? I guess many
would tend to nd the Davidsonian answer to be armative. In
For the classic explication of paradigms in science and their incommensu-
rability, see Kuhn, Thomas (1970), The Structure of Scientic Revolutions, Chicago:
Chicago University Press.
Davidsons main argument can be found in On the Very Idea of a Conceptual
Scheme, Inquiries into Truth and Interpretation.
198 cn.r+rn sr\rx
what follows, I attempt to sow a few seeds of doubt, without chal-
lenging the fundamental framework Davidson has laid out.
The powerful Davidsonian weapon in attacking any relativist the-
sis about objective knowledge is the principle of charity. The applic-
ability of such a principle is grounded in some fundamentally normative
situation, to the eect that we have no choice but to follow certain
rules intersubjectively generated and maintained in accommodating
or coordinating each others causal reactions to shared stimuli so as
to facilitate our social symbiotic existence. But a crucial point then
is that such normative forces (enforcement and reinforcement through
continuous interactions) occur on dierent levels and dierent scales
without xed limits, and manifest themselves to individual conscious-
ness in either implicit or explicit forms (e.g., ranging from instinc-
tive skills of knowing-how to sophisticated symbolic representations).
Every utterance, whether it be a judgment or belief, a command or
a hope, comes gradually to carry more and more corresponding rela-
tionship of commitment and entitlement to other utterances.
extent and strength of such a relationship depends not only on par-
ticular contexts in which this sentence may appear, but also on norms
or rules of the language games in which the speaker may nd him-
self or herself involved. Although the Davidsonian charity logically
guarantees the optimality of truth and coherence for a majority of
beliefs, it does not automatically resolve the operational issue, i.e.,
how to tell a particular true belief from nearby false (or uncertain)
ones that may cohere with the rest of beliefs to dierent degrees
and/or in various ways. In other words, charity is indispensable as
a general starting point and perhaps an ultimate boundary condi-
tion, but it cannot replace the role of many concrete rules of the
game which actually constitute and regulate the intentional relations
and propositional contents of those beliefs involved. The diversity of
such rules and the multiplicity of levels at which they may exist are
simply obvious facts, which help account for the variety and prolif-
eration of games in real life.
Mother tongues or native cultures are normally taken as the grand
language games that shape types of individual consciousness and
forms of life. These games are isolated or independent from each
Such an inferentialist thesis is a natural extention of Davidsonian normative
holism and has been systematically argued in Brandom, Robert (2000), Articulating
Reasons, Cambridge, MA: Harvard University Press.
. r.\irsoxi.x +o xonv.+i\i+v 199
other as far as their causal genesis (say, in terms of geographic envi-
ronments) is concerned. Therefore we cant exclude the possibility
of cultural gaps between them. The problem seems to lie in the
fact that there is no simple and universal method to prove or mea-
sure the depth or width of such gaps. Perhaps one may suggest
that we could use certain statistical success rates of individual inter-
cultural communications as a rough indicator. But what kinds of dia-
logues (e.g., how many rounds of Q&A) could count as successful
One recommendable way of understanding a strange language/
culture without relying on translation is anthropological participatory
observations. If an anthropologist could merge himself deep enough
into the various life games of a primitive tribe which so far had no
exposure to the outside world, he would presumably emerge with
understanding of its oral language as well as its belief systems asso-
ciated with various customs within a limited time span. Contrastively,
however, we may imagine that a member of this primitive tribe sud-
denly came to visit our civilized society. Could he make sense of the
disparate natures or functions of all those essential games of our soci-
ety, such as politics, nance, and academic tenure system, within a
similar time span? Some asymmetry here seems to be obvious. The
oral language of the tribe may not lack some potential or even
explicit structure susceptible to rst-order quanticational logic. Terms
expressing negation, conjunction, and alternation are most likely avail-
able so that the basic inferential operator if . . . then . . . or its
equivalent is likely to exist too. There are good reasons to believe
that the linguistic dierence between a primitive and an advanced
culture is not categorical (i.e., between human language and a non-
human or quasi-human language), but rather of dierences between
higher-order semantic relations or rational norms of the respective
language games. For instance, Ockhams razor is such a high-order
cognitive norm in the games involved in scientic explanation. If the
interpretative failure caused by the above asymmetry between anthro-
pological and reverse-anthropological understanding is not acci-
dental, we could easily conceive that a similar failure will occur
between two primitive tribes of a similar advance level if their native
geographic environments are suciently disparate (e.g., one located
in tropical jungles near the equator while the other is situated in
the tundra of Northern Canada). The rules of the games that have
been developed through their unique responses to the challenges of
200 cn.r+rn sr\rx
survival in such disparate environments, including those regulative
norms accidentally created and yet socially selected because of their
tness to the games appreciation or proliferation, may have shaped
certain parts or structural aspects of their respective belief systems
to such an extent, under the condition of mutual isolation, that when
the two systems meet, the failure of interpretation is doomed to hap-
penthey cannot make sense of everything (signicant) in each others
behavior, i.e., one cannot nd any set of beliefs in ones own belief
system that suce to accommodate or explain a targeted behavior
of the other; or put alternatively, one cannot communicate every-
thing about ones intention to the other in the terms the latter fully
The interpretative asymmetry between a civilized society and a
primitive one can be illuminated by a disparity of another kind. A
civilized society normally grew out of the process of contact, clash/war,
amalgamation among dierent primitive tribes, and outward expan-
sion then into a similar process on a larger scale. So its language
games must also have been under a dynamic process of interaction
and adjustment among its originally heterogeneous components, often
yielding ever higher-order integrations. When such integrations reach
a certain level, the members whose mentality has been shaped by
the norms of the new games will have interpretative resources rich
enough to simulate, approximate, or grasp those highly unique or
specialized parts of the language game of a total stranger from a
primitive tribe. On the contrary, a primitive tribe member can hardly
develop by himself sucient doxastic or conceptual resources to
understand a total stranger, no matter how smart or attentive he or
she may be.
By the same token, two initially isolated civilizations of a similar
sophistication can in principle accommodate each others language
games, at least covering enough aspects or levels of them, once they
encounter one another. This is exactly the point Davidson would
like, and I think correctly, to stress. But the interesting and more
subtle question is whether or not such a cross-cultural interpretation
is necessarily extendable to all aspects and levels of the games on
the other side. That is clearly a question which can hardly be decided
a priori by the principle of charity.
Here are two reasons that make me skeptical about a positive
answer to the above question. First, if there could be an interpre-
tative rupture between two relatively simple or primitive language
. r.\irsoxi.x +o xonv.+i\i+v 201
games, it would be equally hard to exclude the possibility of such a
rupture between two greater language games each of which involves
one rupture element, i.e., one primitive game, as its component,
and never had contact with the other rupture element belonging
to the other greater game. (The sense of rupture here is relative
to the host greater game.) The relationship between components and
the whole game could be so varying as to include the possibility that
the role of one component can neither be replaced nor implied by
any (combination of ) others, despite of the presumable large degree
of integration among most components of the whole. A possible anal-
ogy in a dierent domain is this: an animal species whose genetic
structure lacks certain DNA strings may not be able to digest some
kind of food which is perfectly edible to other species having such
DNAs. Here I certainly cant get into a detailed, and probably fruit-
ful, discussion about certain implicit questions such as how large a
component has to be in order to have such an indispensable and
irreplaceable role, especially in the domain of intentional items. One
thing seems to be clear: correct interpretation requires something
more than merely apparent coherence (as coherence seems to allow
degrees, e.g., one might apparently swallow up the inedible food),
something in the direction of real causal history of engaging and
integrating the object (or its components) in various ways.
As I briey suggested above, any successful integration of new
parts into an old game would have changed or shaped the latter in
some necessary, though often imperceptible, way. In this light, there-
fore, whenever we claim that, say, Chinese culture had assimilated
some particular heterogeneous cultures (or cultural products, such as
Buddhism from India), we can no longer claim, or even imagine,
that the real, original Chinese culture is something like the present
Chinese culture minus those particular exogenous ones. In other
words, the interpretative power or resource of todays Chinese cul-
ture is no longer the same as (or perhaps even hard to compare to)
that of Chinese culture two thousand years or just two hundred years
ago, confronted with a similar alien culture.
What is, then, the implication of this point for the possibility of
interpretative rupture between two grand cultures indicated above?
The implication is the following. When one grand culture faces a
rupture element in the other, it does not literally face the pure ele-
ment itself, as (the past self of ) the other grand culture probably
would when a primitive tribe was rst encountered. For, on the one
202 cn.r+rn sr\rx
hand, the original primitive game to which the rupture element cor-
responds had been merged into the larger game long before, and
on the other hand, that element, either as an organic component of
the grand culture or as some hidden deposit, is no longer indepen-
dently recognizable without appropriate background interpretation of
a much larger picturethis is exactly what is required by the
Davidsonian holism. Now the complexity lies in the likely scenario
that the interpreter from the external grand culture may not be able
to nd sucient reasons to realize or demonstrate that a rupture is
there even if in fact it is. For, as mentioned above, a grand culture
usually has enough interpretive resources to warrant elbow room for
tenable rationalization of almost any seemingly inconsistency in the
object of interpretation.
This complexity naturally leads to my sec-
ond skeptical reason, which happens to involve a similar complexity.
The construction process of the higher-order theories or super-
structure of a civilization contains, generally speaking, two sorts of
ingredients: the rst sort includes those things that are necessary for
explaining or accommodating peoples rudimentary or observational
experiences, such as perceptions of natural regularities, while the sec-
ond sort those contingent or accidental uctuations or deviations
which seem hardly avoidable, say, in personal creation or interper-
sonal transmission of ideas, visions, styles, etc. Now a complexity
emerges from this second sort of ingredients for ultimately a grand
language game or paradigm. Higher-order norms or principles of a
language game may have easily absorbed or incorporated such acci-
dental ingredients, some of which might be very subtle or manner-
ist, in unnoticeable ways and over a long period of time of cultural
accumulation. But once such norms and principles exist or are expli-
cated (in some distinct forms), they must take eect as respective
wholes without possibly distinguishing those necessary ingredients
from the contingent ones at their origin. Moreover, norms are grad-
ually and constantly reinforced or reied over time and with heavy
usage, and some coupling or mutually supportive eects among related
norms may also increase or amplify. All these tendencies may help
Such a complex and even paradoxical situation is probably best reected in
Davidsons own discussion of his so-called paradox of irrationality. See Davidson,
Paradoxes of Irrationality, Problems of Rationality; also see Zheng, Yujian (2001),
A non-substantial approach to practical reason, International Journal for Field Being
(special edition on Whitehead).
. r.\irsoxi.x +o xonv.+i\i+v 203
account for the fact that some originally accidental bifurcations or
forks in personal ideas, thoughts, or imaginations ultimately turned
out to be parts of the hard cores of certain unshakable orthodox
or paradigms in dierent cultures. And from our earlier discussion,
we should add that these more or less integrated normative forces
are causally responsible for constituting as well as regulating the com-
munity members webs of beliefs, or at least many structural aspects
of such webs. No matter whether these initial thought bifurcations
contain any errors or seeds for errors, the holistic paradigms par-
tially brought about by them could become the obstacles, or worse
still, the unconscious pitfalls for cross-cultural interpretations.
I am quite convinced by Davidson of the impossibility of whole-
sale rupture (or incommensurability) between conceptual schemes or
paradigms. However, if my skeptical reasons above hold any water,
i.e., if some temporary, local ruptures of interpretation are inevitable
between language games even of great scale and sophistication, then
we should conclude that the Davidsonian argument against relativism
cannot be applied to the local levels of interpretation. Or, following
the Davidsonian style, one might say that the local failure or incom-
petence of Davidsons theory can make real sense only against the
background or overall validity of his approach to normativity.
A.P. Martinich
I want to discuss two pairs of terms that are verbally identical, that
is, are homonyms. But the two pairs seem to be semantically uncon-
nected. I will suggest that they are in fact semantically connected,
and in ways that I think have not been noticed before. The rst
pair of terms are the meaning (hereafter: c-meaning) of a text and the
corresponding kind of interpretation appropriate to such meaning.
Roughly, c-meaning is one half or aspect of a successful communi-
cation; and hearer interpretation is the other half. The second pair
of terms are the meaning (hereafter: s-meaning) of some thing, state
or event and the corresponding kind of interpretation appropriate
to that kind of meaning.
Roughly, the correct interpretation of some
thing is the knowledge of that things importance.
The reason that these two pairs of meaning and interpretation do
not seem logically connected with each other, notwithstanding their
verbal similarity, is roughly the following: The c-meaning of a text
expresses a peculiar kind of abstract, hypothetical entity,
the communicative meaning of the text; and it depends upon a dis-
tinctively invented entity, a language, which consists of a nite set
of rules that generate (or describe) an innite number of sentences.
In contrast, s-meaning does not essentially express the communica-
tive meaning of anything and does not depend on a language in the
way that c-meaning does.
It is not unusual for a nonlinguistic entity
to have an s-meaning, for example, the Long March or the Cultural
Revolution. Although the s-meaning can only be represented or
I dont use the c- and s- prexes for the two senses of interpretation because
I shall eventually suggest that clear cases of interpretations of c-meaning involve
For ease of exposition, I am pretending that meanings exist.
If Donald Davidson is right to hold that all thought requires language, there
is a sense in which s-meaning requires language.
208 cn.r+rn rion+
expressed with words, the words are not the thing that has the
That there is a dierence between the sense of c-meaning and
the sense of s-meaning is indicated by the fact that for the former,
statement (1):
(1) By It is raining, the speaker meant that it is raining, but it was
not raining.
far from being odd, can be true, because people sometimes say some-
thing false. In short, the fact that a person (or sentence) means some-
thing does not guarantee that the person or sentence is correctly
describing that thing. In contrast, statement (2),
(2) The meaning of the Long March is that Mao Tse-Tungs army
would survive the attacks of the Nationalist Army; but Mao Tse-Tungs
army would not survive the attacks of the Nationalist Army,
is necessarily false, because either the rst conjunct is false or the
second conjunct is false. In short, the truth of an explicit statement
of s-meaning (a sentence of the form The meaning of x is that p)
guarantees that it is true that p. (Of course, the converse does not
hold. The fact that it is true that p does not guarantee the truth of
a statement of the form The meaning of x is that p.)
The contrast I am drawing corresponds to the contrast in Chinese
Case 1: zhe-ge ju-zi shi shen-me yi-si
(This one) (sentence) (is) (what) (meaning)
Case 2: wo xiang-yao tao-lun lun-yu de yi-si [or, yi-yi ]
(I) (want to) (discuss) (the Analects) (of ) (meaning)
Theres a dierence between having an interpretation and giving an interpre-
tation. In giving an interpretation, the speaker must accommodate what he says to
the beliefs of his audience. For the most part, I will be talking about giving an
Alternatively: The meaning of the Long March is that Mao Tse-Tungs army
would not survive the attacks of the Nationalist Army; but Mao Tse-Tungs army
would survive the attacks of the Nationalist Army.
S-meaning is related to H.P. Grices n-meaning (Meaning, in The Philosophy
of Language 4th ed., ed. A.P. Martinich, New York: Oxford University Press, 2001,
pp. 9297, but I am not sure exactly how. The central cases of s-meaning involve
the signicance or importance of something insofar as it relates to a complex of
other things, and depends upon the interests of human beings. The central cases
of n-meaning do not need to be matters of importance and may be independent
of the interests of human beings, e.g., Those clouds mean rain.
ox +vo kixrs or vr.xixo .xr ix+rnrnr+.+iox 209
Before beginning my primary project in this section, the explanation
of what is involved in interpreting c-meaning, a preliminary matter
needs to be discussed. Is there is any important dierence between
the understanding of c-meaning and an interpretation of it? Donald
Davidson, I believe, holds that there is no dierence; and that is my
view as well. His view, as expressed in A Nice Derangement of
Epitaphs, was criticized by Ian Hacking on the ground that under-
standing and interpretation are signicantly dierent. According to
Hacking, for most ordinary utterances, a hearer just understands
them and does not interpret them: The vast majority of things we
say to our peers in ordinary conversation are not interpreted at all.
Now it is certainly true that there are one or more senses of inter-
pretation, according to which interpreting is not merely understanding
what was c-meant.
But what Hacking needed to show was that
Davidson was not using interpretation in any reasonable sense. And
that, I think, is false. There is a sense of interpretation that means
linguistic understanding. To think that there is not is to confuse a
dierence in usage with a dierence of meaning (a dierence in the
cognitive content of the words). The words, ocer and cop (in
one of their senses) mean the same thing, namely, policeman: x is
an ocer if and only if x is a cop. But they are used in dierent
situations. Ocer tends to be used when addressing a policeman
and is slightly formal; and cop tends to be used outside the hear-
ing of a policeman and is informal.
The dierence between mean-
ing and usage could be illustrated with numerous other examples,
for example, mother and mama and father and papa.
As regards c-meaning, the word understand
is typically used
when the speaker
is condent that he has understood a text cor-
rectly and believes that his audience will not disagree with him. The
word interpret is typically used when one of these conditions is
Ian Hacking (1986), The Parody of Conversation, in Truth and Interpretation,
ed. Ernest Le Pore, Oxford: Blackwell, 1986, p. 451.
Hacking, p. 450.
A third word, pig has the same meaning as policeman but is colloquial and
In this and other places, I might have added the phrase, and its cognates.
For the sake of simplicity, I do not.
The speaker had been a reader of the text.
210 cn.r+rn rion+
absent. There is a distinction between the (cognitive) meaning of a
word and its usage.
One might object that the dierence between understanding and
interpretation is that interpretation applies only to complex texts. But
that is not the case. A familiar sign in many parks is Keep o the
grass. But suppose at the bottom of the sign in small letters are the
words: By order of the health department. One interpretation of
this sign is that one should stay o the turf, perhaps because it has
been sprayed with a pesticide. But another interpretation is that
one should not use marijuana since grass is a colloquial term for
In many poems, some of the sentences are easy and some of them
are dicult to understand. One might urge that only the latter sen-
tences require interpretation. But notice that in the rst sentence of
this paragraph, the word understand applied as naturally to the
sentences that supposedly require interpretation as to those that do
not. That understand and interpret mean the same thing can be
illustrated by actual examples of usage in which the author uses both
words about the same text.
That obscure phrase [high crimes and misdemeanors] requires inter-
pretation: we must ask which understanding of it ts best with the most
persuasive overall structure of our constitutional arrangement.
Hacking might object to the linguistic evidence I have been men-
tioning by observing that all that it shows is that some or all cases
of interpretation are cases of understanding, but not the converse.
That objection is strictly correct. But, notice that to make it is to
concede that there is at least a large overlap between understand-
ing and interpretation. So Davidsons use of interpretation to mean
(linguistic) understanding is not far o from ordinary usage, if it is
o at all.
In any case, Davidson could forestall such disputes by saying that
he was using interpretation as a technical term in a sense broad
enough to encompass simple and complex, easy and dicult, cases
of identifying the c-meaning of a text. In addition to rendering objec-
tions based upon some ordinary uses of the word ultimately irrele-
vant, what would be important about this position is that Davidson
Ronald Dworkin, Philosophy and Monica Lewinsky, New York Review of Books,
March 9, 2000; my italics.
ox +vo kixrs or vr.xixo .xr ix+rnrnr+.+iox 211
is giving an account that applies equally to cases of (linguistic) under-
standing or interpretation. That is, he has revealed the underlying
unity of the allegedly dual phenomena.
It now remains for me to show that the same elements required
for understanding or interpreting dicult texts are also required for
understanding easy texts. Davidson in eect gave an elegant proof
of this in Belief and the Basis of Meaning.
I will give an infor-
mal proof by indicating, too briey, that there is a continuum of
cases of linguistic understanding (or interpretation), in which both
kinds of text involve more than just a hearers judgment about the
meaning of the speakers words or about what the speaker meant,
and that the dierence between the two kinds of cases involves the
number or amount of nonlinguistic judgments that are required and
not the kind of requirements for understanding it.
To begin, lets suppose that Mr. Speaker is talking to Ms. Hearer,
who is standing in the doorway of her house. Speaker is standing
on the walkway several feet away. Holding his hand out, Speaker
says, Its raining. Hearer sees some, though not a lot of, water
falling from some distance above Speakers head. It is natural (but
mistaken) to think that Hearers understanding consists of nothing
more than the belief that Speaker said that it is raining. For one thing,
Speaker might be joking or conceivably speaking a language other
than English. So more is required than simply attributing a belief
to the speaker. In the situation just described, Hearer plausibly needs
at least the following beliefs, some of which she previously had and
are merely activated for this occasion and some that are newly
acquired for this occasion:
Interpretation A:
(A-1) Its raining means that it is raining.
(A-2) In uttering, Its raining, Speaker means that it is raining.
(A-3) Speaker believes that it is raining.
(A-4) It is raining.
Notice that the interpretation of the situation involves (1), a belief
about the linguistic part of the world, (2), a belief about what the
speaker meant, (3), a belief about the psychological part of the world,
and (4) a belief about the nonpsychological part of the world.
In The Philosophy of Language, ed. Martinich, pp. 46472
212 cn.r+rn rion+
It is natural to think that an interpreter always comes to these
beliefs in some linear order. One might think that the order of under-
standing is: (A-1), (A-2), (A-3), and (A-4). Or one might think that
they come independently of each other: (A-1) from Hearers knowledge
of English, (A-2) from Hearers belief about Speakers knowledge,
(A-3) from her psychology, and (A-4) coming from observation. I
think both of these beliefs are incorrect.
My guess is that there is no
regular temporal ordering of all four kinds of beliefs.
In addition
to the possible temporal sequence of (A-1) preceding (A-2), (A-2) pre-
ceding (A-3) and (A-3) preceding (A-4) in the process of coming to
the interpretation, it is quite easy to see how Hearers seeing of the
falling water may have led her (tentatively) to believe (A-4), and that
may have helped her to understand and hence to accept (A-3) and
(A-1) and (A-2) in that order. If the ordering of (A-3) and (A-1) prior
to (A-2) seems odd, this may help make it plausible. Hearers belief
in (A-3) and (A-1) may have led her to think that Speaker uttered,
It is raining, and not what Hearer otherwise would have thought
Speaker uttered, namely, Its Rainey [a persons name],
if the Speaker does not know anyone named Rainey.
The psychological process of arriving at interpretation A might
have included other propositions considered but rejected, such as
The following passage from Leo Strauss can be interpreted as asserting the
temporal priority of understanding context, including the genre of the work, to
interpretation: The context in which a statement occurs, and the literary charac-
ter of the whole work as well as its plan, must be perfectly understood before an
interpretation of the statement can reasonably claim to be adequate or even cor-
rect (Persecution and the Art of Writing, Glencoe, IL: The Free Press, p. 30).
Within biblical scholarship, some literary historians claimed that one must
determine the original text of a book before it can be interpreted. A redaction the-
orist might hold that redactors contribution to a text needs to be identied before
it can be interpreted. But, as Barton says, the process of understanding is not so
simply linear ( John Barton, Reading the Old Testament, Philadelphia, The Westminster
Press, 1984, p. 70; see also p. 65).
Since I am hard of hearing, I often rely on what I observe about the non-
linguistic context or know about the preceding linguistic context to largely deter-
mine what words I think I hear and what the speaker means and believes. People
with better hearing and worse eye-sight might rely more on what they hear to
determine what they think they see.
Notice that I might have but did not include Hearers belief about the words
Speaker uttered. It is possible that Speaker said, Israneys here but meant Its
raining and Hearer appropriately corrected Speakers words for the purpose of
understanding what he meant.
ox +vo kixrs or vr.xixo .xr ix+rnrnr+.+iox 213
(B-4) It is not raining [the descending water has a shape not charac-
teristic of rain],
(B-3) Speaker believes that it is not raining (but Speaker is kidding
(B-2) In uttering, Its raining, Speaker means that it is not raining.
Why none of (B-2), (B-3) or (B-4) were accepted is a complex ques-
tion. Briey and roughly, they t less well with the complex Network
of Beliefs with which Hearer came to the interpretive situation and
the sensory input.
Once a person is comfortable with a set of beliefs that serve as
her understanding of the utterance (and the situation more broadly),
the beliefs are accepted as a group (for practical purposes), because
they are mutually supporting. While Hearer may come to believe
(A-2) partially on the basis of believing (A-1), (A-3), and (A-4), it is
equally true that Speaker may come to believe (A-4) partially on the
basis of (A-1), (A-2), and (A-3). Hearers belief that Speaker believes
that it is raining and means that it is raining gives Hearer good rea-
son to believe that it is raining, ceteris paribus. Similar comments apply
to (A-2) and (A-3).
In a simple situation like the one being considered, the brain
processes are so quick that there is no perceptible interval between
hearing and understanding. In more complex situations, there is often
a perceptible time lag, as the person tries to settle on a set of beliefs
that adequately handles the facts. (Hacking thinks of the latter cases
as the only genuine cases of interpretation.)
One might think that information about the nonlinguistic and
nonpsychological world are not relevant to an interpretation. To see
the relevance of (A-4) or some other nonpsychological belief about
the world to the interpretation of Its raining, consider a slightly
dierent situation. Suppose that Speaker had said, It is not rain-
ing. A sensible understanding of this situation is the following:
Interpretation B
(B-1) It is not raining, means that it is not raining.
(B-2) In uttering, It is not raining, Speaker means that it is not raining.
(B-3) Speaker believes that it is not raining.
(B-4) It is not raining.
214 cn.r+rn rion+
In this case, (B-4) is supported by (B-1), (B-2), and (B-3).
As the situation for Interpretation B has been described so far, I
have not said whether anything in Speakers environment suggests
that it is raining or not. (Lets not count Speakers words as part of
the environment.) There are two basic cases to consider. One is that
no water is or seems to be falling on Speaker. In this case, the envi-
ronment gives further support to (B-4), not to mention (B-3). The
other case is that Hearer sees water falling on Speaker. This gives
strong though not conclusive evidence that (B-3) is false and that
(A-4) is true. But in this case (B-2) and (B-3) become doubtful (more
doubtful than (B-1) becomes). And Hearer has various courses of
action available to her. She may change her beliefs with respect to
(B-3) and (B-1). Alternatively, she may act to gather more informa-
tion. By walking out of her doorway, Hearer may discover that
Speaker is standing near a water sprinkler that Hearer could not see
from the doorway. This would account for the falling water.
Each of these changes would probably result in Hearer having a
dierent understanding or interpretation of what Speaker meant.
Let me change the text to indicate perhaps more clearly how judg-
ments about the meaning of the text and judgments about the non-
linguistic facts interact. Suppose the text is GODISNOWHERE.
Two possible interpretations are C and D.
Interpretation C:
(C-1) GODISNOWHERE should be understood as consisting of the
following four words: God, is, now, and here.
(C-2) GODISNOWHERE means that God is now here.
(C-3) By GODISNOWHERE, the author means that God is now
(C-4) The author is a theist.
Interpretation D:
(D-1) GODISNOWHERE should be understood as consisting of the
following three words: God, is, nowhere.
(D-2) GODISNOWHERE means that God is nowhere.
(D-3) By GODISNOWHERE, the author means that God is nowhere.
(D-4) The author is an atheist.
It would be easy to change the facts slightly to get an interpreta-
tion that consisted of (D-1), (D-2), (D-3), and (C-4), that is, an inter-
pretation that a theist seriously asserted that God is nowhere (because
he exists outside of space and time); and another interpretation that
included none of (C-1)(D-4). (For the latter, suppose that the lan-
guage is similar to English but does not use a copula, that a per-
ox +vo kixrs or vr.xixo .xr ix+rnrnr+.+iox 215
son is named Godis is being referred to, that he is being said to
be present at the time of speaking, and the speaker is an agnostic.)
Examples like GODISNOWHERE are relatively frequent in
Chinese, for example,
(a) jin-nian-hao-dao-mei-shao-bu-de-da-guan-si
which could mean
(a-1) jin-nian-hao-dao-mei shao-bu-de-da-guan-si
(This year has bad luck, and it is unavoidable to have a lawsuit.)
(a-2) jin-nian-hao dao-mei-shao bu-de-da-guan-si
(This has good luck, and there are no bad things, and there will be
no lawsuit.)
If we were to consider an interpretation of a long poem, an inter-
pretation would have to bring in many more nonlinguistic facts in
order to give an interpretation of what the poet or the poem meant,
because so much of the c-meaning of a poem is inferred from judg-
ments about propositions that are not closely related to the mean-
ings of the actual words used. Suppose the poet is Russian, or lived
ve hundred years ago and hence had very dierent beliefs from us.
Nonetheless, the requirements for understanding the poem would be
no dierent from understanding the sentence, It is raining.
Lets now consider an example that approximates to the com-
plexity in scholarly interpretation. In such interpretations, it becomes
necessary for the interpreter to make explicit propositions of a kind
that might have been left implicit in simpler examples. I am now
interested primarily in interpretations that are given or presented to
another person, in contrast with interpretations that a hearer merely
has. Often, the meaning of only a few words is made explicit. The
rest, along with other information is merely presupposed. Interpretations
that are given by scholars focus on the points of actual or most likely
disagreement. This is illustrated by two examples of interpretations
of a sonnet by Gerard Manley Hopkins:
It is relatively complicated to explain when one is concerned with what the
poet meant and when one is concerned with what the poem meant. And I pass
over that issue in the main text of this paper.
216 cn.r+rn rion+
Spring and Fall: To a Young Child
Margaret, are you grieving
Over golden grove unleaving?
Leaves, like the things of man, you
With your fresh thoughts care for, can you?
Ah! as the heart grows older
It will come to such sights colder
By and by, nor spare a sigh
Though worlds of wanwood leafmeal lie;
And yet you will weep and know why.
Now no matter, child, the name:
Sorrows springs are the same.
Nor mouth heard of, ghost guessed:
It is the blight man was born for,
It is Margaret you mourn for.
Interpretation I
This poem, Spring and Fall: To a Young Child, is about the abil-
ity of a young child to feel emotions naturally generated by things of
nature, and the inability of a middle aged person to feel those same
emotions. In middle age, the emotional readiness of the child to react
to nature is replaced by an intellectual response. In middle age, the
adult will be able to understand the signicance of Fall, but not be
able to feel it. This interpretation is supported by the title of the poem.
If the normal life span of a human being is 70, as it is proverbially
held to be, and if it is divided into four roughly equal parts, then the
Spring of a person spans the rst eighteen years of a persons life. The
child Margaret addressed in this poem is certainly within that span,
probably somewhere between the ages of ve and ten; so she is in the
Spring of her life. The Fall of a human life would then be between
the ages of thirty-one and forty-ve. This period, which is often referred
to as middle age, is in fact part of the downward slope of a per-
sons life. However, the Fall of a human being is also the period of
her intellectual height. Her undiminished intellectual power and her
experience would give her the wherewithal to make wise judgments.
Reason would predominate over emotion.
The poem begins by indicating that the young girl is grieving over
the falling leaves of autumn. What makes her sad is nothing that she
can rationally comprehend; her fresh thoughts are incapable of ratio-
nally comprehending the signicance of the falling leaves.
When her heart grows older, that is, when she is old enough to
be in the Fall of her life, she will not be able to react emotionally to
the falling leavesIt will come to such sights colderhowever, she
will understand their signicance, as she could not understand it when
she was young.
ox +vo kixrs or vr.xixo .xr ix+rnrnr+.+iox 217
What she will understand is that the falling leaves signify her
own death: It is the blight man was born for, / It is Margaret you
mourn for.
It is not necessary for me to enumerate each proposition of the inter-
pretation and to comment on its type. It is obvious which clauses
or sentences say something about the meaning of the words or what
the poet meant and which say something about the psychology of
the poet or about the world more generally. The sentence The
poem begins by indicating that the young girl is grieving over the
falling leaves of autumn, is semantically similar to the sentences of
our simple interpretations that stated utterance meaning. If greater
explicitness were required, the interpreter could easily have written.
(I-1) Margaret, are you grieving over golden grove unleaving means
that the speaker is asking the addressee Margaret whether she is griev-
ing over the falling leaves of golden grove.
(I-2) By Margaret, are you grieving over golden grove unleaving
Hopkins means that the speaker is asking the addressee Margaret
whether she is grieving over the falling leaves of golden grove.
The interpreter assumes that sentences like (I-1) and (I-2) are unprob-
lematic and hence do not need to be set out explicitly.
The sentences that the interpreter does usewhat would natu-
rally be called an interpretation of the poempresupposes sentences
like (I-1) and (I-2). The focus of Interpretation I are the facts that
the interpreter believes Hopkins is using and that the reader may
not know. The interpretation also claries implications of the poem:
(I-3) What makes Margaret sad is something that she can feel but not
rationally comprehend.
(I-4) When Margaret is old enough, in the Fall of her life, she will
not be able to react emotionally to the falling leaves, but will under-
stand her sadness.
(I-5) Margaret will realize that she is sad about her approaching death.
About half of the rst paragraph provides factual information that
the interpreter thinks is relevant to the interpretation. The last para-
graph, for its part, combines explications of meaning, factual infor-
mation, and further implications intended by Hopkins. The opening
clause, Margaret, are you grieving? and the nal clause, It is
Margaret you mourn for, are hardly more dicult than It is raining.
The intervening clauses vary in diculty. In short, the interpretation
218 cn.r+rn rion+
or understanding of this sonnet is on a continuum with the short
texts we analyzed earlier.
Lets now consider another interpretation of Hopkinss poem in
order to show that the factual information that may justiably be
used in interpretation may not be obvious from the language of the
poem itself. The following interpretation builds on interpretation I,
but then goes beyond it. In short, we are moving further along the
continuum of interpretation in the direction of complexity and
Interpretation J
At one level, the poem, Spring and Fall: To a Small Child, is about
the ability of a young person to feel emotions naturally generated by
things of nature, and the inability of a middle aged person to feel
those same emotions. In middle age, the emotional readiness of the
child to react to nature is replaced by an intellectual response. In mid-
dle age, the child will be able to understand the signicance of the
fall, but not be able to feel it.
One could correctly observe that the normal life span of a human
being is 70, that it can be divided into four roughly equal parts, of
which the Spring would consist of the rst eighteen years of a per-
sons life and the Fall to the years between thirty-one and forty-ve.
This so-called Fall of a human life would coincide with the intel-
lectual height of most people. Their undiminished mental powers and
experience would enable them to make sober judgments. Reason would
predominate over emotion.
But there is much more to the poem than this. Gerard Manley
Hopkins was a Jesuit priest, and we know from biographies and his
letters that he was concerned about the consequences of original sin,
the Christian belief that all human beings come into the world as sin-
ners, and as such, are destined for Hell, unless something or some-
body redeems them. Thus, the word Fall in the title refers to the
Fall of Mankind or the Fall of Adam and Eve. And one of the
principal consequences of original sin is death. The word Spring
refers not so much to the season of the year as to the metaphorical
jumping of a human being into the world. The idea of a somewhat
uncontrolled entrance to the world is central to the thought of some
modern religious thinkers. Soren Kierkegaard talked about a leap of
faith; and. Martin Heidegger, though he ended life as an atheist, stud-
ied for the priesthood as a young man and never lost a kind of mys-
tical understanding of human existence.
If we take Hopkinss interest in original sin seriously, then we will
ox +vo kixrs or vr.xixo .xr ix+rnrnr+.+iox 219
see that while the young girls grieving is emotionally appropriate,
it is not intelligible to her. She does not understand it. The answer to
the rhetorical question, Can you with your fresh thoughts care about
the things of man?, is that of course she cannot. She lacks knowl-
edge about original sin, the corrupted world, and about her own death.
However, her ghost [that is, soul] guessed. As the years go by, she
will lose her innocent emotional response and have them replaced
by an understanding of these things, and also a realization of a self-
centered worry, that what is saddest to her is her own death, It
is Margaret you mourn for.
In Interpretation J, the interpreter is providing information that she
thinks the reader may not know. But this is information that Hopkins
knew about himself at least to some extent and could have known
completely with some eort.
Hopkins wrote his poetry in English a century ago, and there is
a lot of information about England at the time and about his life
and attitudes.
II S-meaning
At some point in the attempt to understand a poem or other text
of some importance a hearers or a readers search for understand-
ing or for an interpretation passes beyond a search for c-meaning
to something else. Interpretation J contains something of this in virtue
of the connection it makes between the poem, the psychology of
Hopkins and Roman Catholic beliefs. But lets consider an example
where this is more explicit. The interpretation is a seventeenth-
century account of Mary Rolandson, an American woman kidnapped
and held captive by Indians. I will rst give a summary of the account
with segments numbered for ease of reference. The italicized parts
stand for elements that are obviously Rowlandsons opinion about
the facts. These are discussed in the interpretation.
A Narrative of the Captivity and Restauration of Mrs. Mary Rowlandson
(1) Indians attacked our settlement Lancaster, Massachusetts in February
1675. They burned several houses and killed many. Several settlers
were stripped naked before being killed by the Indians. One who
oered the Indians money for his life if they would spare him was
220 cn.r+rn rion+
stripped and disemboweled. Thus they went on burning and destroy-
ing before them. The Indians broke into my house and killed or injured
most of the inhabitants, including one of my children, my sister, and
her children. My dogs, which in other circumstances would have
attacked the Indians hid on this occasion.
(2) God did this to make us the more to acknowledge his hand and to see that
our help is always in him.
(3) The Indians laid hold of us and said, Come go along with us;
I told them they would kill me: they answered, that if I were willing
to go along with them, they would not hurt me.
(4) God by his almighty power preserved a number of us from death.
(5) The Indians took me and twenty three other settlers away. Being
totally exhausted, the next day I rode with my child on a horse behind
an Indian.
(6) God was punishing me for my sins but also showing his mercy towards me.
(7) My daughter died nine days later.
(8) The Indians buried her. An Indian oered me a Bible, and said
the other Indians would let me read it. I often took out my Bible and
got consolation from reading it
(9) Tired of a pregnant European womans complaints, the Indians
stripped her naked, sang and danced around her; then, smashed in
her head and that of a child she held, and threw her in a re.
(10) The Indians scavenged in a eld that the English settlers had
harvested. I got two pieces of corn; one was stolen from me. An Indian
gave me some horse liver to eat. I put it on the coals to roast. I ate
half, and half was taken by other Indians.
I broke down and cried in front of the Indians.
One asked me, why I wept, I could hardly tell what to say; yet I
answered, they would kill me: No, said he, none will hurt you.
Then came one of them and gave me two spoonfuls of meal to com-
fort me, and another gave me half a pint of Peas, which was worth
more than many bushels at another time.
(11) I did various work for several Indians, each of whom paid me
for my work. When I was hungry, a squaw showed herself very kindly
to me and gave me a piece of bear; she also gave me some ground-
nuts. Later another squaw laid a skin for me and gave me some
groundnuts and bade me come again.
(12) The Indian party in which my son was kept captive was attacked
by Mohawks, allies of the French. I am glad that my son was not
captured by them, for it might have been worse for him had he been
sold to the French than it proved to be in his remaining with the
(13) The papoose of my Mistress died, and there was one benet
in it, that there was more room. They buried the papoose the next
day, and afterward morning and evening there came a company to
mourn and howl with my Mistress, though I confess, I could not much
condole with them.
ox +vo kixrs or vr.xixo .xr ix+rnrnr+.+iox 221
(14) It rained and I stayed in a wigwam, while many Indians slept
all night in the rain. Thus the Lord dealt mercifully with me many time, and
I fared better than many of them.
(15) After traveling all day, I laid down my load, and went into the
wigwam, and there sat an Indian boiling of horses feet; they had noth-
ing else. I asked him to give me a little of his broth, or water they
were boiling in. He took a dish and gave me one spoonful of corn-
meal porridge, and bid me take as much of the broth as I would.
After three more days of traveling, I had little spirit left. Chief Philip,
came up, took me by the hand, and said, In two weeks, youll be
free and with your husband. I asked if what he said was true. He
said it was. He let me wash and gave me a mirror to look at myself.
Then he gave me food.
Days later an Indian came to me, bid me come to his wigwam. He
fed me many times, even though he had killed two Englishmen.
Whenever I went to his wigwam, he would always give me something,
even though they were strangers to me. Another squaw gave me fresh
pork. I cannot but remember what a sweet, pleasant and delightful
relish that bit had to me, to this day.
(16) The English thought that if they cut down all the Indian corn
plants, they would starve and die from hunger. And as much corn as
the English found, they destroyed. Yet God preserved the Indians for his
hold ends, and permitted the destruction of many of us. Strangely did the Lord
provide for them, so that I did not see one Indian die from hunger. I can but
stand in admiration to see the wonderful power of God in providing such a vast
number of our enemies in the wilderness.
(17) Another thing that I would observe is the strange providence of God, in
turning things about when the Indian was at the highest and the English at the
(18) They mourned (with their black faces) for their own losses, yet
triumphed and rejoiced in their inhumane and devilish cruelty to the
(19) When the Lord had brought his people to this, that they saw no help in
any thing but himself, then he takes the quarrel into his own hand; and thought
they [the Indians] had made a pit, in their own imaginations, as deep as hell for
the Christians that Summer, yet the Lord hurled themselves into it. And the Lord
had not so many ways to preserve them, but not he hath as many to destroy them.
(20) At rst the Indians were all against my going home.. But after-
wards they assented to it and seemed much to rejoice in it; some asked
me to send them some Bread, others some Tobacco, others shaking
me by the hand, oering me a Hood and Scarf to ride in; not one
moving hand or tongue against it.
(21) Thus hath the Lord answered my poor desire. And now God hath granted
me my desire. O the wonderful power of God that I have seen, and the experi-
ence that I have had. So I took my leave of the Indians.
Thus hath the Lord brought me and mine out of that horrible pit and hath set
us in the midst of tender-hearted and compassionate Christians.
222 cn.r+rn rion+
Interpretation of the S-Meaning of The Narrative of the Captivity
The Narrative of the Captivity and Restauration of Mrs. Mary Row-
landson is the account of an ethnocentric seventeenth-century European
woman, who is blind to the misery and humanity of the Indians because
her mind is darkened by her beliefs in a providential God. Her state-
ment of the facts about what happened to her in (3), (5), (8), and espe-
cially (8), (10), (11), and (15) clearly prove that the Indians were humane
to her. Much of (20) is evidence that they felt kindly toward her. (11)
is evidence that at least some of the Indians acted with justice towards
her; and (5), (8), (11), and (15) indicate that they were often merciful.
Notwithstanding all of this evidence, Rowlandson judges the Indians
harshly and unfairly, as in (18). Oddly, in one way her perceptions of
the Indians are not distorted. She accurately reports that the Indians
fed, protected, and even comforted her. What is distorted are her judg-
ments about their characters. She claims they are of the devils party.
This dissonance between the way she was actually treated and her
account of the nature and cause of that treatment is astounding. The
best explanation of her distorted judgment is that her religious faith
darkened her mind, and in two ways.
One was her negative judgment of the Indians. The other was her
belief that God was the genuine cause of all the good things that hap-
pened to her, for example, in (4), (14), (19), and (21), even though the
Indians were usually the agents of them. On the occasions when she
attributes some harmful action to God, she sees them as means to
justiable ends, e.g. (6). Especially strange is Rowlandsons recognition
that her son was probably better o with the Indians than with the
French (12)!
Certainly, the facts reported in (1), (5), (7), and (9) would justify
some negative judgment about the Indians. But such a judgment would
have to be balanced by the fact that the Europeans were as bad or
worse. The English cut down all the Indian corn plants, so that the
Indians would starve and die from hunger (16). This strategy of war-
fare is indiscriminate cruelty. A large part of the Indians she traveled
with were noncombatants, elders, women, and children.
In short, Rowlandson applies a double standard. In (18), she thinks
it exceedingly strange that Indians mourned the deaths of their fellows
and rejoiced at the death of the Europeans. But didnt the Europeans
behave in the same way? Moreover, Rowlandsons indierence to the
death of her mistresss child described in (13) is callous. Her observa-
tion that the childs death had the benet of providing more room in
the teepee is chilling.
This interpretation includes many statements of Rowlandsons
c-meaning, which presuppose statements about the word and sentence
meaning of much of the text. If, contrary to fact, attacked meant
ox +vo kixrs or vr.xixo .xr ix+rnrnr+.+iox 223
celebrated with in (1) or if Rowlandson meant this, then her hatred
and disgust would be unintelligible. The interpretation also includes
many statements about Rowlandsons attitudes and beliefs. But the
focus of the interpretation concerns the s-meaning of her account.
One further point may be made. In addition to the interpretation
in Interpretation of the S-Meaning of The Captivity, Rowlandsons
own text includes an interpretation of the s-meaning of the events
that happened to her, many of which did not involve language, for
example, the killing, the feeding, and much of the comforting. What
caused Rowlandson to believe, say, (1), (3), and (5), the initial facts
about the raid, were things that she saw, heard and smelled. In
theory, she could haveit would have been crazy to do so, but she
could havethought that what she saw were actors dressed up like
Indians and Englishman pretending to ght and kill each other.
Instead, she came to believe or judge that what she saw were Indians
raiding her settlement. The unitalicized portions of the account are
uncontroversial. It would be natural to say on the basis of them that
she understood that the Indians were attacking, and killing and so on.
But the italicized portions are more controversial. Although some
people would still accept them today, most people would not. And
so although one might say that those segments express part of her
understanding of what happened, one is more inclined to say that
they were part of her interpretation of the events, because they are
unlikely to be accepted by the reader. Far from being at variance
with my theory, calling her understanding of the events her inter-
pretation is consonant with my view that there is no principled
dierence between understanding or interpreting language and under-
standing or interpreting the world.
S-meaning is the signicance or importance of something. Like
cases of c-meaning, cases of s-meaning fall along a spectrum, from
the very simple at one end to the highly complex at the other. At
the simple end, interpretations of s-meaning indicate how one thing
is connected to another thing.
And the knowledge of these things
and their connection makes them intelligible and indicates their
signicance. Suppose a parent walks into the kitchen and sees a
broken glass of milk on the oor. She says,
For the rest of this chapter, I will often use thing to abbreviate thing, state,
or event.
224 cn.r+rn rion+
What is the meaning of this? [In Chinese: zhe she shen-me yi-si.]
She obviously is not asking for c-meaning; no one was trying to
communicate anything. Rather, she wants to know more about the
spill, about how the spilled milk occurred or why. She wants to know
how the spill ts in with other facts. When the parent learns that
her twins had been ghting and knocked over the glass of milk, she
has the understanding (or interpretation) she was looking for:
(S-1) The broken glass of spilled milk s-means that the twins have been
This example is of exactly the same kind, I believe, as these famil-
iar philosophical ones.
(S-2) That smoke means that there is a re.
(S-3) Those spots mean that the patient has measles.
(S-4) The new budget means that ination will remain low.
The s-meaning indicates some importance. (S-2), for example, would
be important to a forest ranger, or to people during a drought,
because any re poses the risk of a catastrophe. (S-3) is signicant
or important to the parent whose child has measles or to a pregnant
woman if she has the spots. (S-4) is signicant to members of the
economy. Importance is always relative to a person or population,
and it may be eeting. (S-1) has some importance for a short time,
namely, to the parent who has often lectured her children about not
ghting, especially at the breakfast table. Although spilled milk is
nothing to cry over, it can be source of exasperation. So it has some
In addition to being relative to a population, signicance is also
relative to the events it is connected with. The spilled milk is signicant
with respect to the twins scue, but not with respect to the budget
decit of Tonga.
In simple cases, the answer to a question of the form, What is
the meaning of X? often has meaning in it: The meaning of life
is to love, or The meaning of life is to have as many satisfying
experiences as possible. In complex cases, something is signicant
because it is connected with many other events; consequently a state-
ment of s-meaning requires many sentences to describe it. In such
H.P. Grice (1957), Meaning, in The Philosophy of Language 4th edition, ed.
A.P. Martinich, New York: Oxford University Press, 2000, pp. 927.
ox +vo kixrs or vr.xixo .xr ix+rnrnr+.+iox 225
cases, it would be dicult to have a single sentence of the form,
The meaning of x is that . . . . Often, the word meaning would
not appear at all in the answer to the question. The answer to the
question, What is the meaning of Mao Tse-Tungs Long March?,
might begin.
In 1934, Maos army was encircled by the Nationalist Army led by
Chiang Kai-Shek. Maos army broke through and marched rst west-
ward and then north during 1935. Maos army continued to be men-
aced by the nationalist Army and also by local bands of warriors. The
successful march of 6,000 miles kept the Chinese Communist move-
ment viable.
The word meaning does not occur.
I now want to point out two ways in which c-meaning and s-mean-
ing and hence the two kinds of interpretation are connected.
First, what I have been saying about s-meaning applies to the
understanding or interpretation of c-meaning, as presented in sec-
tion I. To interpret even a short text, a text of one word or one
sentence requires relating it to other things, to propositions about
the speakers belief, the meanings of the words, and the nonlinguis-
tic world. That is, an interpretation of c-meaning at the same time
is an interpretation of the s-meaning of the linguistic event. The
degree of importance of this s-meaning is often low and eeting, but
it exists during the time that the interpretation is being prepared.
If I am right about this, one might be able to nd cases in which
one speaker would ask a question of another speaker that might
appear to be about the latter speakers c-meaning, but in fact is
about the s-meaning. Here is an example:
Helen-1: Wendy came home from the hospital today.
Pat-1: What do you mean?
Helen-2: She had her baby on Monday.
Pat-2: I didnt know that she was pregnant.
Pat certainly knows what the speaker meaning of the sentence of Helen-
1 is, knows that Helen believes that Wendy came home from the hos-
pital, and believes that Wendy did come home from the hospital today.
But then what is Pat-1 asking about? It cant be c-meaning.
226 cn.r+rn rion+
One abstract possibility is that Pat-1 is asking about a kind of
meaning that we have not talked about. That is, there is a sense of
mean that is equivalent to the sense of intend. But Pat-1 is not
asking what Helen-1s intentions are. He knows that her intention
is to inform him that Wendy came home from the hospital today.
What he does not know is how this information is related to other
facts, so that it makes sense or is intelligible to him. Pat-1 is asking
for is additional information that would allow him to connect what
he knows about the c-meaning of Helen-1 with the rest of his beliefs.
Helen does not respond to Pats question with the insulting ques-
tion, What part of Wendy came home from the hospital dont
you understand? She knows that Pat knows the meaning of those
words and everything else we mentioned about its c-meaning. She
knows that Pat needs background information in order to make sense
of the utterance; and that is what she provides in Helen-2. The seg-
ment ends with Pat-2 indicating why he asked Pat-1.
The correlation between c-meaning and s-meaning does not suggest
that c-meaning is reducible to s-meaning. Rather, s-meaning is a
necessary condition for understanding c-meaning, not identical with it.
The second way in which c-meaning and s-meaning are connected
is that one is often confused for the other or they are conated.
This second connection is more important than the rst because
what is at stake here is the clarity scholars have about what they
are saying or doing. The examples that I could give are numerous.
I will mention only a few here.
One of the chief ideas of Quentin Skinner, the foremost historian
of modern intellectual history and the philosophy of history, is that
what the text of a great philosopher means is what his contempo-
rary interpreters took it to mean. Thus, for example, he holds that
Hobbes held neither a deontological nor divine command theory of
natural law, because his contemporaries did not take him to hold
this. This is a thesis about c-meaning. But he sometimes talks about
this procedure in words that suggest that he is talking about s-mean-
ing: Some [historians] are instead concerned with the provision of
interpretations, and thus with the process of placing texts and other
such objects within the elds of meaning from which their own indi-
vidual meanings can arguably be inferred.
It seems that placing
Quentin Skinner, Visions of Politics vol. I, Cambridge: Cambridge University
Press, 2002, p. 10.
ox +vo kixrs or vr.xixo .xr ix+rnrnr+.+iox 227
one work into the eld or context of other works both explains its
c-meaning and shows its s-meaning.
I want to end this lecture with a speculation about a possible con-
fusion of c-meaning and s-meaning, expressed in terms of texts.
Jacques Derrida is famous in some circles and infamous in others
for saying, There is nothing outside the text. What can this sen-
sibly mean? Or, how might he have come, unwittingly, to this posi-
tion? One possibility is that it results from three aspects of his theory:
(a) understanding a text requires knowing the meanings of words
that are not part of the text; (b) it requires knowing facts about the
world; and (c) there is no sharp line dividing what needs to be known
from what does not need to be known. Therefore, there is nothing
outside the text. This speculation is supported to some extent by this
quotation: What I call text implies all the structures called real,
. . . in short all possible referents.
It is also supported by one of
Derridas sympathetic interpreters: There is nothing outside the text
means There is nothing outside the context.
I think it is important to keep text and context distinct (even
though they are relative terms), and important to keep c-meaning and
s-meaning distinct, but also to understand how they are connected.
Jacques Derrida (1988), Limited, Inc., Evanston: Northwestern University Press.
Keith Jenkins (2000), A Postmodern Reply to Perez Zagorin, History and
Theory 39, p. 189.
I want to thank Linton I-Chi Wang for his help on this paper.
Since a principal motive for this collection of papers is to connect Chinese phi-
losophy with Anglo-American philosophy, I want to express my regret that I have
not been able to nd in standard sources available to me treatments in Chinese
philosophy of meaning and interpretation, for example, Chad Hansens Language and
Logic in Ancient China (Ann Arbor, MI: The University of Michigan Press, 1983),
A.C. Grahams Disputers of the Tao: Philosophical Argument in China (La Salle, IL: Open
Court, 1989), and Rudolf Wagners Language, Ontology, and Political Philosophy in China
(Albany, NY: State University of New York Press, 2003). Also, there is no entry
for interpretation in The Encyclopedia of Chinese Philosophy, ed. Antonio Cua (London:
Routledge, 2003), and the few references for meaning are not relevant. I have not
been able to make sense of the principles of interpretation for the Yi-Jing. I am
open to being informed about these matters from Chinese philosophers.
Kim-chong Chong
The metaphor of the moon reected in water, so often encountered
in Buddhist writings, where it suggests something that cannot be
grasped . . . may be used to refer to the perfection of the highest reli-
gious truths, which are so elusive and mysterious that, regrettably, they
can never be fully apprehended. When, however, the same metaphor
alludes to the oating world or the impermanence of all things, there
is implicit fault-nding with the world for being illusory and unreli-
able. Although the image is the same, the judgment embedded in it
diers, one being praise and the other detraction. (Qian Zhongshu,
Metaphors Have Two Handles and Several Sides).
1. Introduction
In an earlier essay, I have discussed the relation between Davidsons
view of metaphor and a strategic use of metaphor to be found in
the Daoist text, the Zhuang-Zi.
This essay came to somewhat ten-
tative conclusions. In particular, I noted that Davidsons view of
metaphor as having no inherent cognitive content seems to be able
to explain Zhuang Zis metaphorical strategy in steering clear of the
espousal of any value. I also suggested that the liberating eect of
this conscious use of metaphor does not seem to agree with the view
of the cognitive linguists, George Lako and Mark Johnson, that
metaphors are unconscious and embodied ways of perceiving the
I am grateful to Cecilia Wee for her comments on an earlier version of this
Qian Zhongshu (1998), Metaphors Have Two Handles and Several Sides, in
his Limited ViewsEssays on Ideas and Letters, Ronald Egan tr., Cambridge: Harvard
University Asia Center, 1998, p. 122.
See Chong, Kim-chong (2006), Zhuang Zi and the Nature of Metaphor,
Philosophy East and West 56:3.
230 cn.r+rn xixr
In the present essay, I shall do the following. First, I develop fur-
ther thoughts about Davidsons view of metaphor. In one respect,
Davidsons conception of metaphor is too narrow. Davidson stresses
that a metaphor cannot be paraphrased wholly without remainder.
But this conception is unnecessary to Davidsons position, because it
easily leads to an essentialist view of metaphor, one that holds that
there must be an inherent meaning to metaphor. At the same time,
however, Davidson maintains that the way to understand metaphor
is through its use, not through the idea of its meaning. This claim
liberates us from the narrow conception, which includes what I shall
refer to as essentialist conceptions of metaphor. Thus, the second
thing I shall do in this essay is to argue against such essentialist con-
ceptions and their application to the understanding of some aspects
of Chinese philosophy. I have already discussed one such concep-
tion in my earlier essay on the Zhuang-Zi, namely, Robert Allinsons.
While I shall briey repeat what I have said there, the main focus
of the present paper will be a more extended discussion of the Lako
and Johnson model. In particular, I shall discuss the application of
the model by Edward Slingerland to philosophical texts such as the
Zhuang-Zi, the Xun-Zi and the Analects.
2. Davidson on Metaphor
Let us begin with an analysis of Davidons view of metaphor. Davidson
argues that if a metaphor has some special cognitive meaning or
content, then it can be paraphrased without remainder. However,
with the exception of dead metaphors, this is not possible. The con-
clusion to be drawn from this is that a metaphor has no special cog-
nitive meaning or content.
The metaphorical image of the moons
reection on water by Qian Zhongshu in the quotation above helps
to illustrate the point. The same image is used to refer to dierent
things, and to make opposite judgments. On the one hand, it is used
to refer to the teachings of the Buddha as an elusive highest reli-
gious truth. On the other hand, it is used to refer to the illusoriness
of the world. If there is some special cognitive content to the metaphor-
Davidson, Donald (1984), What Metaphors Mean, in his Inquiries into Truth
and Interpretation, Oxford: Clarendon Press.
vr+.rnonic.r tsr \rnsts vr+.rnonic.r rssrxcr 231
ical image, this kind of dual and opposite reference would be con-
tradictory. Given an understanding of Buddhist philosophy and the
dierent contexts in which the same metaphor is used, however,
there is no contradiction. The metaphor serves equally well to refer
to an elusive truth that cannot be fully apprehended, and to the illu-
sory world.
On a strict literary denition of metaphor, it could be objected
that an example like the one just mentioned may not be referred to
as a metaphor. Thus if we are saying that the Buddhas teaching
is like the reection of the moon in water, then this would be a sim-
ile, not a metaphor. A metaphor would intimate something from a
statement of identitythe Buddhas teaching is the reection of the
moon in water. To some extent, Davidson himself observes this dis-
tinction between a simile and a metaphor, and denies any simple
explanation of metaphor in terms of simile. Thus, referring once
more to the example of the image of the moons reection, it could
be held that the metaphor is simply a gurative way of expressing
certain likenesses between the Buddhas teachings and the ungrasp-
able quality of the moons reection, or of the illusoriness of worldly
phenomena. Given such an explanation, there would be nothing spe-
cial about this metaphor, since what it intimates can be expressed
clearly, in terms of certain similarities between two terms. In this
usage, the so-called metaphor provides a gurative image that could
be expressed in terms of a simile.
However, I think that it would be a mistake to be bound by the
picture of metaphor as necessarily something that cannot be para-
phrased without remainder. Let us just say that sometimes, a full
paraphrase may be possible, and sometimes not. As we shall see,
there are various language-games of metaphor. Following Wittgenstein,
we may say that there are family resemblances between various
metaphors, including the ways in which they are used, that cannot
be pinned down to any one essential description. Again, in this
regard, the claim that there is always something left over from a
paraphrase of a metaphor is unnecessaryI suspect that this claim
is precisely what leads to the tendency to think that there is some
special cognitive meaning to a metaphor, as I shall illustrate shortly.
Davidsons narrow conception of what a metaphor is diers some-
what from Wittgensteins view of language and meaning as depend-
ing on context and use. But the core of Davidsons position can do
without the narrow conception. Thus, Davidson states that metaphor
232 cn.r+rn xixr
belongs exclusively to the domain of use. It is something brought o
by the imaginative employment of words and sentences and depends
entirely on the ordinary meanings of those words and hence on the
ordinary meanings of the sentences they comprise.
This emphasis
on metaphor as belonging to the domain of use and the imagina-
tive play of words and sentences not only opens the way to a broader
usage of the term metaphor. It also puts us on guard against any
tendency to view metaphor as having any inherent, special mean-
ing. This is of a piece with the tendency to posit a particular essence
to phenomena of language, a tendency that the later Wittgenstein
argued against. In what follows, I shall examine some attempts to
posit a particular essence to metaphor with reference to examples
from Chinese philosophy.
3. Essentialist Views of Metaphor
We begin with Robert Allinsons claim that the buttery dream in
the Zhuang-Zi has an inherent meaning.
In translation, the original
account of the dream goes like this:
Last night Chuang Chou [Zhuang Zi] dreamed he was a buttery,
spirits soaring he was a buttery . . . and did not know about Chou.
When all of a sudden he awoke, he was Chou with all his wits about
him. He does not know whether he is Chou who dreams he is a
buttery or a buttery who dreams he is Chou. Between Chou and
the buttery there was necessarily a dividing; just this is what is meant
by the transformations of things.
Like Davidson, Allinson is impressed by the fact that a metaphor
cannot be paraphrased without remainder. However, he draws a
dierent conclusion. He deduces, instead, that a metaphor must have
an inherent, intuitive content. And this content can be grasped by
an intuitive cognitive capacity of the human mind.
Ignoring the
latter claim, let us examine the claim that a metaphor must have
Davidson, 1984, p. 247.
Allinson, Robert (1989), Chuang-Tzu for Spiritual Transformation: An Analysis of the
Inner Chapters, Albany: State University of New York Press.
Graham, A.C. (1981), Chuang-tz5: The Seven Inner Chapters, tr., London: George
Allen and Unwin, p. 61.
Allinson, 1989, p. 26.
vr+.rnonic.r tsr \rnsts vr+.rnonic.r rssrxcr 233
an inherent, intuitive content. Thus, Allinson claims that the buttery
dream inherently presents the idea of beauty together with the high-
est positive value that is attached to it, and symbolizes metamor-
phosis from an inferior position to a superior one, from low to high,
and so on.
There is more than one reason for rejecting Allinsons claim. First
of all, it does not follow that if a metaphor cannot be paraphrased
without remainder, then it must be the case that it has some inher-
ent content that has not been propositionally captured. This is one
instance of trying to posit an essence to a phenomenon of language
where perhaps there is none. As we have seen, another conclusion
is possible. That is, as Davidson argues, there is simply no such
inherent or special content. Second, dierent interpretations of the
buttery dream are possible.
Davidson holds as a corollary of reject-
ing the idea of a special cognitive meaning, that a metaphor can
intimate any number of things.
This should not, I think, be taken
as a license to say that any interpretationincluding Allinsonsis
as good as any other. The use of metaphor has its context. As we
have seen, the image of the moons reection in the water is used
and understood in a certain way, within the context of Buddhist phi-
losophy. But we can imagine the image as having another use and
another understanding within, say, the context of the poetry of
Western Romanticism. There is a third reason to reject Allinsons
attribution to the Zhuang-Zi as placing the highest value on beauty.
That is, instead of imposing the highest value upon anything, Zhuang
Zi steers clear of espousing any particular value. We shall see why
he would want to do this and the strategy that he adopts, shortly.
In looking at this strategy and its context, we would be looking at
a particular imaginative use or a language-game of metaphor.
Before embarking upon a description of this strategy, we should
note another account of metaphor as an unconscious mapping or
Allinson, 1989, p. 73.
Mller, Hans-Georg (1999), Zhuang Zis Dream of the ButteryA Daoist
Interpretation, Philosophy East and West 49:4, p. 442, reads the dream in terms of
one of the central themes in the text, the relation between life and death. The state
of being in a dream and the state of being awake are equally authentic. Similarly,
the states of being dead and of being alive. Being in one state, one is unaware of
the other.
Davidson, 1984, p. 262. Davidson says that there is no limit to what a
metaphor calls to our attention . . .
234 cn.r+rn xixr
projection of embodied spatial relations onto the concepts that are
used in our everyday lives. The projection is described in terms of
metaphorical schemas. This is the theory of conceptual metaphor
expounded by George Lako and Mark Johnson in a series of jointly
and separately authored works.
This theory has recently been applied
to Chinese philosophy by Edward Slingerland.
With specic refer-
ence to the Zhuang-Zi, for instance, Slingerland notes that the con-
cept of the Self takes the form of several metaphorical schemas. It
is only necessary to cite a few of these.
For instance, there is the general Subject-Self schema, consisting
of a split between the subject and the self, and where the self can
be represented by a person, object, location, and so on. Other schemas
are said to be special cases of this basic, general schema. Thus, there
is the Locational Self schema, arising from interactions with bounded
spaces and containers. Here, one talks of the self in terms of a con-
tainer in which certain things can be stored. Or else the container
can be one part of the self, such as the heart-mind. As an example,
Slingerland mentions a passage in the Zhuang-Zi that talks of the fast-
ing of the heart-mind, and where the heart-mind is likened to a
stomach that can be made tenuous or empty through metaphorical
fasting. Once the fasting is complete, the only thing left will be the
qi, which is in turn described as being so tenuous a substance that
it has space to receive things and serve as a reservoir for the Way
to gather.
Another schema is the Essential Self schema, where
there is locution of what is essential to the self. This is combined
with the container schema, where certain things are said to be prop-
erly internal to the self, and others are properly external to the self.
Directly after this list, Slingerland states that The power of this
Lako, G. and Johnson, M. (1980), Metaphors We Live By, Chicago: The University
of Chicago Press; Johnson, Mark (1987), The Body in the Mind, Chicago: The University
of Chicago Press; Lako, G. and Johnson, M. (1999), Philosophy in the Flesh, New
York: Basic Books.
Slingerland, Edward (2003), Eortless ActionWu-wei as Conceptual Metaphor and
Spiritual Ideal in Early China, New York: Oxford University Press; Slingerland, Edward
(2004), Conceptions of the Self in the Zhuang-Zi: Conceptual Metaphor Analysis
and Comparative Thought, Philosophy East and West 54:3.
For the metaphorical schemas described in the next paragraph, see Slingerland,
2004, pp. 32830.
Slingerland, 2004, p. 329. For the ancient Chinese, the heart-mind or xin
(sometimes translated as mind or heart) is the organ capable of both cognitive
and aective functions.
vr+.rnonic.r tsr \rnsts vr+.rnonic.r rssrxcr 235
metaphor schema is that it motivates a variety of entailments that
have crucial soteriological signicance and yet that can be accepted
without need for justication or argument by anyone familiar with
the use of containers.
Slingerland is right that these locutions of the Self are present
in the Zhuang-Zi. However, this does not go far enough to help us
to understand how Zhuang Zi uses metaphors in the text. Take the
fasting of the heart-mind, for instance. This is correctly described as
an aspect of the self that is referred to in terms of a container. Under
this schema, the heart-mind can be said to be either empty or full.
A further question can still be asked, namely, what motivates this
locution in the Zhuang-Zi? According to the model of conceptual
metaphor adopted by Slingerland, there is no conscious motivation.
Instead, the motivation is the result of an unconscious mapping of
our sensori-motor experiences with locations and containers in every-
day life (source domain) onto the linguistic and conceptual (target)
domain. Under the theory, metaphors are non-propositional, in the
sense of a causal mechanism of projection. This explanation, how-
ever, does not mention the creative employment of a metaphorical
strategy in the Zhuang-Zi. Insofar as the theory of conceptual metaphor
fails to mention this strategic use of metaphor in the Zhuang-Zi, it
does not do justice to its philosophical aim and content. An appre-
ciation of this strategy would make it hard to accept Slingerlands
claim that the signicance of the heart-mind can be accepted with-
out need for justication or argument by anyone familiar with the
use of containers. Zhuang Zi refers to people who fail to empty
their heart-minds of distinctions. And not everyone familiar with the
use of containers would be able to come up with the creative use
of metaphor in argument, in the way that Zhuang Zi does.
Slingerland, 2004, p. 331. The list of things said to be internal and external
to the self is not essential to the present discussion. For the internal, we have
Heaven, qi, spirit, Virtue, true self, the numinous. For the external, we have (the)
Human, full heart-mind, knowledge or scheming, fame or achievements, cultural
standards, likes and dislikes, life and death, the (political) world, the physical form.
The list as it stands may not be intelligible without some explanation, but any
attempt to explain these terms would take us away from the issue at hand.
236 cn.r+rn xixr
4. The Metaphorical Strategy of the Zhuang-Zi
In the text, Zhuang Zi mentions the endless disputes of philosophers
such as the Confucians and the Mohists, and the restless states of
their heart-minds (xin ).
Zhuang Zis aim is to attain clarity
(ming ) of the heart-mind that involves being detached from these
disputes. As part of this strategic aim, he never directly criticizes the
Confucians and the Mohists. Instead, he uses the voice of charac-
ters like Nanguo Ziqi who describes the disputants as restlessly ghting
with their heart-minds. The disputes are likened to sounds that erupt
in the hollows when the wind blows, or to the peeps of birds. Although
these sounds arise and stop spontaneously, the disputes of men can-
not cease because they have a xated heart-mind (cheng xin). In other
words, they are dead set against allowing other perspectives.
Despite this strategy of speaking through another voice, however,
Zhuang Zi is conscious of the logical status of his own words. That
is, if the words of the philosophers are no better than the sounds of
the hollows or the peeps of birds, what is the status of Zhuang Zis
own words? And isnt Zhuang Zi similarly caught up in emotional
entanglement? In response to this, Zhuang Zi adopts at least two
metaphorical strategies: First, he conceives of the heart-mind as a
mirror or still water that (ideally) reects but does not store what-
ever comes before it. Second, he applies a certain empty structure
to his words by using what has been translated as goblet words (zhi
yan ). Both strategies are related, with the common goal of main-
taining the clarity of the heart-mind.
This conception of the clarity of the heart-mind in terms of the
metaphor of the mirror dierentiates it from the clarity of discursive
discourse. The aim of Zhuang Zis discussion is to clear the heart-
mind of disturbing impurities, namely, the storage of distinctions. It
is not to heighten the heart-minds capacity to make distinctions, nor
on that basis, to state any particular doctrine. The project, in other
words, is not one of critical inquiry but of stilling the heart-mind.
In the following paragraphs that describe Zhuang Zis metaphorical strategy,
I have borrowed from a more detailed and sinological account that I have given
in (2006), Zhuang Zi and the Nature of Metaphor, Philosophy East and West 56:3.
The account is mainly of the second chapter of the rst seven chapters of the
Zhuang-Zi, usually referred to as The Inner Chapters, and generally believed by
scholars to be written by Zhuang Zi himself.
The stillness or clarity of the heart-mind is elsewhere described as being in
vr+.rnonic.r tsr \rnsts vr+.rnonic.r rssrxcr 237
The second, related, metaphorical strategy consists in the use of
goblet words. The zhi ( ) is a wine goblet that tips when it is full,
and rights itself when empty. Shuen-fu Lin has described its meta-
phorical use in the Zhuang-Zi as follows:
The basic meaning of chih [zhi ] is the wine goblet. Chih-yen [zhi yan]
is used as a metaphor for the Taoist ideal use of the mind [xin, heart-
mind] in making speech. The point that is emphasized is that chih
a drinking vessel used as a metaphor for the mindis originally empty
and gets temporarily lled with liquida metaphor for wordswhich
comes from a larger wine container only when the occasion requires
one to do so. Chih-yen, then, is speech that is natural, spontaneous,
unpremeditated, always responding to the changing situations in the
ow of discourse. If one can engage only in this sort of verbal act,
one can keep his mind perpetually in a state of naturalness, harmony,
transparency, and emptiness. Chih has also been interpreted by schol-
ars as a pun on the character chih [zhi ], meaning uneven, irregular,
and random. Understood as such, chih-yen means irregular and ran-
dom words, referring especially to the random comments made by
the implied author . . . on the stories as well as by the characters
within the stories themselves. In terms of the organization of ideas,
each of the Inner Chapters does appear to have this random and
haphazard quality.
Let us see how the structure of Zhuang Zis words resembles the
operation of the goblet. The strategy here is to take a particular dis-
tinction (good/bad, right/wrong, this/that, allowable/not allowable,
self/other, being/nonbeing, beginning/no beginning, true/false, and
so on) and through an outpouring of paradoxes and innite regresses,
empty the heart-mind of the distinction. The following example
comes immediately after Zhuang Zi says that he does not know
whether his words t into the category of other peoples:
an empty mode of fasting and as forgetting. See Oshima, Harold (1983), A
Metaphorical Analysis of the Concept of Mind in the Chuang-tzu, in Experimental
Essays on Chuang-tzu, edited by Victor Mair, Honolulu: Center for Asian and Pacic
Studies, University of Hawaii Press, p. 77; Cua, Antonio (1998), Forgetting Morality,
in his Moral Vision and Tradition, Washington D.C.: The Catholic University Press
of America.
Lin, Shuen-fu (1988), Confucius in the Inner Chapter of the Chuang Tzu,
Tamkang Review, Vol. 18, No. 14, p. 384. See also his (1994) The Language of
the Inner Chapters of the Chuang Tzu, in The Power of Culture Studies in Cultural
History, edited by W.J. Peterson, A.H. Plaks, and Ying-shih Y, Hong Kong: The
Chinese University Press.
238 cn.r+rn xixr
However, let me try making my statement. There is a beginning. There
is not yet beginning to be a beginning. There is a not yet beginning
to be a not yet beginning to be a beginning. There is being. There
is nonbeing. There is not yet beginning to be nonbeing. There is a
not yet beginning to be a not yet beginning to be nonbeing. Suddenly
there is nonbeing. But I do not know, when it comes to nonbeing,
which is really being and which is nonbeing. Now I have just said
something. But I dont know whether what I have said has really said
something or whether it hasnt said something.
The innite regresses have the eect of taking away what was initially
a clear distinction, in this case, between beginning and no beginning
on the one hand, and between being and nonbeing on the other.
Zhuang Zi seems happy with the position that although he has said
something, it could also be said that he has not said anything. Either
he has said something or he has said nothing. But, paradoxically, in
saying that he has said nothing, he has said something. In contem-
porary logic, this is known as a semantic paradox, of which a well
known example is the liars paradox: the cognitive content of This
sentence is false (that is, what it says), cannot be determined.
This strategy enables Zhuang Zi to steer clear of the espousal of
any particular value. This is cashed out, for instance, in terms of
the notion of use. In the Zhuang-Zi, we nd stories about deformed
men, gnarled trees and other objects which are conventionally des-
ignated as useless. In each case, Zhuang Zi posits some other per-
spective from which what has been deemed useless may yet be seen
to have some use. More broadly, what is conventionally regarded as
deformed or ugly may still have the power to attract, or in some
instances, enables one to remain free and to stay clear of harms
way. However, this is not an absolute lesson, as we learn from the
story of the goose that could not cackle (hence deemed useless because
it could not warn its owner of intruders) and was chosen for dinner
over another that could. This story comes directly after Zhuang Zi
tells his disciples about the tree that is deemed worthless and is there-
fore able to live out its years.
Watson, Burton (1968), The Complete Works of Chuang Tzu, Columbia University
Press, p. 43.
Watson, 1968, p. 209. This story is discussed in some detail by Allinson, 1989,
Chapter 12.
vr+.rnonic.r tsr \rnsts vr+.rnonic.r rssrxcr 239
5. Conceptual Metaphor and Chinese Philosophy
Let us resume the earlier discussion of the theory of conceptual
metaphor, as applied to the Zhuang-Zi. I have agreed with Slingerland
that there are certain metaphorical schemas under which the self is
regarded as a container in the Zhuang-Zi. Still, this does not explain
the strategic use of metaphor in the text. The claim that the use of
metaphor in the Zhuang-Zi is unconscious in an embodied way fails
to do justice to the creative employment of metaphorical argument
that is found there. As we have seen, the argument is consistently
held together by the metaphors of the heart-mind as a mirror, and
of goblet words, to logically avoid the particular espousal of any
value in propositional terms. Whether Zhuang Zi succeeds in his
attempt to avoid the statement of any proposition is another mat-
ter. But it should be clear that the sense in which Zhuang Zis words
purport to be non-propositional has to do with the strategy of these
and other metaphorical devices.
It has nothing to do with any
unconscious causal mechanism of projection.
As part of this strategy, it is important to note that the standard
of coherence and consistency is an important issue for Zhuang Zi.
In contrast, under the conceptual metaphor model, this does not
seem to present any problem. Referring to some inconsistencies
between dierent metaphorical schemas, Slingerland states that
Although these various schemas are at times literally inconsistent,
they are generally not incompatiblethat is, they serve to supple-
ment one another and thereby t together to form a coherent con-
ception of the self.
Similarly, he states that As we shall see,
although the various schemas for the self in the Zhuang-Zi are liter-
ally inconsistent, they are not incompatible, and in fact the infer-
ence patterns that they provide t together to motivate a coherent
soteriological strategy.
Presumably, under the theory of concep-
tual metaphor, there is no problem with these inconsistencies, since
the metaphorical schemas are held to be embodied and hence non-
propositional. Nevertheless, there remains talk of inference patterns
See Lin, 1988, for a description of other literary and metaphorical devices
used in the Zhuang-Zi.
Slingerland, 2004, p. 328.
Slingerland, 2004, p. 328.
240 cn.r+rn xixr
that are supposed to t together to motivate a coherent soterio-
logical strategy. If we follow the strategy of metaphorical argument
of the use of goblet words that I have described for the Zhuang-
Zi, and the logical context as well as the context of clarity in which
it is situated, however, there is no need to resort to this forced talk
about the compatibility of inconsistent schemas. That is, Zhuang Zi
either succeeds in maintaining consistency, or he does nothe can-
not have it both ways. Whatever may be the case, the strategy of
goblet words, as we have seen, aims to avoid inconsistency.
The same assertion about the coherence of inconsistent schemas
is made with reference to another philosopher, Xun Zi. Referring
to the Human Nature is Bad chapter of the Xun-Zi,
The fact that some of these metaphoric schemas are literally incon-
sistent does not present a problem for Xun Zi or the reader because
they are conceptually coherent by virtue of their similar or comple-
mentary entailments. For instance, whether our inborn nature is a
place that we leave and to which we do not return or a thing
that we lose and cannot recover, the basic entailment is the same:
that, as we might say in English, there is no going back. Similarly,
although the portrayal of human nature as a substance always shared
by everyone contradicts the metaphor of its being something that we
irrevocably lose, these schemas do not come into direct conceptual
conict because they have very dierent targets (equal opportunity vs.
cannot regress), which means that none of their entailments directly
contradict each other. That is, we could understand human nature as
something shared at birth while still realizing that it is lost as we
mature. In this sense, the HUMAN NATURE AS SHARED MATE-
RIAL metaphor makes explicit an entailment that is at least consis-
tent with, and perhaps implied by, some of the other metaphors: we
all leave from the same place or have the same raw material to
work with.
Contra Slingerland, it is important whether Xun Zi is talking of
human nature as a thing that is lost, or as a place that is left
behind. If there is no going back, this locution belongs to a place
For translations of the Xun-Zi, see Knoblock, John (1988, 1990, 1994), Xun-
Zi: A Translation and Study of the Complete Works, 3 vols., Stanford: Stanford University
Press. The Nature is Evil (or Nature is Bad) chapter is in the 3rd volume,
Chapter 23. Also, Watson, Burton (1963), Hsn Tzu: Basic Writings, New York:
Columbia University Press.
Slingerland, 2003, p. 37.
vr+.rnonic.r tsr \rnsts vr+.rnonic.r rssrxcr 241
and not a thingXun Zi (as Slingerland interprets him) would have
to make up his mind whether he is talking of a place or a thing.
Similarly for the notion of human nature as a shared substance or
as a thing that is irrevocably lost. It is insucient to point out that,
happily, these have either complementary entailments or dierent
targets. Debates over human nature, both ancient and modern, are
notoriously controversial. It matters whether human nature is a sub-
stance, whether it is held to be good or bad, in what sense equal
or unequal, and so on. And if it is held that this substance is irrev-
ocably lost, what does this imply? If what is implied is that we can-
not regress, then what does this in turn mean, and how are we to
proceed in social and political terms? Does the impossibility of regress
mean that we forge ahead and recreate ourselves, and if so, what
are the implications for equality, and so on. Thus, the situation may
not be as convenient as Slingerland makes it out to be. On the con-
trary, the possibility of conict is ever present because of the social
and political implications of the dierent schemas of human nature.
These diculties arise out of Slingerlands interpretation of Xun
Zi. As in the case of the Zhuang-Zi, this interpretation employs the
theory of conceptual metaphor. In evaluating this interpretation, we
should look at the context of Xun Zis arguments. A main target of
Xun Zis critique in the Human Nature is Bad chapter is certain
arguments put forward by Mencius. For example, Xun Zi provides
a diagnosis of Menciuss idea that human nature is good as being
due to a belief in an original, pristine state that is thought to be
inseparable from the resource of goodness, in the same way as cer-
tain sense organs are inseparably linked with their functions (the ears
with hearing, the eyes with sight). Having provided this diagnosis,
he proceeds to deate this idea by saying that the moment one is
born, one has already moved away from whatever original simplic-
ity there may beif indeed there was ever such a state in the rst
place. Thus, Xun Zi is not adhering to a position of human nature
as a location, as in a starting point of a journey and to which one
does not return. Instead, he is debunking what he sees as an assump-
tion held by Mencius. That is, he is denying the existence of a pris-
tine, moral state. Similarly, the idea that Xun Zi holds that human
nature is bad or evil has to be qualied in various ways. First,
there is no intrinsic evil or badness to human beings, in the way
that it is the substance of human beings to be evil or bad. Instead, it
is a fact that human beings have various desires, and under conditions
242 cn.r+rn xixr
of scarcity and non-regulation, the circumstances in which human
beings nd themselves would be very bad indeed. Second, instead
of any particular substance, Xun Zi talks of the cognitive and instru-
mental capacities that allow human beings in general to discover
certain regularities in the world, as well as to create constitutive rules
of ritual that will govern their proper interaction. As part of this,
human beings are said to have the capacity to congregate and make
distinctions. However, human beings have dierent characters and
dierent abilitiesnot everyone is able to shape their capacities to
the same extent. Thus, not everyone has the ability to become a
sage. Everyone may have equal cognitive and instrumental capaci-
ties to begin with. But, according to Xun Zi, it does not follow that
everyone has the same abilities.
To take stock of my argument so far: I have argued that the the-
ory of conceptual metaphor is unable to account for a metaphori-
cal strategy within the Zhuang-Zi, one that attempts to avoid being
stuck with the propositional adherence to any particular value. The
non-propositional sense of Zhuang Zis words can be understood
through the metaphorical device of the heart-mind as a mirror, and
of goblet words. It does not lie in what is alleged by the theory of
conceptual metaphor, that is, the unconscious mapping of the embod-
ied spatial relations of everyday life onto concepts. Furthermore, it
is important to the metaphorical strategy that I have described that
there be no self-contradiction. The theory of conceptual metaphor,
on the other hand, claims that the metaphorical schemas it describes
are inconsistent, but not incompatible. Quite apart from the ques-
tion of its coherence, this notion of inconsistency without incom-
patibility raises the question of whether the interpretations oered
in terms of conceptual metaphor are adequate, as I have illustrated
in the case of Xun Zi. The point should be made more generally.
I have given a detailed account of Xun Zis arguments in relation to Mencius
in Chong (2003), Xunzis Systematic Critique of Mencius, Philosophy East and West
53:2. Note that Slingerland does allow one instance of a metaphorical schema that
entails not sitting well with the other schemas. This is the schema that Human
nature is a human agent that is bad, and we know that it is bad because it desires
or wishes to be good. The problem with this is that it entails saying that human
nature has internal tendencies toward being good, precisely because it is bad. See
Slingerland, 2003, pp. 3738. For an explanation of what this point implies for
Xun Zis position, see Cua, A.S. (1978), The Quasi-empirical Aspect of Hsn Tzus
[Xun Zis] Philosophy of Human Nature, Philosophy East and West 28:1.
vr+.rnonic.r tsr \rnsts vr+.rnonic.r rssrxcr 243
Sometimes, problems of interpretation can arise, quite independently
of the schemas. Thus, whether a particular schema is legitimately
imposed on a certain position depends on understanding that posi-
tion. In other words, the schema cannot, by itself, establish the inter-
pretation that it proposes.
Let us illustrate this with reference to one more example, this time
from the Analects. Consider the following example of metaphorical
usage in this text. In the passage 3.8 of the Analects, Confuciuss dis-
ciple, Zixia, asks for an explanation of the the following lines from
the Odes: Her entrancing smile dimpling, Her beautiful eyes glanc-
ing, Patterns of colour upon plain silk. The following dialogue
between Confucius and Zixia ensues:
The Master said, The colours are put in after the white.
Does the practice of the rites likewise come afterwards?
The Master said, It is you, Shang [Zixia], who have thrown light
on the text for me. Only with a man like you can one discuss the
In his reading of this passage, Slingerland says, Just as all of the
cosmetics in the world are of no avail if the basic lines of the face
are not pleasing, so is the renement provided by ritual practice of
no help to one lacking in good native substance. It is this entailment
that explains Confuciuss concern that cultural adornment be rmly
rooted in its native substrate . . .
There is more than one descrip-
tion of this native substrate in Slingerlands account. First, Slingerland
refers to this as constituting some basic emotions that form the
root of the ritual forms. For instance, that it is important not to lose
touch with emotions such as grief when mourning, instead of being
overly concerned with the formal details. I would agree with Slingerland
on this aspect of the Analects, and this is something worth stressing
in the ethics of Confucius, although in doing so, there is no need
to mention any notion of a native substrate.
However, Slingerland supplements this with a second description
of what constitutes the native substrate. Thus, he cites passages
that would indicate some kind of innate tendency toward the good
Lau, D.C. (1979), Confucius: The Analects, tr., Harmondsworth: Penguin.
Slingerland, 2003, p. 54.
Slingerland, 2003, p. 72.
244 cn.r+rn xixr
or even that innate orientation toward the good is a universal qual-
Slingerland is concerned to show that there is a paradox of
virtue in the Analects: one must already be virtuous in order to
acquire virtue. Also, he nds a tension in the Analects between what
he calls the adornment metaphor, and the craft metaphor. In the
former, the rites do not alter the native substrate, but instead can
only be said to depend for their eectiveness on the prior existence
of the substrate. The craft metaphor, on the other hand, connotes
a radical shaping of a material that has no particular prior mass or
or that requires violent reshaping of the original material.
In describing these metaphors, Slingerland indicates the anities with
Mencius (human nature is originally good) and, at the same time,
with Xun Zi (goodness is the result of accumulated eort and not
the result of any originally good material). Slingerland is not wrong
in pointing to these anities. It is possible to read the Analects in
either a Mencian way, or in a Xunzian way. Again, however, a fur-
ther question arises as to whether it is legitimate to read the ethical
philosophy of the Analects in either of these ways.
For instance, an interpretation of the term zhi as native substrate
and the conception of the rites as an embellishment or adornment
of this substrate is not self-evident. Thus, there is a reading that is
quite the opposite of the reading (in passage 3.8) that it is the white
that comes rst. It could be held that the colors are rst laid down,
and white is used to enhance the colors instead. This reading means
that although one has to have some character (color) in the rst
place, the rites (white) are still required to rene ones character.
Under this reading, the renement is not an adornment of any-
thing, but involves a process of practicing and learning. In taking
this reading, the Qing dynasty philosopher Dai Zhen rmly rejects
any reference to the practice of the rites in terms of adornment. For
him, adornment takes place when a persons feelings are wearing
thin and yet he continues to work on the appearance of things.
As Dai Zhen sees it, the so-called adornment metaphor would be
inappropriate to a description of the purpose of the rites as an edu-
Slingerland, 2003, p. 72.
Slingerland, 2003, p. 55.
Slingerland, 2003, p. 53.
Chin, Ann-ping and Freeman, Manseld (1990), Tai Chen [Dai Zhen] on Mencius:
Explorations in Words and Meaning, trs., New Haven: Yale University Press, p. 156.
vr+.rnonic.r tsr \rnsts vr+.rnonic.r rssrxcr 245
cational and self-cultivational medium. For him, the rites play a more
active and positive role in the formation of character.
It would be a mistake to think that in rejecting the metaphor of
adornment, we are thereby endorsing the craft metaphor. The the-
ory of conceptual metaphor oers these two alternatives: either one
is talking of human nature as being based upon a native substrate,
in which case one is saying that human nature rests upon a certain
universally good substance which needs to be adorned, or else one
talks of human nature as being forged or crafted. However, it is fal-
lacious to assume that we must read Confucius in either of these
modes. Certainly, the metaphorical schemas of adornment and craft
cannot be self-justifying, and the theory of conceptual metaphor can-
not legitimize such a reading independently of other means of under-
standing the Analects. It should be stressed that the point is not that
the interpretation that I put forward is correct, while Slingerlands
is not. Instead, the point is that conceptual metaphor cannot by itself
adequately do the work of interpretation, or force the interpretation.
No doubt Slingerland is right in pointing to certain locutions of, and
anities between, the Chinese philosophers. However, further inves-
tigation that goes beyond the parameters of conceptual metaphor is
necessary, before any proper judgment can be made about the ade-
quacy of an interpretation.
6. Conclusion
It might seem that in doing comparative philosophy, one takes ones
pick: one interprets from some perspective, and there are dierent
models to work with. Thus, it might be held that I am criticizing
the application of the Lako and Johnson model of conceptual
metaphor to Chinese philosophy from the perspective of Davidsons
theory of metaphor. However, so it might be held, the same game
can be played from either side. But what I have said against the
various interpretations of Allinson and Slingerland can be said inde-
pendently of Davidsons view of metaphor. I could have illustrated
the problems with Allinsons and Slingerlands readings of the
Zhuang-Zi, for instance, without any reference to Davidson (or even
In any case, I have not agreed with everything that Davidson says
about metaphor. For instance, that one must make a demarcation
246 cn.r+rn xixr
between metaphor and simile, and that a metaphor cannot be para-
phrased without remainder. It has seemed to me that this latter belief
is unnecessary to the understanding of metaphor, and furthermore,
I have diagnosed this belief as a cause of the tendency to claim that
there is a special cognitive content to metaphor. In addition, Davidsons
corollary idea that a metaphor can intimate any number of things
has to be qualied, if we are not to take this as meaning that any-
thing goes in the interpretation of metaphor. If it is the case that
anything goes, we would have no reason to disagree with, say,
Allinsons reading of the buttery dream.
Despite these dierences, however, I think that Davidsons stress
on metaphor as belonging to the realm of use strikes the right note,
and this is where the signicance of his view of metaphor lies. In
this connection, I have highlighted the fact that Zhuang Zi has a
highly imaginative metaphorical strategy that is missed or inade-
quately referred to by Allinson and Slingerland in their accounts of
metaphor. Generally speaking, the reason why they miss or inade-
quately describe this strategy is that they share a tendency to believe
that there must be an essence to metaphor. I have counterposed
against this what I believe Zhuang Zi, Xun Zi and Confucius were
doing, in the particular contexts within which they were situated. In
other words, I have denied that the theory of conceptual metaphor
is adequate to the understanding of these philosophers. This is not
to say that the structure that the theory of conceptual metaphor
attributes to the way we use metaphors in our daily lives is entirely
wrong. Instead, I have argued against the tendency to apply what
is regarded as essential to metaphor under this theory to all cases.
In particular, I have argued that its application to the examples we
have looked at has not helped their proper understanding. In short,
there is no essence to metaphor. We have to look at the contexts
and arguments in which a particular metaphor is used to appreci-
ate its application and signicance. As Wittgenstein says, instead of
trying to grasp the essence of the thing, one must always ask one-
self: is the word ever actually used in this way in the language-game
which is its original home?
Wittgenstein, Ludwig (1968), Philosophical Investigations, G.E.M. Anscombe tr.,
Oxford: Basil Blackwell, Part I, no. 116.
Yang Xiao
The focus of this paper is on the word and in the title of this
volume Davidsons Philosophy and Chinese Philosophy. I believe scholars
of Chinese philosophy should engage with contemporary philosophy
of language by drawing out its implications in the context of early
Chinese philosophy and language; reading the Analects with Davidson
can shed light on both the communicative practice in the Analects
and Davidsons philosophy of language.
Reading classical Chinese texts with Davidson should also help us
to see why we should not draw conclusions about the nature of com-
municative practice in early China based on observations about the
grammatical and semantic features of classical Chinese. I shall call
this style of reasoning the grammatical approach to pragmatics.
Some scholars have made arguments about what the Chinese can or
cannot do with the classical Chinese language, and their arguments are
based solely on their observations on the grammatical and semantic
An early version of this paper was presented at the conference Davidsons
Philosophy and Chinese Philosophy: Constructive Engagement in Beijing on June
89, 2004. I wish to thank my fellow participants at the conference for their very
helpful comments, particularly A.P. Martinich, Michael Krausz, Koji Tanaka, Stephen
Angle, David Wong, Yujian Zheng, and Samuel Wheeler. My special thanks go to
Bo Mou, both for his admirable work as the organizer of the conference, and for
his unfailing support and patience as the editor of this volume. I also wish to thank
David Keightley, P.J. Ivanhoe, A.P. Martinich, Robert Ashmore, Carine Defoort,
Michael Puett, and Anna Xiao Dong Sun; I am deeply indebted to their insight-
ful comments on earlier versions of this paper. I would like to express special thanks
to my wife Anna, from whom I have learned a great dealperhaps more than I
initially wanted!about what an astonishingly wide range of things a seemingly
innocent utterance in everyday life can mean. My wife is also the connection that
brought me to Davidson. In the late 1990s I moved to Berkeley to be with her,
who was a student there, and audited two of Davidsons seminars. I didnt realize
until later that these seminars have changed my philosophical life. I dedicate this
paper to the memory of Donald Davidson.
248 cn.r+rn +rx
features of classical Chinese. Following Davidson, I suggest that we
should make the pragmatic turn by focusing directly on peoples
linguistic or communicative practice, namely the utterances of sen-
tences in concrete occasions on which the sentences are put to work.
More specically, I will argue against two assumptions in the gram-
matical approach. The rst is what I shall call the empirical assumption,
which asserts that, since classical Chinese is not an inected lan-
guage, it does not have any linguistic device to indicate grammati-
cal moods. The second is what I shall call the mood-force correlation
thesis, which claims that grammatical moods and pragmatic forces
are closely correlated. In other words, the grammatical features, or
any conventional features in general, determine how linguistic expres-
sions can be used pragmatically. Obviously, the correlation thesis
enables one to derive conclusions about pragmatic forces from obser-
vations about grammatical moods.
It is through these two assumptions that I shall engage with an
important debate between Dummett and Davidson in contemporary
philosophy of language. The debate is regarding the relationship
between the grammatical moods of a sentence and the pragmatic
forces of the utterance of the sentence. Dummett endorses the mood-
force correlation thesis that there is a strict correlation between mood
and force, and that illocutionary force is always conventional. Davidson
rejects the thesis, and argues that neither force nor ulterior purpose
of an utterance is governed by linguistic conventions.
Peter Strawson
predicted in 1969 that the conict between the communication-
intention-based pragmatics and convention-based formal semantics
has been, and would continue to be, the Homeric struggle at the
heart of the philosophy of language.
This debate between Dummett
and Davidson can be seen as a continuation or unfolding of this
The Dummett-Davidson debate is anticipated by an earlier debate between
Austin and Strawson concerning the issue of whether force is always conventional;
please see Strawson, P.F. (1971a), Intention and Convention in Speech Acts, in
Logico-Linguistic Papers, London: Methuen, pp. 14969. Dummett has tried to defend
Austin against Strawson; please see Dummett, Michael (1995), Force and Convention,
in The Philosophy of P.F. Strawson, edited by Pranab Kumar Sen and Roop Rehha
Verma, New Delhi: Indian Council of Philosophical Research, pp. 6693, and
Strawsons reply in the same volume, pp. 4037.
Strawson, P.F. (1971b), Meaning and Truth, in Logico-Linguistic Papers, London:
Methuen, pp. 17089. For an excellent brief history of contemporary philosophy
of language from this perspective, please see Martinich, A.P. (1997), Philosophy of
nr.rixo +nr ANALECTS vi+n r.\irsox 249
What I want to show in this paper is that this debate has impor-
tant implications for the study of Chinese philosophy and language.
If Davidson is right, we would have to say that, based on observa-
tions about the grammatical and semantic features of the Chinese
language, one cannot draw any conclusion about what the Chinese
speakers can or cannot do with Chinese sentences; one would have
to base ones arguments directly on observations about the pragmatic
features of their communicative practice.
In Section 1, I will introduce the basic concepts of mood and
force, and I will introduce the grammatical approach, as well as
the two basic assumptions of the approach, namely the empirical
assumption and the mood-force correlation assumption. In Section
2, I will rst respond to the empirical assumption by arguing that
classical Chinese has its own linguistic devices (such as ending par-
ticles) to indicate moods; I will then discuss the limits of this response.
In Section 3, I will discuss Davidsons argument against Dummetts
mood-force correlation thesis and his conventionalist theory of force.
In Section 4, I will use a passage from the Analects to illustrate
Davidsons point that there is no convention of sincerity, which is
at the heart of his argument against Dummetts conventionalist the-
ory. Section 5 concludes the paper by briey exploring the impli-
cations of what Davidson calls the principle of the autonomy of
linguistic meaning.
In the English language we indicate grammatical mood by inecting
the form of the verb. For example, the verb phrase to be immedi-
ately put into practice has at least four inections, which signify
indicative, interrogative, imperative, and subjunctive moods:
Language, in Routledge History of Philosophy, Volume X Philosophy of Meaning, Knowledge
and Value in the Twentieth Century, edited by John Caneld, London and New York:
Routledge, pp. 1138. Martinich convincingly shows that such a conict between
the two approaches to language starts with the debate between Russell and Strawson.
As he puts it, Behind Strawsons objection [to Russells theory of description] is a
view of language that is radically dierent from Russell. For Russell, words and
sentences are the fountains of meaning. For Strawson, people using words and sen-
tences are. For Russell, semantics is the primary object of linguistic study. For
Strawson, it is pragmatics, how people use words (p. 18).
250 cn.r+rn +rx
(1) . . . is being immediately put into practice . . .
(2) . . . is [it] immediately being put into practice?
(3) . . . should be immediately put into practice . . .
(4) . . . were to be (could have been) immediately put into practice . . .
These inected phrases are known as mood-indicators. When they
appear in complete sentences, they indicate four grammatical moods of
a sentence:
(1a) Indicative sentence:
What has just been learned is being immediately put into practice.
(2a) Interrogative sentence:
Is what has just been learned being immediately put into practice?
(3a) Imperative sentence:
What has just been learned should be immediately put into practice.
(4a) Subjunctive (counterfactual) sentence:
What has just been learned were to be (could have been) imme-
diately put into practice.
As we can see, the mood of a sentence is a syntactic feature of the
sentence, independent of any actual uses of the sentence. The mood
of a sentence remains the same even when the sentence is being
used to do dierent things in dierent situations; this is because the
mood is a formal feature of the syntactic structure of a sentence.
Now when a sentence is uttered by a speaker in a specic situa-
tion, the speaker is using it to do certain things. We need another
term to refer to what the speaker is doing with the sentence; the
term is the illocutionary force of the utterance, or simply the force
of the utterance. For example, when the utterance of a sentence is
being used to issue an order, we say that the force of the utterance
is to issue an order.
When one utters the above four sentences, (1a)(4a), one can do
at least four dierent things:
(1b) making an assertion
(2b) asking a question or making a request
(3b) oering advice (issuing an instruction, an order, or a command)
(4b) expressing a wish (or regret)
The way I presented these examples might have given the impres-
sion that there is a strict correlation between the grammatical moods
of a sentence (the interrogative, indicative, imperative, or subjunctive),
and the forces of the utterances of the sentence (asking a question,
describing a fact, oering a piece of advice, or expressing a wish).
nr.rixo +nr ANALECTS vi+n r.\irsox 251
It is indeed true that, when we ask a question, we often use an inter-
rogative sentence; when we describe a fact, we often use an indica-
tive sentence, and so on. Nevertheless, is it really the case that the
interrogative sentences are always used to ask questions, just as imper-
ative sentences are always used to issue an order? In other words,
is there a strict correlation between mood and force?
In Moods and Performances, which was rst presented at a con-
ference in 1976, Davidson tries to answer these questions. He thinks
that the questions can be formulated a little dierently, because the
relationship between mood and force can also be seen as about the
relationships of two ways of classifying utterance:
The moods classify sentences, while uses classify utterances; but the
moods indirectly classify utterances, since whatever distinguishes sen-
tences can be used to distinguish utterances of them. So we may ask,
what is the relation between these two ways of classifying utterances;
how are assertions related to utterances of indicative sentences, for
example, or commands to utterances of imperative sentences?
What Davidson argues against is the mood-force correlation thesis,
which claims that the associated classes of utterances are identical:
utterances of imperatives are commands, utterances of interrogatives
are question-askings, etc.
If Davidson is right that there is no strict
correlation between mood and force, then we should not try to deter-
mine what people can or cannot do with English sentences by look-
ing at the grammatical features of the language; the fact that English
is an inected language with a variety of linguistic devices to indi-
cate grammatical moods becomes unimportant and irrelevant.
Let us now turn to some examples in Chinese. In order to illus-
trate what I call the grammatical approach, let us take a look at a
classical Chinese sentence from 11.22 of the Analects:
(C) Wen si xing zhi .
Davidson, Donald (1984a), Mood and Performance, in Inquiries into Truth and
Interpretation, Oxford: Oxford University Press, pp. 10910.
Davidson, 1984a, p. 110.
All quotations from the Analects are to book and passage numbers in Yang,
Bojun (1980), Lun-Yu-Yi-Zhu [The Analects with Translations and Comments], Beijing:
Zhonghua shuju. All translations in this paper are my own, but I have been aided
greatly by the existing English translations by Simon Leys and D.C. Lau.
252 cn.r+rn +rx
The rst character wen means to hear; the second character si means
this; the third character xing means to practice or to be put into practice;
the last character zhi means itreferring, in this case, what has
just been heard. One may translate the sentence as something like
this: Having heard it, then immediately put it into practice. Or,
What has just been learned should be immediately put into prac-
tice. But there is a problem here: This is just one of many possi-
ble translations.
Part of the problem comes from the fact that neither classical nor
modern Chinese is an inected language. If one believes that inection
is the only way to indicate the mood of a sentence, then there is
no way to determine the mood of this sentence. Hence we can have
at least four English translations for the original Chinese sentence:
(1a) What has just been learned is being immediately put into practice.
(2a) Is what has just been learned being immediately put into practice?
(3a) What has just been learned should be immediately put into practice.
(4a) What has just been learned were to be (could have been) imme-
diately put into practice.
That is to say, the Chinese sentence wen si xing zhi in itself allows
it to be translated into any of these English sentences, each with a
dierent grammatical mood. How do we make sense of these gram-
matical dierences between the Chinese and English languages? One
may argue that, because there are no mood-indicators in classical
Chinese, people must have been confused about illocutionary forces
in ancient China. Or one may conclude that certain speech-acts
(such as expressing a wish) cannot be done, due to the absence of
There would be more possible translations if we take into account time and
number. The English language indicates time and number by inection at every
occurrence of a verb or noun. As A.C. Graham has pointed out, even though
Chinese verbs and nouns have no inection, this does not mean that the classical
Chinese language does not have its own devices to indicate them. In fact, Chinese
indicates time and number by particles only when time and number is relevant. As
Graham argues, we need to be told whether an event is past, present, or future
no more often than is indicated by the temporal particles of Chinese. The idea
that there are confusions in early Chinese thought due to the absence of tense and
singular or plural seems to me quite untenable. At some places one has trouble
rendering into English without committing oneself to tense or number, but this is
merely a translators problem (Graham, A.C. [1978], Later Mohist Logic, Ethics and
Science, Hong Kong: The Chinese University Press, pp. 289). What I argue here
is that what Graham says about time and number also applies to mood in classical
nr.rixo +nr ANALECTS vi+n r.\irsox 253
the corresponding mood-indicators (such as the subjunctive mood
indicator) in classical Chinese.
Two scholars are representative of this kind of reasoning. Alfred
Bloom has argued that, since the Chinese language does not have
a linguistic device for conterfactuality, the Chinese do not have coun-
terfactual thinking. Moreover, since argumentation needs counter-
factuals, the Chinese are incapable of argumentation.
this line of thinking, one can make a similar argument that the
Chinese cannot express wishes, because expressing wishes also needs
Chad Hansen is a much more inuential scholar; he has famously
claimed that the ancient Chinese do not have concepts of sentence,
belief, or truth, and that they never use sentences to describe facts
or to express truths or beliefs; their words are only used to guide
peoples behaviors. His arguments are based on observations about
the syntactical dierences between Chinese and English sentences.
In his observations, he focuses on certain grammatical features of
classical Chinese, which, in comparison to English, are obviously dis-
tinctive and unique. Here is a summary of some of these features:
(1) The absence of sentence function marking in classical Chinese:
The absence of sentence function marking, . . . and the use of predi-
cate-only sentences contribute to viewing all words as having only a
naming function and to the failure to distinguish the sentence as a
functional composite linguistic form.
(2) The lack of grammatical inections in classical Chinese:
Chinese does not have grammatical inections, which in Western lan-
guages, draw attention to the sentence as a compositional unit. . . .
Chinese theories of language did not concentrate on sentences because,
simply, classical Chinese sentencehood is not syntactically important.
Another manifestation of the lack of grammatical inections in clas-
sical Chinese, according to Hansen, is that Classical Chinese does
not have explicit descriptive and prescriptive forms, which can be
Bloom, Alfred H. (1981), The Linguistic Shaping of Thought: A Study in the Impact
of Language on Thinking in China and the West, Hillsdale, New Jersey: L. Erlbaum.
Hansen, Chad (1985), Chinese Language, Chinese Philosophy, and Truth,
Journal of Asian Studies 44 (3), p. 516.
Hansen, 1985, p. 500.
254 cn.r+rn +rx
easily expressed by the inections of the verbs in English.
on these observations, Hansen concludes that the linguistic practices
and theories in China and English-speaking countries are radically
These observations about the dierences between Chinese and English
syntax explain (from a Chinese point of view) why we place so much
emphasis on the sentence, or (from our point of view) why Chinese
philosophers do not.
Classical Chinese does not have explicit descriptive and prescriptive
forms. Students of comparative translation, therefore, will nd huge
chunks of text that one translator renders in declarative English and
another in imperative English. Behind this apparent ambiguity, I sug-
gest, lies this assumption about the function of language. All language
functions to guide behavior.
One of the most striking characteristics of these arguments by Bloom
and Hansen is that they never look directly at the linguistic and
communicative practice; instead they focus on the grammatical fea-
tures of Chinese sentences, and end up with a conclusion about the
nature of Chinese linguistic practice.
Let us now return to our ear-
lier example from 11.22 of the Analects, (C) wen si xing zhi , to illus-
trate this point.
As we have shown, the Chinese verb xing in (C) has no inection,
whereas the English verb phrase to be put into practice has at least
four inections, which correspond to four grammatical moods. That
is to say, for this one Chinese sentence (C), there can be at least
four English translations: (1a), (2a), (3a), and (4a). Let me reiterate
(1a) and (3a) as follows:
(1a) What has just been learned is being immediately put into practice.
(3a) What has just been learned should be immediately put into practice.
Note that (1a) is a descriptive, indicative English sentence, and (3a)
is a prescriptive, imperative English sentence. Like Hansen, one may
feel compelled to conclude that the English-speaking people can dis-
Hansen, Chad (1992), A Daoist Theory of Chinese Thought, Oxford: Oxford University
Press, p. 51.
Hansen, 1985, p. 500.
Hansen, 1992, p. 51; the emphasis is Hansens.
When Hansen says, My hypothesis is that real dierences in the languages can
explain dierences in the popular theories of language (Hansen, 1992, p. 25;
emphasis added), he means the grammatical dierences.
nr.rixo +nr ANALECTS vi+n r.\irsox 255
tinguish between two functions of language: stating facts and guid-
ing behavior, whereas the Chinese see language as having only one
function, .i.e., to guide behavior.
In general, if one does take a grammatical approach, one might
be tempted to argue that, since Chinese is not an inected language,
and since Chinese verbs do not have the grammatical moods such
as indicative, interrogative, imperative, and counterfactual moods,
the Chinese are either incapable of doing things such as making an
assertion, asking a question, issuing an instruction, or expressing a
wish, or they are incapable of telling them apart.
Two assumptions in these arguments are problematic. The rst is
what I have called the empirical assumption, which is that the classical
Chinese language does not have any linguistic device to indicate the
grammatical moods. The second is what we have called the mood-
force correlation thesis, which claims that grammatical moods and prag-
matic forces are closely correlated; in other words, the grammatical
features of linguistic expression determine how they can be used
pragmatically. The correlation thesis enables one to derive conclu-
sions about pragmatic forces from observations about grammatical
moods. In the rest of the paper, I shall deal with these two assump-
tions in turn.
It is relatively easy to respond to the empirical assumption. Bloom
and Hansen seem to presuppose that inection is the only way for
a language to have mood-indicators. However, empirical evidence
shows that classical Chinese has dierent grammatical devices to indi-
cate moods, one of which is through ending particles. These are
words at the end of sentences that have no substantive meaning in
This argument has its logical problems. For example, one cannot derive homo-
geneity from ambiguity, i.e., from the fact that Chinese expressions can have ambigu-
ous or multiple interpretations, one cannot draw the conclusion that there is only
one interpretation. More concretely, if wen si xing zhi can be read either as descrip-
tive or prescriptive utterances, one cannot conclude that this means it is always pre-
scriptive, i.e., to always guide behavior. If one sticks to the grammatical approach,
then, in order to claim that all the uses and functions of sentences in a language
L are to guide behavior, one would have to show that all sentences in L are un-
ambiguously prescriptive ones.
256 cn.r+rn +rx
themselves, and their only function is to indicate the moods of the
Although the transcribers and editors of the Analects do not know
the technical term of ending particles or particles, they do have
a systematic way of using a variety of ending words to indicate the
moods of sentences. One example is the particle hu . It has no
substantive meaning when it appears at the end of a sentence, but
it has an important grammatical function, which is to indicate that
the sentence is an interrogative one. For example, junzi shang yong
(A gentleman prizes courage) is an indicative sentence, but
if we add hu at the end of it, we get junzi shang yong hu
(Does a gentleman prize courage?), which is an interrogative
sentence, used by Zilu to ask a question in 17.23. Another example
is Guan Zhong zhi li (Guan Zhong knows the rituals), which
is indicative, whereas Guan Zhong zhi li hu (Does Guan
Zhong know the rituals?) is interrogative, and is used to ask a ques-
tion in 3.22.
Another interrogative ending particle is zhu . Sentences with
the ending particle zhu are also often used to ask questions:
13.15. Duke Ding asked: One single maxim that can lead a country
to prosperity, is there such a thing (you zhu )? Confucius replied:
[. . .].
Duke Ding said: One single maxim that can ruin a country, is there
such a thing (you zhu )? Confucius replied: [. . .].
As we can see, we can ask whether something exists when the par-
ticle zhu is paired with the verb you (there is) to form the fol-
lowing sentence:
(Q ) You zhu ?
Is there such a thing?
To answer the question, one can give a positive reply by saying:
(A) You zhi .
There is such a thing.
We can nd such a pattern in another passage:
7.35. The Master was gravely ill. Zilu asked permission to oer a
prayer. The Master said: Is there such a practice (you zhu )? Zilu
said: Yes, there is ( you zhi ), and the prayer goes like this: We
nr.rixo +nr ANALECTS vi+n r.\irsox 257
pray to you, spirits from above and spirits from below. The Master
said: In that case, I have been praying for a long time already.
Now we have responded to the empirical assumption on its own
ground by showing that there exist other linguistic devices to indi-
cate grammatical moods in classical Chinese. However, this gram-
matical response is not satisfactory in many ways. I shall mention
just two problems here. The rst is that certain types of mood-indi-
cators are absent in the Analects; for example, although we can nd
interrogative and indicative particles in the Analects, we cannot nd
any imperative particles. Now let us again take the sentence wen si
xing zhi as an example. Earlier in the paper, we mentioned
that there are four possible English translations, two of which are
imperative and indicative sentences. However, there is no gram-
matical or semantic feature in the original sentence that gives us any
information about whether it is indicative or imperative, because in
the pre-Qin and Han periods there was no ending particle indicat-
ing the imperative mood. It is only in much later periods (the Tang
and Song dynasties) that new ending particles such as zhuo and
hao were invented to indicate the imperative mood.
Had we
found wen si xing zhi zhuo in the Analects, we would have
been able to say that this is an imperative sentence.
The second problem is that, with regard to the particles we do
nd in the Analects, even though they can indicate grammatical moods
of the sentences, they do not always tell us what the pragmatic forces
of the utterances are. In other words, there is no correlation between
mood and force in classical Chinese. For example, if we look at all
the sentences that end with the interrogative particle hu in the
Analects, we would nd out that these sentences are not always being
used to ask questions or make requests. Confuciuss very rst utterance
in the rst passage of the Analects has the ending particle hu. One
translator correctly renders it as follows: The Master said: To learn
something and then to put it into practice at the right time: is this
not a joy?
Although it is a grammatically interrogative sentence,
For a variety of examples of such imperative particles, see Luo, Ji (2003), Bei-
Song-Yu-Qi-Ci-Ji-Qi-Yuan-Liu [The Mood-Indicators in the Northern Song Dynasty
and Their History], Chengdu: Bashu shushe, pp. 14076 and pp. 2308. The use
of zhuo as an imperative particle can still be found in many regional dialects today
in China (pp. 14452).
I am using Simon Leys translation here. The translation reects faithfully the
258 cn.r+rn +rx
it is obviously a rhetoric question, which can be readily expressed
by an indicative sentence: It is a joy to learn something and then
put it into practice at the right time. Let us take the interrogative
mood-indicator zhu as another example. The particle zhu appears
14 times at the end of a sentence in the Analects. It turns out that
they are not always being used to ask questions or make requests:
Sometimes the sentence is used to ask a question (see 6.6, 7.35, 9.13,
13.1, and 13.15); sometimes it is just a rhetorical question, which is
equivalent to an assertion (6.30, 12.11, and 14.42).
How do we make sense of these cases? Should we conclude that
classical Chinese is uniquely dierent from all the other languages
because the mood-force correlation thesis does not apply to it? I
believe this is where Davidson comes in. These examples from the
Analects sharply highlight the issue that is at the heart of the Dummett-
Davidson debate, in which Davidson argues against Dummetts mood-
force correlation thesis. For Davidson, communication is possible only
because there is no correlation between the grammatical features of
a language and what people can do with it. Hence, if Davidson is
right, there is nothing unique about the lack of such a correlation
in classical Chinese.
One of Davidsons early arguments against the mood-force correla-
tion thesis is based on the existence of counterexamples. After hav-
ing cited a passage from Dummett, in which Dummett gives his
version of the correlation thesis,
Davidson comments:
[W]hat bothers me is the implied claim that assertion and the indica-
tive mood can be this closely identied. For there are many utterances
of indicative sentences that are not assertions, for example indicative
sentences uttered in play, pretense, joke, and ction; and of course
assertions may be made by uttering sentences in other moods. (Utterances
grammatical mood of the original Chinese sentence. But in order to emphasize that
the force of the utterance is actually a rhetorical question, a better translation might
be: To learn something and then to put it into practice at the right time: isnt this
a joy?
See Dummett, Michael (1973), Frege: Philosophy of Language, London: Duchworth,
pp. 315, 316.
nr.rixo +nr ANALECTS vi+n r.\irsox 259
of Did you notice that Joan is wearing her purple hat again? or
Notice that Joan is wearing her purple hat again may on occasion
simply be assertions that Joan is wearing her purple hat again.) And
similarly for other moods; we can ask a question with an imperative
or indicative (Tell me who won the third race, Id like to know your
telephone number), or issue a command with an indicative (In this
house we remove our shoes before entering).
There is a passage from the Analects that is similar to Davidsons last
example (I use Simon Leyss translation here):
13.18 The Governor of She declared to Confucius: Among my peo-
ple, there is a man of unbending integrity: When his father stole a
sheep, he denounced him. Confucius said: Among my people, men
of integrity do things dierently: a father covers up for his son, a son
covers up for his fatherand there is integrity in what they do.
This is another counterexample to the mood-force correlation the-
sis, because the indicative sentences here are used to issue a normative
instruction, just as in Davidsons example, In this house we remove
our shoes before entering.
Davidson is aware that it is not enough to refute Dummetts mood-
force correlation thesis simply by giving a list of counter-examples;
he has to respond to Dummetts conventionalist version of the the-
sis, which is supposedly capable of dismissing these counterexamples.
I now turn to Davidsons arguments against Dummetts conven-
tionalist theory of force.
In Mood and Performance, Davidson mentions that Dummett
can explain away the counterexamples by saying that they are all
deviant, abnormal or non-serious cases. Dummett claims that it is
normal, natural or serious that indicative sentences are always
used to make assertions, imperative sentences are always used to
issue commands, and interrogative sentences are always used to ask
questions. Davidson argues that Dummetts solution doesnt work:
It is easy to see that an appeal to what is serious or normal does
not go beyond an appeal to intuition. It is no clue to the seriousness
of a command that it is uttered in the imperative rather than the
indicative; similarly, a serious question may be posed in the impera-
tive rather than the interrogative mood. And if normal means usual,
or statistically more frequent, it is dubious indeed that most indicatives
Davidson, 1984a, p. 110.
260 cn.r+rn +rx
are uttered as assertions. There are too many stories, rote repetitions,
illustrations, suppositions, parodies, charades, chants, and conspicuously
unmeant compliments. And in any case the analysis of mood cannot
plausibly rest on the results of this sort of statistical survey.
According to Davidson, a way for Dummett to rescue his thesis is
to drop the concepts of normal or serious cases, and switch to a
conventionalist version of the thesis. Instead of saying that an asser-
tion is an indicative sentence uttered in the normal case, Dummett
could say that an assertion is an indicative sentence uttered under
conditions specied by convention.
Dummett has a specic proposal about the convention for assertions,
which is that assertion consists in the (deliberate) utterance of a sen-
tence which, by its form and context, is recognized as being used
according to a certain general convention.
But for Davidson, this
is just the denition of assertion, not the convention of assertion:
This [proposal of Dummetts] also seems to me to be wrong, though
in a somewhat dierent way. What is understood is that the speaker,
if he has asserted something, has represented himself as believing it
as uttering a sentence he believes true, then. But this is not a con-
vention, it is merely part of the analysis of what assertion is. To assert
is, among other things, to represent oneself as believing what one
Therefore, for Davidson, the real issue is: Can there be a conven-
tion that can always tell us whether a speaker believes in what she
utters? To this question, Davidsons answer is no. His argument goes
like this. Let us assume, for the sake of argument, that we do have
an assertion-indicator such as Freges assertion sign. That is to say,
we have a sign that is not just the formal equivalent of the indica-
tive mood, but also a conventional sign of the force of assertion.
Suppose that we always use this strengthened mood whenever we
make an assertion. Davidson then argues,
Davidson, 1984a, p. 111.
Dummett, 1973, p. 311.
Davidson, Donald (1984b), Communication and Convention, in Inquiries into
Truth and Interpretation, Oxford: Oxford University Press, p. 270. For Dummetts
rebuttal, please see Dummett, Michael (1993), Mood, Force, and Convention, in
The Seas of Language, New York: Oxford University Press, pp. 20323. But I can-
not deal with Dummetts rebuttal here.
nr.rixo +nr ANALECTS vi+n r.\irsox 261
It is easy to see that merely speaking sentence in the strengthened
mood cannot be counted on to result in an assertion; every joker, sto-
ryteller, and actor will immediately take advantage of the strengthened
mood to simulate assertion. There is no point, then, in the strength-
ened mood; the available indicative does as well as language can do
in the service of assertion. But since the indicative is not so strong
that its mere employment constitutes assertion, what must be added to
produce assertion cannot be merely a matter of linguistic convention.
In another essay, Communication and Convention, Davidson for-
mulates the argument as follows:
It is clear that there cannot be a conventional sign that shows that
one is saying what one believes; for every liar would use it. Convention
cannot connect what may always be secretthe intention to say what
is truewith what must be publicmaking an assertion. There is no
convention of sincerity. If literal meaning is conventional, then the
dierence in the grammatical moodsdeclarative, imperative, inter-
rogative, opativeis conventional. These dierences are in the open
and intended to be recognized; syntax alone usually does the job. What
this shows is that grammatical mood and illocutionary force, no mat-
ter how closely related, cannot be related simply by convention.
Note that Davidsons conclusion that there cannot be conventional
indicators for assertion applies to all languages or linguistic practices.
It is not a unique feature of the Chinese language that it does not
have assertion-indicators or force-indicators, because no language
Davidson, 1984a, p. 113.
Davidson, 1984b, p. 270. Here we should be very careful not to take Davidson
as saying that the illocutionary force is a purely private, interior, and mental act.
Elsewhere he does make it clear that this is not what he means: The argument
[for the autonomy of linguistic meaning] has a simple form: mood is not a con-
ventional sign of assertion or command because nothing is, or could be, a con-
ventional sign of assertion or command. The reason for this, it should be stressed,
is not that the illocutionary force of a speech act is a purely mental, interior, or
intentional aspect of the act (1984a, p. 114; emphasis added). Nevertheless, Davidsons
point certainly has to do with the fact that speech act has a mental, interior, or
intentional aspect. Right after the passage cited above, Davidson adds, Of course
assertion or command must be intentional, as must meaning in the narrow sense.
But it is part of the intention that the act should be interpreted as assertive or com-
manding, and therefore part of the intention that something publicly apparent should
invite the appropriate interpretation (p. 114).
262 cn.r+rn +rx
Davidsons point that there is no linguistic convention of sincerity
(or insincerity) can be illustrated through a very interesting example
from the Analects:
17.4 The Master went to Wucheng [where Ziyou was the governor],
where he heard the sound of string instruments and hymns. He was
amused and said with a smile (wan er er xiao ): Why use an
ox-cleaver to kill a chicken? Ziyou replied: Master, in the past I
have heard you say: Gentlemen who cultivate the Way love people;
ordinary people who cultivate the Way are easy to govern. The
Master said: My friends! What Ziyou said is true. My earlier remark
was a joke.
Let me list Confuciuss two utterances, as well as the statement he
makes in the past, as follows:
(1) Why use an ox-cleaver to kill a chicken?
(2) Gentlemen who cultivate the Way love people; ordinary people
who cultivate the Way are easy to govern.
(3) My friends! What Ziyou said is true. My earlier remark was a joke.
There can be a variety of interpretations of the forces of Confuciuss
utterances. Let us start with (1). Many readers would agree with
Ziyous taking (1) as an assertion, and understand (1) as saying that
Ziyou should not have bothered to cultivate ordinary people with
music and rituals. This is analogous to killing a chicken with an ox-
cleaver; Confucius wants to say that Ziyou only needs to cultivate
the gentlemen. But this is inconsistent with Confuciuss own teach-
ing, which Ziyou cites in his response to the Master.
Can there be any conventions to determine that (1) must be an
assertion? As we can see in the passage, the editors of the Analects
actually report that Confucius says it with a smile (wan er er xiao).
Does this indicate that it is not an assertion, but an ironic remark?
Can we then say that it is a convention of insincerity that a speaker
is making an ironic remark rather than an assertion when the speaker
utters the sentence with a smile? It is indeed the case that people
do sometimes tell a joke with a smile, and yet they can also make
an assertion with a smile, and tell a joke with a straight face. One
In 6.14, it is mentioned that Ziyou is the governor of Wucheng. I use Simon
Leyss translation here.
nr.rixo +nr ANALECTS vi+n r.\irsox 263
may claim that the convention of insincerity should include more
factors in order to be a real convention; for example, it is possible
that Ziyou does not detect the force of (1) because he doesnt pay
attention to the tone of Confuciuss utterance.
Davidson does not deny that these factors (such as tone and ges-
ture) play key roles in the making and detecting of assertions, but
he argues that they cant conclusively determine whether the speaker
is sincere:
It must also be conceded that interpreters and speakers of a language
are generally able to tell when an assertion has been made, and that
this ability is an essential part of their linguistic competence. Furthermore,
knowledge of linguistic and other conventions plays a key role in the
making and detecting of assertions. Costume, stance, tone, oce, role,
and gesture have, or may have, conventional aspects, and all these
elements can make a crucial contribution to the force of an utterance.
We may easily allow all this without agreeing that merely by follow-
ing a convention, indicative or imperative utterances become assertions
or commands.
Therefore, our conclusion has to be that there is no convention
telling us that Confuciuss utterance (1) must be an assertion or that
it must be a joke. That is to say, we cannot say that Ziyou must
have got it wrong in his reaction to (1) when he takes it as an asser-
tion, and is puzzled by the fact that it is not consistent with (2). In
fact, it is an entirely plausible scenario that Ziyou has got the force
of (1) right. That is to say, Confucius originally does make an asser-
tion when he utters (1). It is only after Ziyou points out that (1) is
not consistent with (2), which is Confuciuss own belief in the impor-
tance of using rituals and music to cultivate ordinary people, that
Confucius gives a retrospective articulation of the force of the remark,
claiming that (1) is actually intended as an ironic remark and should
have not been taken seriously in the rst place. Since a joke can-
not be inconsistent with an assertion, Confucius can thus explain
away the seemingly contradiction between (1) and (2).
Now let us turn to the interpretations of Confuciuss utterance (3),
which is a remark about the force of (1). There are two possible
readings of the force of (3). The rst is to assume that it is an asser-
tion. The second is to assume that it is still an ironic remark.
Davidson, 1984a, pp. 1123.
264 cn.r+rn +rx
If we take (3) My earlier remark was a joke as an assertion, we
would have to agree that Confuciuss earlier remark, (1), was indeed
a joke. How do we make sense of this? Li Zhi (15211602), a
Ming dynasty scholar, has suggested that Confucius is extremely
pleased about what he has seen in Wucheng, where Ziyou is the
governor, and that is why he intentionally makes a fan yu
(ironic remark).
In other words, Li Zhi here is alluding to an inter-
esting phenomenon, which is that when people are extremely pleased
they often feel that they have to use irony to express it. But Ziyou
fails to understand Confuciuss intention, and, as Li Zhi puts it, gets
very serious about it. Confucius then has to get serious as well,
making an assertion regarding the force of (1) to clarify his intention.
Can we determine conclusively that (3) must be an assertion? As
Dummett would suggest, we can get clues from the manner in which
(3) is uttered, or from certain linguistic conventions such as the gram-
matical indicators. The editors report that the Master speaks with a
smile when he utters (1), but they do not say anything about the
manner in which the Master utters (3). Its simply The Master said.
However, as we have argued earlier, this kind of description does
not necessarily mean that Confucius is making an assertion.
Now let us look at the grammatical indicators, in this case, the
ending particles. In the original Chinese version of (3), we nd two
sentences with ending particles: there is the particle ye at the
end of the sentence What Ziyou said is true, and there is the par-
ticle (er ) at the end of the sentence My earlier remark was a
joke. Traditional Chinese scholars agree that ye and er are two
typical indicative particles ( jue-ci ).
But this does not mean that
these two utterances in (3) must be assertions. As Davidson has
argued, we cannot say that the indicative sentences are always used
Li, Zhi (1975), Si-Shu-Ping (Comments on the Four Books), Shanghai: Shanghai
renmin chubanshe, p. 146.
Li, 1975, p. 146.
For example, we nd the following observation from the preface to a Yuan
Dynasty monograph on particles: hu , yu , ye , zai , fu , these
are inquisitive particles (yi-ci ); yi , er , yan , ye , these are assertive
particles ( jue-ci ) (Lu, Yiwei [1988], Zhu-Yu-Ci-Ji-Zhu [Collected Comments
on the Particles], edited and commented by Wang Kezhong, Beijing: Zhonghua
shuju, p. 183). Scholars believe that Lu Yiwei was the rst to write a book-length
study of the particles. We do not know much about the author except that the
book was written no later than 1324.
nr.rixo +nr ANALECTS vi+n r.\irsox 265
to make assertions in normal and serious situations. Davidsons
point becomes especially obvious in our case: There is absolutely no
use to appeal to the notion of the serious situations, because whether
this very situation is a serious one is exactly what we are trying to
determine here.
Would the appeal to convention help? Dummett suggests that an
assertion is an indicative sentence uttered under conditions specied
by linguistic conventions. Our case here shows that Davidson is right
to claim that Dummetts suggestion wont work. If there is any con-
vention that can help us decide whether (3) is an assertion or a joke,
it would have to be the non-linguistic ones, such as the conventional
image of Confucius as always being deadly serious, or the conven-
tional wisdom that the Analects is a collection of Confuciuss sincere
moral instructions and commands. Christopher Harbsmeier has shown
convincingly that, in the Analects, we can often nd Confucius teas-
ing his students, playing with words, and amusing people by saying
things jokingly. Contrary to the conventional image, Confucius is
actually an impulsive, emotional, and informal man, a man with
wit and humor, a man capable of subtle irony with an acute sensi-
bility for subtle nuances.
In his comments on 17.4, Harbsmeier
suggests that Confucius is probably still joking when he makes his
last remark: My earlier remark was a joke. In other words, Confucius
might have been joking all the way through.
I have argued elsewhere that traditional Chinese scholars also believe that the
syntactic features of a sentence do not determine the pragmatic uses of the sen-
tence. Here is one example. In his 1687 commentary on Lu Yiweis book on par-
ticles, the Qing Dynasty scholar Chen Lei says, [The ending particle] zai indicates
interjection, or interrogation, or interruption, or assertion, or just the completion
of the utterance. We should always look at the total context of the speech and text;
we cannot just focus on the word zai (Lu, 1988, p. 17). Other Qing scholars
who have written on particles, such as Yuan Renlin, Wang Yinzhi, and Liu Qi,
have made similar observations. It can be argued that Chinese scholars do not nec-
essarily see these particles as just grammatical mood indicators. In fact, when the
Chinese scholars write about particles, they discuss not only their syntactic func-
tions, but also (and even more often) their functions and uses in composition, style,
persuasion, rhetoric, and argumentation.
Harbsmeier, Christoph (1990), Confucius Ridens: Humor in the Analects, Harvard
Journal of Asiatic Studies 50, 1990, p. 131. I am grateful to David Keightley for hav-
ing urged me to read this article.
One way to make sense of this possibility is to think about Harold Pinters
plays. Besides the Analects, Pinters plays can be read as another massive set of exam-
ples of how conventionally simple utterances, such as I dont know, You are right,
Yes, Well, or non-utterances (pause, silence) are able to do a wide range of things
266 cn.r+rn +rx
Relying on Davidsons arguments, we can see how it is possible to
make dierent judgments about whether certain utterances in the
Analects are assertions or jokes.
Obviously, whether an utterance is
a joke or an assertion has great implications when we interpret a
text. I believe what Bernard Williams has to say about how to inter-
pret Platos Theaetetus applies to the Analects as well: If we are going to
get the most from reading one of Platos dialogues, we have to keep
in close touch with its tone, sustaining a sense of what is a joke,
what is merely provisional, what is being tired out or tried on.
Let me summarize my arguments in this paper. If we take a prag-
matic perspective, we will realize that it is not enough to know the
grammatical mood of the sentences, for the mood is the feature of
a sentence that remains the same, regardless of the dierent situa-
tions in which it is uttered, whereas the force is the feature of the
utterance that varies from situation to situation. To put the point in
a nutshell, there is no strict correlation between grammatical mood-
indicators of a sentence and the forces of the utterances of the sen-
tence, because the speaker can always intend to use the sentence to
do things that are not determined by its grammatical or conven-
tional features. That is to say, we must take into account the total
speech situation in which the speaker makes the utterance.
In other words, the grammatical or conventional features of lin-
guistic expressions do not determine how they can be used prag-
matically, and this applies to all languages. Davidson takes this general
in our daily, domestic life. The force of a sentence in a script, as every good direc-
tor or actor knows, is not determined by the literal meanings of the sentence or
any other linguistic conventions, and hence can always be interpreted (and deliv-
ered) dierently. Directors and actors thus can always have a new interpretation of
a play in a new production.
Elsewhere I have shown that generations of commentators in China have made
dierent judgments regarding the forces of the utterances in the Analects; please see
Xiao, Yang (2006), The Pragmatic Turn: Articulating Communicative Practice in
Early China, Oriens Extremus.
Williams, Bernard (1999), Plato, New York: Routledge, p. viii.
The term total speech situation is from Austin, J.L. (1975), How to Do Things
with Words, Cambridge: Harvard University Press, Second Edition, p. 52 and
p. 148. He emphasizes that [i]t is important to take the speech-situation as a
whole (p. 138).
nr.rixo +nr ANALECTS vi+n r.\irsox 267
point as a basic trait of human language and calls it the autonomy of
linguistic meaning:
What this argument illustrates is a basic trait of language, what may
be called the autonomy of linguistic meaning. Once a feature of lan-
guage has been given conventional expression, it can be used to serve
many extra-linguistic ends; symbolic representation necessarily breaks
any close tie with extra-linguistic purpose. Applied to the present case,
this means that there cannot be a form of speech which, solely by
dint of its conventional meaning, can be used only for a given pur-
pose, such as making an assertion or asking a question.
In another place, Davidson states the thesis of the autonomy of lin-
guistic meaning as follows: Once a sentence is understood, an utter-
ance of it may be used to serve almost any extra-linguistic purpose.
An instrument that could be put to only one use would lack auto-
nomy of meaning; this amounts to saying it should not be counted
as a language.
People were quite shocked when they rst heard Davidson declar-
ing at the end of his 1986 essay A Nice Derangement of Epitaphs
that there is no such thing as a language. This sentence, taken
out of its context, does sound shocking. But let us cite the whole
passage here:
I conclude that there is no such thing as a language, not if a language
is anything like what many philosophers and linguists have supposed.
There is therefore no such thing to be learned, mastered, or born
with. We must give up the idea of a clearly dened shared structure
which language-users acquire and then apply to cases. And we should
try again to say how convention in any important sense is involved in
language; or as I think, we should give up the attempt to illuminate
how we communicate by appeal to conventions.
There is nothing one should feel shocked about if one is familiar
with Davidsons thesis of the autonomy of meaning, rst defended
in his 1976 essay Mood and Performance. One should see clearly
Davidson, 1984a, pp. 1134. Please also see 1984b, p. 274.
Davidson, Donald (1984c), Thought and Talk, in Inquiries into Truth and
Interpretation, Oxford: Oxford University Press, p. 164.
Davidson, Donald. (1986), A Nice Derangement of Epitaphs, in Philosophical
Grounds of Rationality: Intentions, Categories, Ends, edited by Richard Grandy and Richard
Warner, Oxford: Clarendon, p. 174. This paper, with comments by Ian Hacking
and Michael Dummett, is also included in Truth and Interpretation: Perspectives on the
Philosophy of Donald Davidson, edited by Ernest LePore, Oxford: Blackwell, 1986.
268 cn.r+rn +rx
that Davidson is simply saying that we should not focus on words
and sentences, or the conventions and rules that are supposed to
govern them; instead, we should look at peoples communicative
practice, and how they actually do things with words and sentences.
And no convention can capture our communicative practice. Since
classical Chinese can be easily perceived as lacking a clearly dened
shared structure which language-users acquire and then apply to
cases, reading the Analects together with Davidson should make
and I hope, has madeit much easier for us to see that we should
give up the attempt to illuminate how we communicate by appeal
to conventions.
Chung-ying Cheng
Introductory Remarks
In this article I shall advance in the rst part the Donald Davidsons
theory of truth interpretation in terms of the Convention T as orig-
inally established by Tarski in his formalization of semantics for nite
Davidson has expanded the Tarskis formal-semantic
notion of truth interpretation and applies to natural languages in
order to show how the meaning of a sentence could be given in
terms of truth conditions of the sentence. This is an ingenious move.
But we wish to ask whether the Davidsonian approach to interpre-
tation of truth in terms of the Tarskian Convention T is sucient
to explain meaning and truth in a living natural language like the
Chinese and whether it is itself should be further broadened for both
semantic and hermeneutical representation of our understanding real-
ity based on experience. In other words, we shall ask whether an
ontological re-interpretation and application of Convention T could
serve an important purpose of ontological and cross-ontological under-
standing that I call onto-hermeneutical understanding.
See Donald Davidsons book (1984), Inquiries into Truth and Interpretation, Oxford:
Oxford University Press (1991 printing), p. 85. Specically the Essay 4 is titled
Semantics for Natural Languages, Essay 1, Theories of Meaning and Learnable
Languages and Essay 5 In Defense of Convention T are also directly relevant.
A. Tarski proposes the idea of Convention T in his article The Concept of Truth
in Formalized Languages, published in Polish in 1933, in German in 1936, and
in English in 1956 (in Logic, Semantics, Metamathematics, Oxford: Clarendon Press).
See my explanation of the notion of onto-hermeneutical interpretation in my
article Inquiring into the Primal Model: Yijing and the Onto-Hermeneutical Tradition,
in Journal of Chinese Philosophy, vol. 30, double issue 34, pp. 285288. In a situa-
tion where both understanding and communication are important, one must inter-
pret the situation in light of ones experience of reality in the situation together
with both ones belief and knowledge of the situation. It is an interpretation that
is both hermeneutical because it involves understanding of language, and ontolog-
ical because it involves reference to a reality.
272 cn.r+rn rrr\rx
In the second part of the article I shall advance ve basic prin-
ciples for understanding meaning and truth in Chinese language in
view of Chinese philosophy as a theory of truth. To do this is to
show how meaning and truth in Chinese language or more precisely
how meaning and truth of sentences made in Chinese language could
have a theory of truth that explain their meaning and truth in a
framework of Chinese philosophical reection formulated in the same
language. Thus we may regard these ve principles as a theory of
truth of the language, namely a theory that gives the categories of
truth conditions for whatever meaningful sentences can be made in
the language. There is no doubt that I am inspired in reformulat-
ing the ontology of Chinese language in this manner by Donald
Davidson in view of his work on the semantics for natural languages
as mentioned above. Davidson of course was himself inspired by
Tarski s work on formal semantics that is developed for the pur-
pose of clarifying the meaning structure of sentences in a formal lan-
guage in terms of the truth conditions of the sentence to be given
in the same language (as a meta-language). It is conceivable that the
meta-language could be a dierent kind of language other than the
object language, but the sentences of the object language must be
assumed to be translatable into the meta-language in order to speak of
the truth conditions of the later.
In this sense the ve principles of
a theory of truth for Chinese language belong to the meta-language
of Chinese language, which would direct how a meaningful sentence
in the Chinese language should be translated. The novelty of my
approach is that I make the assumption that I have come to a truth
theory of those ve principles by way of comprehensive observation
and comprehensive reection on the use of the language.
Tarskis Convention T demands that each sentence of object has a translation
in the meta-language. Tarski also proposes that natural languages are essentially
inter-translatable. Davidson mentions this in his essay In Defense of Convention
T (Ibid., p. 72). Davidson also argues for rejecting the idea of conceptual scheme
in his article On the Very Idea of a Conceptual Scheme, Essay 13 in his book
Inquiries into Truth and Interpretation. The insight here is that any two systems of lan-
guage have to face reality as a common base for each system to make its inter-
pretation, of not only the reality that is transcendentally presupposed by any system,
but of each other in light of ones interpretation of the reality. An alternative lan-
guage therefore receives a putative translation and interpretation in a given lan-
guage in light of the given language s experience of reality.
rnov rox.rr r.\irsoxs tsr or cox\rx+iox + 273
An immediate question arises: How about a language that can-
not be translated into the given meta-language? The reply is: One
has to interpret the alien language in terms of what we know about
our own language in light of whatever evidence or reason there is
available. This is what Quine has suggested in his radical transla-
tion of an unknown word Gavagai in a local language in view of
the collateral information a linguist may gather in a situation of
observation or confrontation.
After explaining these principles, we may see how Chinese phi-
losophy plays a part in explaining the meaning and references of
Chinese language in general. For our purpose we just assume that
Chinese philosophy could be said to play the role as a theory of
truth for the Chinese language, because it could lay down the truth
conditions for sentences said in the language. The truth conditions
could be conveyed by the same sentence the name of which is used
in the object language. It could be also conveyed by a dierent sen-
tence in the meta-language that interprets the sentence in the object
language due to ones understanding. One must also bear in mind
two matters in Chinese philosophy concerning things in the world.
For one thing, the concept of truth is not treated independently of
our concepts or talk of the real, because there is no intention to
separate truth from reality that is in turn not separable from expe-
rience of natural things in the world. For another, the lack of a uni-
versal copula in the Chinese language to make reference to an abstract
universal being can only be explained by the fact that Chinese phi-
losophy begins with observi